Risk Analysis, Vol. 21, No.

1, 2001

Evaluation of the Physical Hazards Associated with

Two Remedial Alternatives at a Superfund Site
Paul K. Scott,1* Alex Pittignano,2 and Brent L. Finley3

This study presents an evaluation of the risks due to the physical hazards associated with two
remedial alternatives for a former chemical manufacturing facility in New Jersey. Both the
on-site and off-site risk of work-related fatalities during remedy implementation and the risks
of accident or accident-related fatalities during the off-site transport of site-related materials
were evaluated. The two remedial alternatives evaluated were on-site containment and excavation with off-site incineration. The risk of at least one fatality due to a work-related accident was estimated for on-site activities associated with each remedial alternative, and for
off-site incineration. The risks of at least one accident and of one accident-related fatality
were calculated with accident and fatality data from the U.S. Department of Transportation.
In addition, the risk of at least one accident that might potentially affect a natural resource
(e.g., river, lake, or national park) was evaluated. This evaluation indicates that the risk of a
work-related fatality is over an order of magnitude higher, and the risk of an accident or accidentrelated fatality is over three orders of magnitude higher, for the excavation/off-site incineration remedial alternative than for the on-site containment alternative. Overall, this study indicates that the physical hazards associated with excavation and off-site incineration are much
greater than those associated with on-site containment for this site. Therefore, if a choice between the two remedial alternatives were to be made based solely on physical hazards and accident risk, the on-site containment alternative would be more protective of human health
and the environment than the excavation/off-site incineration alternative.
KEY WORDS: Physical hazard; work-related fatality; accident risk; evaluation of remedial alternatives

1. INTRODUCTION

remedies. Current U.S. Environmental Protection

Agency (USEPA) guidance(1,2) recommends methods
for quantitatively assessing chemical health risk using
standard exposure and risk assessment techniques.
There is, however, no guidance for quantitatively
evaluating the nature and degree of the physical hazards (injuries and fatalities) that might be associated
with remedy implementation. In addition, few peerreviewed papers present the quantitative assessment
of physical hazards associated with remedy implementation.(325) One of these studies, Hoskin et al.,(3)
concluded that the risk of one fatality occurring during typical remediation activities, such as excavation

The National Oil and Hazardous Substances

Contingency Plan (NCP) identifies several criteria
that must be evaluated when selecting a remedial alternative for a Superfund site. One of these criteria
requires an assessment of the potential short- and
long-term health risks associated with the proposed
1

Exponent, Inc., Pittsburgh, PA.

Chemical Land Holdings, Inc., East Brunswick, NJ.
3
Exponent, Inc., Oakland, CA.
*Address correspondence to Paul K. Scott, Exponent, Inc., 106
Bailey Avenue, Pitsburgh, PA 15211; pscott@exponent.com.
2

53

0272-4332/01/0200-0053$16.00/1 2001 Society for Risk Analysis

54
or capping, may be higher than those associated with
chemical exposure at hazardous waste sites. Cohen et
al.(4) compared the risk of worker fatality calculated
by Hoskin et al. with the cancer risks incurred by residents living near a hazardous waste site using the
Years-of-Potential-Life-Lost (YPLL) metric. They
found that the YPLL for remediation workers was
higher than that for the nearby residents. Given the
tremendous number of man-hours that are required
to implement some of the longer term remedial options in addition to the significant use of heavy machinery and, in some cases, the nations highway and
rail system (for transporting contaminated material),
it would seem that in many instances the physical hazards of remedy implementation might significantly
exceed the chemical risk. Therefore, to fully comply
with the spirit of the NCP, a quantitative physical hazard evaluation should be an important component of
the risk management decision-making process for
remedy selection at Superfund sites.
A complete evaluation of physical hazards
would consider on-site remediation activities and any
off-site treatment or disposal activities involved in the
remedy. On-site physical hazards consist of worker injuries and fatalities, and the probability of these incidents occurring can be estimated in a fairly straightforward manner. Consideration of off-site activities
can become very complex. If a significant amount of
material must be transported to (e.g., clean fill) or
from (e.g., contaminated debris) the site, then worker
and nonworker (e.g., pedestrians) transportation accidents and fatalities must be considered. These could
consist of road, rail, and shipping transportation incidents. In addition, the potential for spills of contaminated materials into heavily populated or pristine
areas would need to be considered. Additionally, one
must consider worker accidents and fatalities associated with treatment of the contaminated material at
an off-site location (landfilling, incineration, etc.).
This article presents a case-study comparison of
the physical hazards associated with two different remedial alternatives for a former 2,4,5-Trichlorophenoxyacetic acid manufacturing facility in New Jersey. The
primary purpose of remediation at this site is to reduce exposures to dioxin-contaminated soil and other
site material (building walls, scrap, etc.). The estimated
amount of dioxin-contaminated material ranges from
94,720 to 167,500 tons. The first alternative, which is
the USEPA-approved remedy for the site, involves
on-site containment and capping of contaminated material. The second involves excavation, transportation,
and off-site incineration of contaminated material.

Scott, Pittignano, and Finley

The second alternative was rejected by the USEPA in
the Record of Decision (ROD) for the site because it
was deemed to pose too great a chemical health risk to
workers and residents. Nonetheless, it would be useful
to determine whether a physical hazard analysis supports the USEPAs decision. Also, such an evaluation
would be beneficial because these two alternatives
represent the two general categories of remedial options that are often considered for hazardous waste
sites. One alternative emphasizes on-site containment
with minimal material movement or exposure, while
the other involves significant removal and treatment.
In this analysis, the risk of worker fatalities and
transportation-related fatalities and accidents that
could occur during implementation of the two remedies are examined. Also evaluated is the potential for
spills into natural resource areas. This analysis is focused on a comparison of physical hazard risk and
does not include the risks associated with chemical
exposure during remediation.
Because the interest is in comparing the relative
differences in physical hazards between the two remedies, the probabilities of at least one fatality and of
at least one accident were used to measure risk.
While other measures could be used to compare the
relative physical hazard risks between these two remedies, these two are the simplest to understand, easiest to calculate, and have been used previously for a
similar comparison of another site.(3) If physical hazard and chemical health risks were being compared
together, the YPLL or similar risk measures would be
more appropriate.(4) The article concludes with an
evaluation of the merits of these two remedies and a
discussion as to how such quantitative analyses might
be better incorporated into the decision-making process in the future.
2. METHODS
A detailed list of the activities that are necessary
to implement each remedy is presented in Table I. In
addition, the following other assumptions were made
for the physical hazard calculations. Currently, the
only stationary incinerator in the United States permitted by the USEPA to incinerate dioxin-contaminated material is located in Coffeyville, Kansas. For
this exercise, it was assumed that this facility meets all
of the appropriate regulatory requirementscurrent
Resource Conservation and Recovery Act (RCRA)
permit, Maximum Achievable Control Technology
(MACT) standards, and so forthto incinerate dioxincontaminated material at the time of remedy imple-

Evaluation of Physical Hazards

55

Table I. Activities Necessary to Implement Each Remedial Alternative

On-site containment
Installation of a trench cutoff wall (slurry wall) and flood wall
Drum and tank content stabilization
Groundwater treatment and extraction
Off-site disposal of groundwater treatment residuals
Demolition of remaining structures
On-site placement of containerized materials and debris
Cap installation

Excavation/off-site incineration
Installation of a new bulkhead sheet piling around the laid-lock perimeter
of the site
Dewatering of the site
Excavation of site soil
Demolition of structures
Decontamination of scrap metal
Sorting and sizing of debris prior to shipment
Importation and placement of clean fill
Transportation of excavated soil and other site-related material to
incinerator in Coffeyville, KS
Incineration of site-related material

mentation. It was also assumed that site-related material would be transported to Coffeyville via rail; recyclable metals would be transported to a local recycling
center via truck; and clean fill would be imported from
a nearby location via truck. For this alternative, the volume of soil to be excavated (i.e., the volume of soil and
other site-related material that exceeds a cleanup
level of 1 ppb) is 97,374 yd3, or 167,500 tons.
2.1. Evaluation of Work-Related Fatality
The source of work-related fatality data used in
this analysis is Causes of Death in the Workplace,(2)
which contains work-related fatality rates for several
occupations by Standard Industrial Classification
(SIC) job title. This book compiles the work-related fatality rates from the Bureau of Labor Statistics Supplementary Data System (SDS), which contains selected
states worker compensation files. While Leigh(6) does
not present the work-related fatality rates specifically
for hazardous waste site remediation workers, fatality
rates are presented for the trades and occupations that
are needed to implement a remedial alternative. While
other sources of work-related fatality data exist,(7,8)
these sources were not used because they do not contain fatality rates for each specific trade and occupation needed, and they are not substantially different
than those from Leigh.
To calculate the probability of at least one workrelated fatality, the methods presented by Hoskin et
al.(3) were used. When the probability of failures, in
this case fatalities, is small and the number of trials
(opportunities for fatality) is large, the Poisson probability distribution can be used to model the number
of deaths per year given the expected number of
deaths per person-year (m).(3) The expected number
of deaths can be calculated as

m 5

^H

i51

Fi,

(1)

where Hi is the number of person-years associated

with job category i, and Fi is the fatality rate associated with job category i.
The probability of at least one fatality occurring
can be calculated based on the Poisson probability
distribution function as
Probability of at least one fatality 5 1 2 e

2m

. (2)

The fatality rates used for this analysis do not account

for the possible detrimental effects of the personal
protective equipment (PPE) that workers are required to wear by Occupational Safety and Health
Administration (OSHA) regulations. These effects
include heat stress, reduced vision and hearing, and
loss of mobility. Compared with unequipped individuals, workers who must use PPE will be subject to a
considerable risk of heat stress and will most likely
experience a decline in job performance due to the
loss of mobility.(3) Consequently, the actual fatality
rates for remediation activities performed with PPE
should be higher than those for unequipped workers.
Person-year estimates for both remedies were
based on standard engineering practices and sitespecific factors, for example, structures to be demolished, size of the site, and so forth. The person-year estimates for the truck driver are based on the number
of miles driven during the remedy and the assumed
speed of the truck (55 mph). The person-year estimates for the railroad worker are calculated assuming
one worker per 6-day trip to the incinerator. The incinerator person-years are calculated based on an
incineration rate of 30 tons/day and the total tonnage
of site-related materials to be incinerated. Table II

56

Scott, Pittignano, and Finley

Table II. Calculation of Risk of Fatality Due to Work-Related Hazards for the On-Site Containment and
the Excavation/Off-Site Incineration Remedial Alternatives

On-site containment
SIC code
11
85
245
245
245
245
394
424
421
436
436
436
441
441
481
666
715
751
751
780

Excavation/off-site
incineration

Job category

Fatality ratea

Projected
person-yearsb

Weighted
fatality ratec

Projected
person-yearsd

Weighted
fatality ratec

Engineering Staff
Safety officer
Project manager
Superintendent
Company employees
USEPA representative
Timekeeper
Crane operator
Cement finisher
Pile driver
Equipment operator
Treatment plant operator
Foreman
Piling foreman
Mechanic
Incinerator worker
Truck driver
General laborer
Decon attendant
Railroad worker

3.28E-05
0.00E100
5.94E-05
5.94E-05
5.94E-05
5.94E-05
3.70E-06
1.77E-04
6.60E-05
2.16E-04
2.16E-04
2.16E-04
1.17E-04
1.17E-04
5.40E-05
2.33E-04
3.88E-04
3.29E-04
3.29E-04
7.62E-04

7.50
2.15
2.15
2.15
8.50
2.95
2.15
0.35
0.31
1.15
8.68
2.19
3.68
0.22
2.06
NA
1.06
21.4
1.30
NA

3.52E-06
0.00E100
1.83E-06
1.83E-06
7.22E-06
2.50E-06
1.14E-07
8.85E-07
2.92E-07
3.54E-06
2.68E-05
6.76E-06
6.16E-06
3.60E-07
1.59E-06
NA
5.89E-06
1.01E-04
6.11E-06
NA

11.7
3.35
3.35
3.35
13.3
4.60
3.35
0.35
NA
0.86
13.6
3.61
5.75
0.16
3.21
3,150
2.41
33.4
2.03
74.2

1.15E-07
0.00E100
5.98E-08
5.98E-08
2.36E-07
8.21E-08
3.73E-09
1.86E-08
NA
5.57E-08
8.78E-07
2.34E-07
2.02E-07
5.67E-09
5.20E-08
2.20E-04
2.81E-07
3.30E-06
2.00E-07
1.70E-05

Total projected person-years:

Expected number of fatalities per person-yeare:
Probability of at least 1 fatalityf:

70
1.23E-02
1.22E-02

3,330
8.09E-01
5.55E-01

Note: SIC 5 standard industrial classification; NA 5 not applicable to the remedial alternative.
a
Fatality rates as listed in Leigh.(6)
b
Person-year estimates based on standard engineering practices.
c
Weighted fatality rate is (percent of total person-years) 3 (fatality rate).
d
Person-year estimates based on standard engineering practices and volume of soil to be excavated, transported, and incinerated.
e
Expected number of fatalities is (sum of the weighted fatality rates across job category)/(total projected person-years).
f
Probability of at least 1 fatality is 1 2 exp[2(expected number of fatalities per person-year)].

presents the job-specific person-year estimates, the jobspecific fatality rates, and the probabilities of at least
one fatality for these two alternatives. In addition,
Fig. 1 presents the probabilities for at least one workrelated fatality for the two remedial alternatives.
2.2. Evaluation of Transportation-Related
Fatalities and Accidents
In addition to the work-related fatalities for all
worker categories, the physical hazards associated
with other endpoints specifically related to transportation, such as vehicle accidents, fatalities during vehicle accidents, and accidents that could affect natural resources, were evaluated. To estimate the total

number of accidents, the total number of truck or rail

miles traveled was multiplied by an accident-permile-year rate from the U.S. Department of Transportation (DOT).(9,10) These values are summarized
and published by the DOT every year. This yields an
expected number of accidents, or m, which is then
converted to a probability of at least one accident according to Equation (2) in the previous section.
Fatalities during accidents were calculated in a
similar manner. For fatalities during accidents, a statespecific or railroad-specific fatality-per-mile-year
rate taken from the DOT is multiplied by the total
number of miles traveled,(9,10) yielding an expected
number of fatalities during accidents, or m. These fatality rates include the truck drivers, railroad work-

Evaluation of Physical Hazards

57

Fig. 1. Summary of work-related fatalities associated with each remedial alternative.

ers, truck and rail passengers, drivers of other vehicles involved in the collision, and pedestrians. Based
on the fatality data, 96% of railroad fatalities involved persons not employed by the railroad, and
47% involved highway rail-crossing incidents.(10)
The probability of vehicle accidents in an area of
natural resource importance during remedy implementation can also be estimated. For the purposes of
this analysis, a natural resource is defined as a major
water body, state/national park, or state/national forest. Any form of transport that brings site-related material within 0.5 mile of a natural resource is assumed
to constitute a potential scenario in which a vehicle
accident could affect a natural resource, and the human and ecological receptors associated with it. The
0.5-mile radius was selected because it is the standard
radius of effect used in transportation risk assessment.(11213) The per-mile accident rates used to evaluate accidents in natural resource areas are the same
as those used for all accidents. The exception is that
the number of miles used for this endpoint is equal to
the total number of miles the vehicle travels through
a natural resource area, rather than the total mileage
per trip.
The only significant volume of transportation associated with the on-site containment remedial alternative involves the transport by truck of groundwater
residuals to a landfill, and of scrap metal to a recycling center. While a small amount of residual material will have to be incinerated (,1%), this amount is

not significant, and transport to the incinerator was

not evaluated for this alternative. Transportation accidents under the excavation/off-site incineration remedial alternative could occur during transport of
materials to the recycler, the transport of clean-fill to
the site, and the transport of contaminated materials
by rail from the site in New Jersey to the incinerator
in Kansas. Using the soil volume estimates (excavated and imported), the vehicle miles traveled (to
the clean-fill source and to the recycler), and the appropriate incident rates for New Jersey yields the
traffic accident rates for transporting clean fill to the
site, and for transporting material to the recycler.
Table III presents the total mileage, mileage
through natural resource areas, accident rates, and
probabilities of at least one accident for the two alternatives. Table IV presents the total mileage, accident
fatality rates, and probabilities of at least one accident-related fatality. Figures 2 and 3 present the
probabilities of at least one accident and of at least
one accident-related fatality, respectively.
3. DISCUSSION
Table II presents the person-years associated
with each worker category for the on-site containment and excavation/off-site incineration remedial
alternatives. Based on this table, implementation of
the excavation/off-site incineration alternative would
require person-years that far exceed those of the on-

58

Scott, Pittignano, and Finley

Table III. Calculation of Risk of Transportation-Related Accidents for the Two Remedial Alternatives

Scenario

Total
mileage
(miles)

On-site containment alternative

Waste to landfill
36,560
Metal to recycler
80,736
Total

Mileage
Expected number
in natural
of accidents
resource
Expected number
in a natural
areas
Accident rate
of accidents
resource area
(miles) (accidents/mile)a,b
(accidents)c
(accidents)c

Risk of
at least 1
accidentd

Risk of at
least 1 accident
in a natural
resource aread

8,480
NA

1.28E-09
1.37E-09

4.68E-05
1.11E-04

1.09E-05
NA

4.68E-05
1.11E-04
1.58E-04

1.09E-05
NA
1.09E-05

Excavation/off-site incineration alternative

Clean fill to site
265,562
NA
Metal to recycler
2,880
NA
Soil to incinerator 1,498,650 515,000
Total

1.37E-09
1.37E-09
3.43E-06

3.64E-04
3.95E-06
5.14E100

NA
NA
1.77E100

3.64E-04
3.95E-06
9.94E-01
9.94E-01

NA
NA
8.29E-01
8.29E-01

Note: NA 5 Not applicable because no natural resources are located along route.
a
Accident rate for any type of vehicle accident is mileage-weighted rate based on rates for each state or each railroad line.
b
Accident rate for accidents involving potential natural resource impacts same as those for total accidents. Trip mileage includes areas of
route near a surface water body, state/national park, or state/national forest.
c
Expected number of accidents is (accident rate) 3 (total mileage) or (accident rate) 3 (mileage in natural resource areas).
d
Risk of at least 1 accident is 1 2 exp[2(expected number of accidents)] and 1 2 exp[2(expected number of accidents in a natural resource area)].

site containment alternative. This remedial alternative would require over 6.5 million man-hours, or
3,330 person-years to implement. Conversely, the onsite containment alternative would require approximately 140,000 man-hours, or 70 person-yearsonly
2% of the time needed for the excavation/off-site incineration alternative. The largest contribution to the
total person-years needed for the excavation/off-site
incineration alternative is from the incinerator workers, who contribute almost 95% of the total personyears. Therefore, the primary cause for the large
number of person-years necessary for this remedial

alternative is the time and labor needed to incinerate

the soil and contaminated material after it is removed
from the site.
The job-specific fatality rates also provide a useful point of comparison for the two remedial alternatives (Table II). Two of the occupations with the highest fatality ratesthe incinerator operator and the
railroad workerare needed for the excavation/offsite incineration alternative, but not for the on-site
containment alternative. The large number of railroad worker and incinerator operator person-years
required to complete the excavation/off-site inciner-

Table IV. Calculation of Risk of Transportation-Related Fatalities for the Two Remedial Alternatives
Accident fatality rate
(deaths/mile)a

Expected number
of fatalities (deaths)b

Risk of at least
1 fatalityc

36,560
80,736

7.76E-10
1.63E-09

2.84E-05
1.32E-04

2.84E-05
1.32E-04
1.60E-04

Excavation/off-site incineration alternative

Clean fill to site
265,562
Metal to recycler
2,880
Soil to incinerator
1,498,650
Total

1.63E-09
1.63E-09
1.65E-06

4.33E-04
4.69E-06
2.47E100

4.33E-04
4.69E-06
9.16E-01
9.16E-01

Scenario
On-site containment alternative
Waste to landfill
Metal to recycler
Total

Total mileage
(miles)

Accident rate for any type of vehicle accident is mileage-weighted rate based on rates for each state or each railroad line.
Expected number of accidents is (accident fatality rate) 3 (total mileage).
c
Risk of at least 1 fatality is 1 2 exp[2(expected number of fatalities)].
b

Evaluation of Physical Hazards

59

Fig. 2. Summary of transportation-related accidents associated with each remedial alternative.

ation alternative, coupled with their relatively high

fatality rates, are the primary reasons why the jobrelated fatality risk for this alternative is over an order
of magnitude higher than that for the other. These
risks indicate that the probability of at least one fatality occurring during the implementation of the excavation/off-site incineration alternative is approximately 45 times higher than that for the other
remedial alternative. In addition, 95% of the total
weighted fatality rate for the excavation/off-site incin-

eration alternative is due to the incinerator worker,

while the worker category with the largest contribution for the on-site containment alternativethe general laboreronly contributes about 57% to the total
weighted fatality rate.
Even more dramatic are the differences between
the two remedies regarding the probability of at least
one transportation accident, accident-related fatality,
or accident in a natural resource area. As shown in
Tables III and IV and Figs. 2 and 3, in all cases the

Fig. 3. Summary of transportation-related accident fatalities associated with each remedial alternative.

60
probability of these events occurring during the excavation/off-site incineration alternative is at least
5,000 times greater than in the on-site containment
remedial alternative. Also, the probability of these
events occurring during the excavation/off-site incineration alternative is over 50%, a very high likelihood caused by the large number of trips involved.
Hence, it is clear that the excavation/off-site incineration alternative would be associated with a significantly greater number of deaths and accidents.
It is also important to note that trucking the soils
to the incinerator would not significantly decrease
the risk of fatality or accident. In fact, the risk of at
least one accident occurring during truck transport of
site-related materials to the incinerator is higher than
that for rail transport. Assuming the same mileage
per trip, it would take 7,614 one-way, 22-ton truck
trips to the incinerator to haul the entire tonnage of
site-related material to the incinerator. Using an
overall accident rate for the United States of 9.49 3
1027 accidents per mile per year,(9) this gives an accident risk of over 99% and an expected number of accidents of 10.5.
Finally, it should be noted that while the on-site
containment alternative can be implemented over a
relatively short time frame, the excavation/off-site incineration alternative would require several years to
complete. Based on an incinerator burn rate of 30
tons/day, 5,580 days or 15.3 years of continuous 24hour/day burning would be required to incinerate all
the site-related material. Given the inherent scheduling conflicts with rail transportation, the storage limitations of the incinerator, and periods of incinerator
inoperation, it is likely that 40 years or more might be
required to incinerate all the site-related material. In
short, not only would more total fatalities occur during the implementation of the excavation/off-site incineration remedial alternative, they would continue
to occur for a time period far beyond the time interval necessary to complete the on-site containment remedial alternative.
The actual numbers of chemical spills and catastrophic accidents were not considered here but
could be quantified. While the chances of such incidents occurring under the on-site containment alternative are minimal, there are several plausible scenarios under which persons or natural resources
could be acutely exposed to site-related material during the shipment of over 100,000 tons of highly contaminated material across seven states. Other accidents or releases could also occur as a result of
incinerator failure.

Scott, Pittignano, and Finley

Finally, the issue of significant property damage
was not assessed here, but is a plausible scenario. Specifically, under the excavation/off-site incineration alternative, significant subsidence could occur due to
the excavation of the majority of surface soil from the
site, resulting in damage to the foundations of surrounding properties. Such an episode would not occur under the on-site containment alternative.
3.1. Uncertainties Associated with Physical Hazard
Risk Calculations
There are several uncertainties related to using
the worker fatality data presented in Leigh(6) to estimate the physical hazard risks during remedy implementation. These uncertainties are discussed in detail
in Hoskin et al.(3) The impact of using these fatality
rates versus fatality rates from other sources such as
Table 2 from the Bureau of Labor Statistics,(7,8) Leigh,(14)
or Meng(15) is minimal because the fatality rates used in
this analysis are less than twofold (50% to 70%) lower
than the rates from those other sources.
The main uncertainty in the estimate of at least
one accident in a natural resource area is the definition of a natural resource. For these risk estimates,
the potential for an accident in a natural resource
area was assumed to be possible only when the truck
or railroad car traveled within 0.5 miles of a major
water body, national/state park, or national/state forest. Therefore, if the natural resource area was not a
water body or a designated national or state park or
forest, it was not included in the analysis. In all likelihood, the number of miles traveled through or near a
natural resource will be higher because other areas of
natural resource importance, such as wetlands, parks
and reserves, and unused land not designated as a
park or reserve, were not included because they are
not readily identifiable from route maps.
Another uncertainty in calculating the probability of at least one accident or one accident-related fatality is the total rail mileage used, which was based
on the assumption that site material is shipped every
day during remedy implementation. The amount of
soil that can be incinerated per day, however, is much
smaller than the amount that can be excavated and
shipped each day. In addition, there is a limit to the
amount of contaminated soil that can be stored
awaiting incineration. If the capacity of the incinerator to store and burn the soil is used as the limiting
factor rather than how much can be shipped from the
site per day, many smaller shipments of contaminated
material are needed for this remedial alternative,

Evaluation of Physical Hazards

whether the material is shipped by rail or by truck.
Thus, the actual probability of at least one accident,
fatality, or accident in an area of natural resource importance will be much higher than the estimates presented here.
Because the probability of at least one accident
is used as a risk measure, the previous uncertainty
could lead to an absurd solution to mitigate this
source of risk. Because the probability of at least one
accident is based on the total mileage, lowering the
total mileage can lower this risk. For example, if all of
the waste is placed in one trainload, the probability of
at least one accident will fall to approximately 5.0 3
1023. The severity of any accident would be higher,
however, given the larger amount of dioxin-contaminated material carried on this train. Because this risk
measure is insensitive to the severity of an accident, it
is important to understand the assumptions behind
the calculation. While the probability of one accident
will be higher if smaller shipments of contaminated
material are made, the severity of these accidents will
be lower due to the decreased shipment size.
Finally, the main limitation of this analysis of
physical hazards is that the physical hazard risks cannot be compared directly to the human health and
ecological risks associated with exposure to the dioxincontaminated material. Because the probability of at
least one fatality or at least one accident is used as the
risk measure, the physical hazard risks are not directly comparable to the human health and ecological
risks associated with chemical exposures. To combine
physical hazard risk with chemical exposure risks, other
risk measures, such as YPLL or Quality-Adjusted Life
Years can be used.(4) To incorporate natural resource
losses, risk measures would have to be converted into
the dollar values of the predicted losses.
4. CONCLUSIONS
In this study, the risks of a work-related fatality, an
accident during transport, and an accident-related fatality were calculated and compared for two remedial
alternativesan on-site containment alternative and
an excavation/off-site incineration alternative. This
evaluation indicated that the physical hazards associ-

61
ated with the excavation of site soil, transportation, and
off-site incineration were much higher than those associated with on-site containment. Thus, the on-site containment alternative is more protective of worker
health and safety, and is much less likely to lead to a potential accidental release of contaminated material offsite than the off-site incineration alternative.

REFERENCES
1. U.S. Environmental Protection Agency (USEPA). (1988).
Guidance for conducting remedial investigations and feasibility
studies under CERCLA (USEPA 540/G-89/004). Washington,
DC: Office of Emergency and Remedial Response.
2. U.S. Environmental Protection Agency (USEPA). (1991).
Risk assessment guidance for Superfund: Vol. 1. Human health
evaluation manual (Part C, Risk evaluation of remedial alternatives) Interim. (PB92-963334). Washington, DC: Office of
Emergency and Remedial Response.
3. Hoskin, A. F., Leigh, J. P., & Planek, T. W. (1994). Estimated
risk of occupational fatalities associated with hazardous waste
site remediation. Risk Analysis, 14, 10111017.
4. Cohen, J. T., Beck, B. D., & Rudel, R. (1997). Life years lost at
hazardous waste sites: Remediation worker fatalities vs. cancer deaths to nearby residents. Risk Analysis, 17, 419425.
5. Mar, T., Frost, F., & Tollestrup, K. (1993). Physical injury risk
versus risk from hazardous waste remediation: A case history.
Regulatory Toxicology and Pharmacology, 17, 130135.
6. Leigh, J. P. (1995). Causes of death in the workplace. Westport,
CT: Quorum Books.
7. U.S. Bureau of Labor Statistics. (1994). Census of fatal occupational injuries. Washington, DC: Author.
8. U.S. Bureau of Labor Statistics. (1995). Census of fatal occupational injuries. Washington, DC: Author.
9. U.S. Department of Transportation. (1996). Highway statistics
1995. Washington, DC: Federal Highway Administration.
10. U.S. Department of Transportation. (1996). Accident/incident
bulletin no. 164. Calendar year 1995. Washington, DC: Federal
Railroad Administration.
11. U.S. Department of Transportation. (1990). Present practices of
highway transportation of hazardous materials (FHWA-RD-89018). Washington, DC: Federal Highway Administration.
12. Mills, G. S., & Neuhauser, K. S. (1999). Statistical evaluation of
population data for calculation of radioactive material transport accident risks. Risk Analysis, 19, 613619.
13. Mills, G. S., & Neuhauser, K. S. (1998). Truck transport of
RAM: Risk effects of avoiding metropolitan area. In Proceedings of the 12th International Conference on the Packaging and
Transportation of Radioactive Materials (PATRA 98). Paris,
France.
14. Leigh, J. P. (1985). Fatal occupational injuriesTexas, 1982.
Morbidity and Mortality Weekly Register, 34(10), 130134, 139.
15. Meng, R. (1991). How dangerous is work in Canada? Estimates of job-related fatalities in 482 occupations. Journal of
Occupational Medicine, 33, 10841090.