Hazard Ranking System Issue Analysis:
Sites with Unknown Waste Quantity
MITRE
-------
Hazard Ranking System Issue Analysis:
Sites with Unknown Waste Quantity
Lawrence M. Kushner
August 1986
MTR-86W83
SPONSOR:
U.S. Environmental Protection Agency
CONTRACT NO.:
EPA-68-01-7054
The MITRE Corporation
Metrek Division
7525 Colshire Drive
McLean, Virginia 22102-3481
-------
Department Approval:
MITRE Project Approval:.
7
ii
-------
ABSTRACT
This report presents a review of the status of hazardous waste
sites with unknown waste quantity under the EPA Hazard Ranking System
(HRS). Data are presented on the number of sites with unknown waste
quantity, the distribution of their HRS scores as compared to the
distribution of HRS scores for sites at which the waste quantity is
known, and the most frequently designated site activities for unknown
waste quantity sites. The significance of the waste quantity factor
score in the HRS is discussed and the consequences of the current
policy of assigning a waste quantity default value of 1 are examined.
Several possible alternatives to that policy are identified, examined
and commented on.
iii
-------
TABLE OF CONTENTS
Page
LIST OF ILLUSTRATIONS vii
LIST OF TABLES viii
1.0 INTRODUCTION 1
1.1 Background 1
1.2 Issue Description 3
2.0 APPROACH 7
3.0 UNKNOWN WASTE QUANTITY SITES THROUGH UPDATE #5 9
3.1 Total Number and Distribution by HRS Score 9
3.2 Site Activities 12
4.0 SENSITIVITY OF HRS SCORE TO WASTE QUANTITY 17
4.1 Structure of the HRS Scoring System 17
4.2 Statistical Analysis of Sensitivity 19
4.3 Analysis of the Data Through Update #5 20
5.0 DISCUSSION OF ALTERNATIVES 25
5.1 Continuing Present Use of a Default Value of 1 25
5.2 Using a Different (But Fixed) Default Value 28
5.3 Using a Site-Activity-Dependent Default Value 29
5.4 Getting Sampling Data from Which Waste Quantity 37
Could be Estimated
5.5 Dropping Waste Quantity as a Scoring Factor 38
6.0 SURROGATES FOR WASTE QUANTITY 41
7.0 DISCUSSION AND RECOMMENDATIONS 45
-------
LIST OF ILLUSTRATIONS
Figure Number Page
3-1 Sites Through Update #5, by HRS Score and 10
Knowledge of Waste Quantity
3-2 Distribution of HRS Scores for All Sites 11
and Sites With Unknown Waste Quantity
(Through Update #5)
4-1 Sensitivity of HRS Scores <28.50 to Waste 21
Quantity Default Value
5-1 Frequency of Waste Quantity Factor Values 26
Through Update #5 (Sites With Unknown
Waste Quantity Excluded)
5-2 Distribution of Sites by Selected Site 27
Activity (Through Update #5)
5-3 Distributions of Waste Quantity Factor 30
Values for Sites With Known Waste Quantity
and Designated Solely as Landfill (Municipal),
Landfill (Commercial/Industrial), Surface
Impoundment, or Waste Piles--(Through
Update #5)
5-4 Distributions of Waste Quantity Factor 31
Values for Sites With Known Waste Quantity
and Designated Solely as Spill, Containers
and Drums, Tanks (Above Ground) or Tanks
(Below Ground)--(Through Update #5)
5-5 Distributions of Waste Quantity Factor 32
Values for Sites With Known Waste Quantity
and Designated Solely as Chemical Processing/
Manufacturing, Wood Preserving, or Other
Manufacturing/Industrial--(Through
Update #5)
7-1 Distribution of Waste Quantity for Sites 50
Through Update #5 (Sites With Unknown Waste
Quantity Excluded)
Vll
-------
LIST OF TABLES
Table Number Page
1-1 Hazardous Waste Quantity Factor Values 5
3-1 Distribution of HRS Scores for All Sites 13
and Sites With Unknown Waste Quantity
(Through Update #5)
3-2 Percentage of Unknown Waste Quantity Sites 14
in the NPL Data Base (Historical Development)
3-3 Frequency of Site Activity Designations-- 15
Sites With Unknown Waste Quantity (Through
Update #5)
4-1 Effect of Altering WQ Default Value on HRS 22
Scores of Sites With HRS Score <28.50
(Through Update #5)
5-1 Parameters of Waste Quantity Factor 34
Distributions for Various Types of Sites
(Through Update #5)
5-2 Waste Quantity Default Values, by Type of 36
Site
viii
-------
1.0 INTRODUCTION
1.1 Background
The Comprehensive Environmental Response, Compensation, and
Liability Act of 1980 (CERCLA) (PL 96-510) requires the President to
identify national priorities for remedial action among releases or
threatened releases of hazardous substances. These releases are to
be identified based on criteria promulgated in the National
Contingency Plan (NCP). On July 16, 1982, EPA promulgated the Hazard
Ranking System (HRS) as Appendix A to the NCP (40 CFR 300; 47 FR
31180). The HRS comprises the criteria required under CERCLA and is
used by EPA to estimate the relative potential hazard posed by
releases or threatened releases of hazardous substances.
The HRS is a means for applying uniform technical judgment
regarding the potential hazards presented by a release relative to
other releases. The HRS is used in identifying releases as national
priorities for further investigation and possible remedial action by
assigning numerical values (according to prescribed guidelines) to
factors that characterize the potential for any given release to
cause harm. The values are manipulated mathematically to yield a
single score that is designed to indicate the potential hazard posed
by each release relative to all other releases. This score is one of
the criteria used by EPA in determining whether the release should be
placed on the National Priorities List (NPL).
-------
During the original NCP rulemaking process and the subsequent
application of the HRS to specific releases, a number of technical
issues have been raised regarding the HRS. These issues concern the
desire for modifications to the HRS to further improve its capability
to estimate the relative potential hazard of releases. The issues
include:
• Review of other existing ranking systems suitable for
ranking hazardous waste sites for the NPL.
• Feasibility of considering ground water flow direction and
distance, as well as defining "aquifer of concern," in
determining potentially affected targets.
• Development of a human food chain exposure evaluation
methodology.
• Development of a potential for air release factor category
in the HRS air pathway.
• Review of the adequacy of the target distance specified in
the air pathway.
• Feasibility of considering the accumulation of hazardous
substances in indoor environments.
• Feasibility of developing factors to account for
environmental attenuation of hazardous substances in ground
and surface water.
• Feasibility of developing a more discriminating toxicity
factor.
• Refinement of the definition of "significance" as it
relates to observed releases.
• Suitability of the current HRS default value for an unknown
waste quantity.
• Feasibility of determining and using hazardous substance
concentration data.
• Feasibility of evaluating waste quantity on a hazardous
constituent basis.
-------
• Review of the adequacy of the target distance specified in
the surface water pathway.
• Development of a sensitive environment evaluation
methodology.
• Feasibility of revising the containment factors to increase
discrimination among facilities.
• Review of the potential for future changes in laboratory
detection limits to affect the types of sites considered
for the NPL.
Each technical issue is the subject of one or more separate but
related reports. These reports, although providing background,
analysis, conclusions and recommendations regarding the technical
issue, will not directly affect the HRS. Rather, these reports will
be used by an EPA working group that will assess and integrate the
results and prepare recommendations to EPA management regarding
future changes to the HRS. Any changes will then be proposed in
Federal notice and comment rulemaking as formal changes to the NCP.
The following section describes the specific issue that is the
subject of this report.
1.2 Issue Description
In applying the HRS, a release is regarded as threatening public
health, welfare, or the environment by migration of hazardous
•jf
material away from the site by three possible routes --ground
The HRS also takes into account the threat from direct contact
with the waste and from fire and explosion; however, only the
threat from migration away from the site is used for ranking the
sites for possible remedial action.
-------
water, surface water, and air. A migration score for each applicable
route is calculated by scoring the site with respect to a number of
factors that characterize (a) the hazardous substances at the site,
(b) the containment of the hazardous substances, (c) the potential
for migration of the hazardous substances from the site by that
route, and (d) the presence and proximity of targets (i.e., human
populations or sensitive ecological systems or environments).
Finally, the individual route scores are combined to give an overall
HRS score for the site. The scoring system is structured so that HRS
scores can range from 0 to a maximum of 100, the higher the score the
higher the relative threat ascribed to the site. Under present
policy, sites with HRS scores greater than or equal to 28.50 are
included on the National Priorities List (NPL) and are eligible for
remedial funding under CERCLA.
One of the scoring factors for waste characteristics is waste
quantity (WQ) . It is defined as the total amount of hazardous waste
at a site, other than that which is fully contained and therefore can
not migrate. Hazardous waste is defined as wastes or other materials
that, as received at the site, contain one or more hazardous
substances as defined in CERCLA, Section 101. Thus, any non-
hazardous component of such waste is included in HRS estimates of
waste quantity. The values that can be assigned to WQ range from 0
to 8, according to Table 1-1.
-------
TABLE 1-1
HAZARDOUS WASTE QUANTITY FACTOR VALUES
Tons/Cubic Yards
0
1-10*
11-62
63-125
126-250
251-625
626-1250
1251-2500
2500
No. of Drums
0
1-40*
41-250
251-500
501-1000
1001-2500
2501-5000
5001 10000
>10000
Assigned Value
0
1
2
3
4
5
6
7
8
In practice, a quantity of hazardous waste greater than 0
but no more than 10 tons/cubic yards or 40 drums is given an
assigned value of 1.
Source: EPA, Appendix A of the National Contingency Plan
(47 FR 31219, July 16, 1982, or 40 CFR 300).
-------
As discussed more fully in later sections, the importance of
waste quantity in the HRS is two-fold. It is one of several scoring
factors which bear on the magnitude of the daily dose of a released
substance that an exposed individual might receive, and/or the length
of time (several hours to many years) that the exposure might last.
More importantly, however, it is the only scoring factor that scales
positively with both of them.
Since CERC1A generally applies to abandoned or inactive sites,
it was to be expected that there would be sites at which it would be
difficult, if not impossible, to make an actual determination, or
even a credible estimate, of waste quantity, even though there might
be clear evidence that hazardous wastes were present. To deal with
such sites, EPA adopted the policy of assigning WQ a default value
of 1. This policy has now been identified by EPA staff as warranting
review and analysis. This report presents such a review and
analysis. It also identifies, examines, and comments on a number of
alternative policies available to EPA for scoring unknown waste
quantity sites.
-------
2.0 APPROACH
The review described in this report involved the following
steps:
• developing data on the number of sites with unknown waste
quantity, their distribution by HRS score, and their most
frequently observed site activities (e.g., landfill, surface
impoundment, etc.);
• examining the sensitivity of the HRS scores for such sites
to the choice of default value;
• examining the consequences of adopting alternative policy
options available to EPA, including:
continuing the present policy,
using another (fixed) default value,
using a default value that could be different from site
to site depending on the site activity,
getting sampling data at the site from which waste
quantity could be estimated,
dropping WQ as a scoring factor;
• examining the feasibility of identifying one or more
surrogates for waste quantity that could be used when no
credible estimate was available; and
• drawing conclusions and making recommendations (if any) that
are justified by this study.
The statistical data used throughout this report were obtained
from the computerized NPL data base maintained by MITRE. The data are
for sites in the data base through NPL Update #5 for which both the
&
HRS scores and supporting data from site investigation reports had
been entered into the automated data base. There were 1286 such
These would include Preliminary Assessments, Site Inspection
Reports and other investigative reports of a similar nature.
-------
sites. In addition, remedial investigation reports (RIs) for unknown
waste quantity sites were reviewed in order to see what additional
information they might provide concerning waste quantity at those
sites.
The NPL data base includes all of the sites whose HRS scores have
been calculated and were submitted to EPA Headquarters. Since 1983,
EPA guidance has only directed that sites with HRS scores greater than
25.00 be submitted to Headquarters as NPL candidates; thus, since the
adoption of that guidance, sites with HRS scores less than 25.00 have
for the most part not been included in the data base. As a result,
the distributions shown in this report, while valid for the set of
1286 sites used in the study, are not fully reflective of the universe
of sites that were scored. Where this may be particularly important
to the discussion, special note is made in the text. An additional
point is that not all of the data in the NPL data base have been
subjected to quality assurance checks. While this may affect the
presented numerical information to some degree, the effects should not
be of sufficient magnitude to influence the report's major findings.
-------
3.0 UNKNOWN WASTE QUANTITY SITES THROUGH UPDATE #5
3.1 Total Number and Distribution by HRS Score
Of the 1286 sites in the NPL data base through Update #5 that
JL.
constituted the data base for this study, 965 have scores > 28.50
and 321 have scores < 28.50 (Figure 3-1). Of the total, 269 showed no
information for waste quantity and had been scored with the waste
quantity default value of 1. Sites with unknown waste quantity make
up 21.2 and 19.9 percent, respectively, of the sites above and below
the 28.50 cut-off score. The data discussed in the next paragraph
show that this approximately equal representation of unknown waste
quantity sites in both groups is simply an accidental consequence of
the choice of 28.50 as the cut-off point for the NPL. It does,
however, suggest that the combination of a cut-off HRS score of 28.50
and a WQ default value of 1 is not resulting in gross under- or
over-representation of sites with unknown waste quantity on the NPL.
(As will become clear in subsequent discussion, however, some under-
representation probably does occur.)
Figure 3-2 shows the distribution of HRS scores for all sites,
and for sites with unknown waste quantity. As is to be expected, use
of a default value of 1 for waste quantity keeps the sites with
This figure is larger than the number of sites actually proposed
or promulgated on the NPL through Update #5, because it contains
some sites that are ineligible for listing; for instance, sites
regulated under the Resource Conservation and Recovery Act (RCRA).
-------
8
CO
0) CM
® Al
$ CO
O
Irt
CO
. OJ
v
$
IS
i
000
900
800
700
600
500
400
300
200
100
100
200
300
400
snn
—
—
—
—
—
—
__
"••
—
—
JJJ
MlS
—
^•269
JE2ZZ Sues V\
^•1286
/ith Unknown
y//S/> Waste Quantity
FIGURE 3-1
SITES THROUGH UPDATE #5, BY HRS SCORE AND
KNOWLEDGE OF WASTE QUANTITY
10
-------
CO
&
cr>
0)
_Q
300
250
200
150
100
50
o
^
I
^
Total Number of Sites
in HRS Score Range
Number of Sites With Unknown
Waste Quantity in That Range
n
co
o>
o
in
co
CO
t
o
in
co'
en
o
in
co
co
C\J
I
o
in
co
CO
OJ
I
o
in
co
CM
o
in
CO
CJ
en
••t
CO
CO
o
in
co
o
in
CO
co
CO
o
in
CO
in
o
in
CO
CO
LO
I
o
in
CO
in
en
s
t
o
in
co
in
CO
co
o
in
s
o
in
co'
CD
s
o
in
HRS Score
FIGURE 3-2
DISTRIBUTION OF HRS SCORES FOR ALL SITES AND SITES
WITH UNKNOWN WASTE QUANTITY (THROUGH UPDATE #5)
-------
unknown waste quantity out of the highest score ranges. The largest
number of unknown waste quantity sites occurs in the 33.50-38.49
score range. In this range, their percentage, 34.2 percent, is also
higher than in any other score range (Table 3-1).
Table 3-2 presents data on the occurrence of sites with unknown
waste quantity in the historical development of the NPL. Since the
promulgation of the original NPL list (Update #0) and NPL Update #1,
there has clearly been an increase in the percentage of sites with
unknown waste quantity among those being considered for subsequent
updates. This is not surprising, since one might expect that the
first sites reported would have been those about which the most was
known. The data suggest that in the future one might expect from
about 20 to 30 percent of the sites proposed by the states or by EPA
to have an unknown waste quantity.
3.2 Site Activities
The most frequently designated site activities for sites with
unknown waste quantity are given in Table 3-3. The single most
frequently cited activity is "well field," being identified for
nearly 25 percent of the unknown waste quantity sites.* Landfills,
As a matter of site-scoring policy, well fields are defined as
having an unknown source, and therefore, by implication, the waste
quantity is unknown. The fact that only 95.7 percent, rather than
100 percent, of the sites designated as well fields are shown as
having an unknown waste quantity reflects two sites scored early
in the NPL program at which the source was unknown, but waste
quantity was estimated from ground water plume data.
12
-------
TABLE 3-1
DISTRIBUTION OF HRS SCORES FOR ALL SITES AND SITES
WITH UNKNOWN WASTE QUANTITY (THROUGH UPDATE #5)
HRS Score
Range
0.00
3.50
8.50
13.50
18.50
23.50
28.50
33.50
38.50
43.50
48.50
53.50
58.50
63.50
68.50
73.50
- 3.49
- 8.49
- 13.49
- 18.49
- 23.49
- 28.49
- 33.49
- 38.49
- 43.49
- 48.49
- 53.49
- 58.49
- 63.49
- 68.49
- 73.49
- 78.49
TOTAL
Total Number
of Sites
21
64
42
62
62
70
210
260
174
113
95
60
29
11
11
2
1286
Sites With
Unknown WQ
3
15
7
11
14
14
59
89
51
6
0
0
0
0
0
0
269
Percentage With
Unknown WQ
14.3
23.4
16.7
17.7
22.6
20.0
28.1
34.2
29.3
5.3
0.0
0.0
0.0
0.0
0.0
0.0
20.9
13
-------
TABLE 3-2
PERCENTAGE OF UNKNOWN WASTE QUANTITY SITES
IN THE NPL DATA BASE
(Historical Development)
Percentage With
Update Identification Unknown Waste Quantity
0 14.5
1 20.3
2 30.6
3 42.9
4 17.9
5 or later 33.6
*As of August 4. 1986.
Calculated with respect to the number of sites
reporting sufficient data to make a Judgement
whether the waste quantity was known or unknown.
14
-------
TABLE 3-3
FREQUENCY OF SITE ACTIVITY DESIGNATIONS—SITES WITH UNKNOWN WASTE QUANTITY (THROUGH UPDATE #5)
Percentage of Sites
Number of Sites With Percentage of Sites With That Site
Unknown WQ That With Unknown WQ Activity That Are
Site Activity Are So Designated* That are So Designated* Unknown WQ Sites
Well Field (Ground Water Plume)
Landfill (Municipal)
Landfill (Commercial/Industrial)
Surface Impoundment
Waste Piles
Spill
Containers/Drums
Tanks (Above Ground)
Tanks (Below Ground)
Chemical Processing/Manufacturing
Wood Preserving
Electroplating
Other Manufacturing/Industrial
66
48
42
40
9
22
31
11
16
12
6
9
33
24.5
17.8
15.6
14.9
3.3
8.2
11.5
4.1
5.9
4.5
2.2
3.3
12.3
95.7**
28.1
12.0
10.0
8.2
16.4
9.9
9.9
21.9
11.8
12.0
23.1
23.2
*Because many sites are given multiple activity designations and because activities that were
designated only infrequently have been omitted from this listing, the figures in the first column
need not sum to the total number of sites with unknown waste quantity. Similarly, those in the
second column need not sum to 100 percent.
**As a matter of site-scoring policy, well fields are defined as having an unknown source, and
therefore, by implication, the waste quantity is unknown. The fact that only 95.7 percent,
rather than 100 percent, of the sites designated as well fields are shown as having an unknown
waste quantity reflects two sites scored early in the NPL program at which the source was unknown,
but waste quantity was estimated from ground water plume data.
-------
surface impoundments, and waste piles (which are all land disposal
sites) represent possibly as many as 50 percent of all unknown waste
quantity sites. (See footnote, Table 3-3.) Except for the case of
well fields, the percentage of unknown waste quantity sites among all
of those sites designated by a given activity is reasonably
consistent with the overall percentage of unknown waste quantity
sites (i.e., 20.9 percent).
16
-------
4.0 SENSITIVITY OF HRS SCORE TO WASTE QUANTITY
4.1 Structure of the HRS Scoring System
The HRS score is the normalized square root of the sum of the
squares of the ground water (S_T), surface water (SCTT) , and air
gw ow
(Sa) migration route scores; i.e.,
(s2 +s2 +sV/2
EW
HRS score = —2
1.73
Each migration route score is itself the normalized product of
three scores; one characterizing the likelihood of release (RC), one
for waste characteristics (WC), and one for targets (T); i.e.,
Si = ki(RCi)(WCi)(Ti),
where the subscript i designates a particular route, and k^ is a
normalization factor for each route. Waste quantity enters into the
calculation of (WC) in essentially the same way for all routes. For
either of the water routes, for instance,
(WC) = (toxicity/persistence factor value)
+ (WQ factor value).
Values for the toxicity/persistence factor can be from 0 to 18.
Waste quantity factor values can be from 0 to 8, following the
*&
guidance given in Table 1-1.
Because the toxicity/persistence factor and the waste quantity
factor are additive, the percentage effect on (WC) of a given change
A route for which the waste quantity factor value is 0 is not to
be scored, since presumably there is no hazardous waste present
that is available for migration by that route.
17
-------
in the waste quantity factor value is inversely proportional to the
magnitude of the toxicity/persistence factor value. For example, if
the toxicity/persistence factor value is 4, then a change in the
waste quantity factor value from 2 to 8, results in a 100 percent
increase in (WC), from 6 to 12. If the toxicity/persistence factor
value is at its maximum, 18, then a change from 2 to 8 in the WQ
factor value results in only a 30 percent increase in (WC), from 20
to 26. In the air route case, the range over which the percentage
change in (WC) can vary is even greater because a different scoring
scheme is used for that pathway, but the inverse relationship still
holds.
Since the waste characteristics score enters into the
computation of each route score as a multiplicative factor, a given
percentage change in it will result in the same percentage change in
the route score. If only one route is being scored, the percentage
change in it will be carried forward as the same percentage change in
the HRS score. If more than one route is being scored and the same
percentage change occurs in all of them (not an unlikely circumstance
if the ground water and surface water routes are the only ones being
scored), then that percentage change will also appear in the HRS
score. If the percentage change is different from route to route,
then the percentage change in the HRS score will be an intermediate
18
-------
value, but tending toward the percentage change in the highest
scoring route.
4.2 Statistical Analysis of Sensitivity
In an earlier study (unpublished), MITRE carried out a
statistical analysis of the sensitivity of the HRS score to each of
the scoring factors in the hazard ranking system. The analysis
examined only that set of randomized factor values for single route
sites that would result in an HRS score equal to 28.50. Each factor
value was systematically varied by + 0.5 unit, holding all others
constant, and the effect on the HRS score was determined. The
analysis showed that the HRS score was less sensitive to the WQ
factor than to any other factor when the route was ground water or
surface water. In the air route case, the sensitivity of the HRS
score to the WQ factor was somewhat more important, but still not
large. The numerical findings were that a + 0.5 unit change in waste
quantity factor value would change the HRS score for ground water or
surface water sites by 4.7 and 5.8 percent, respectively, and by 7.7
percent for an air route site. Because of the limitations of that
study, these numerical results are suggestive only, but the study's
conclusion that the HRS score is relatively insensitive to changes in
the WQ factor value is valid.
19
-------
4.3 Analysis of the Data Through Update #5
For the purpose of this study, the most operationally
significant way to examine the sensitivity of HRS score to waste
quantity is to look at the effect of a change in the WQ default value
on the HRS scores of unknown waste quantity sites whose HRS scores
were less than 28.50; specifically, how many would have their scores
increased to 28.50 or greater if a WQ default value other than 1 were
used. (Sites with scores > 28.50 were not looked at. A default
value of 1 is effectively a minimal value and using any other value
could only raise the HRS score. Raising the HRS score would be of
little practical consequence for those sites whose scores were
already equal to or greater than 28.50 since they would already be on
the NPL and eligible for remedial funding.) Figure 4-1 and Table 4-1
show the results of such a manipulation of the data for the sites
through Update #5. The data show that of the 64 unknown waste
quantity sites with HRS scores currently < 28.50, a minimum of 7 and
a maximum of 28 could score > 28.50 if a default value greater than 1
were chosen. An increase of 28 sites corresponds to an increase of
2.9 percent in the number of sites with HRS score > 28.50. The
reader is reminded that because the NPL data base is incomplete for
sites with HRS scores < 25.00, the points on the curve for the higher
factor values are probably lower than they would be if the data base
were more complete. An estimate of the extent to which the data base
20
-------
30
Is *
O. §0
JT Al
D)
-------
ro
TABLE 4-1
EFFECT OF ALTERING WQ DEFAULT VALUE ON HRS SCORES OF SITES
WITH HRS SCORE <28.50 (THROUGH UPDATE #5)
SITE NAME
Littlerield Township Dump
Lenawee Disposal Service, Inc. Lf
Duston Road
Duck 1 sland Landf i 1 1
Blauvelt Wei 1 Field
Wurtsmith Air Force Base
Ka Ikaskia
Cnnob Park Site
Phoenix Nike
Arundel Corp, Rd . Site
Pearl City Landri 1 1
Hoi den Landri 1 I
Arena 2 (Dioxin)
Harvey Residential Wet Is
Sunrise Landfill
Atlanta Wei Is Vi 1 lage
Araonk Village Wells
East nan Residence
Saukvi 1 le Wei 1 Field
Barrels, Inc.
Bachman Road Wells
Green Street, Houlton
Mechanic Street/Houl ton
Rosati Site
Hoopa Veneer
Ottawa/Luminous Proc.
Lansdowne Radiation Site
Estes Landri 1 1
Schaffer Res. Wei Is
Bedford VII lage Wei Is
Skandla Resident' 1 Wei Is
Fenburg Duap
Ski 1 jam Residence
Bristol Ounp
Sanitary Transfer k Landfill Co.
Placerville (Vanadium)
Sawplt Mi 1 1 (Vandium)
Comonvea 1 th Ca s Co .
Port of Monroe Landfill
Wood ford Road Quarry
Ford Road Ouap
Mount Pleasant
Central LF Authority
1 ron River
Rotron, Inc.
Deer Lake
Klngman Alrpt Industrial Area
Hooksett Landf 1 1 1
Henderson Farm/Littleton
Charlevolx Chemical Co.
Foremost -McKesson
Mr. Harold Curran
LOCATION
Oden
Adr i an
Sa lem
Hamilton Township
Rockland County
Oscoda
Ka Ikaska
R i chmond
Jacksonvi 1 le
G 1 en Bu rn 1 e
Pea r 1 C i ty
Hoi den
Imper i a 1
Ha rvey
Way land Township
Atlanta
Town of North Castle
Rayland
Saukvi 1 le
Lans i ng
Oscoda
Hou 1 ton
Hou 1 ton
Rosa t I
Hoopa Va 1 ley
Ottawa
Lansdowne
Phoen i x
Schaffer VI 1 lage
Bed f o rd
Skandla
Find lay
Cleveland
Bristol
Koshkonong
Place rv 1 M e
Sawp 1 t
F ram 1 ogham
Monroe
Elyrla
Clyrla
Mt. Pleasant
Preltung Township
I ron County
Saugert les
Marquette
Klngman
Hooksett
Littleton
Che r 1 evo 1 x
Henry County
Davenport
CURRENT
HRS
SCORE
28,'lfl
28.31
27.28
27.111
27.13
26.29
26. 13
26.05
25.97
25.53
25.115
24.27
211. 10
23.95
23.06
23.02
23.01
23.01
23.01
22.69
22.1)3
21.97
21.97
21.76
20.90
20.76
20.32
18.97
17.12
17.10
17.10
16.50
16.32
16.18
16.02
16.00
16.00
15.l»8
15.16
11.33
11.12
10.93
10.50
9. HO
9.35
9.20
6.45
8.34
8.27
*8.1<4
7.68
7.34
"2"
30.26
30.08
29.38
29.63
29.22
28.31
27.77
28.06
27.96
26.87
26.78
25.78
25.37
25.45
24. 8M
30. 12
214.77
27.95
24.77
23.89
29.32
23.1I(
23. lit
23.61
22.00
22.58
22.33
19.97
25.16
18.1(2
18.U2
18.00
18.01)
17.01)
17.32
16.95
16.95
16.67
15.98
12.01)
11.70
13.37
20.25
10.1)5
9.85
9.70
8.89
8.99
8.72
26.1)5
8.09
7.73
"-:
32.
31.
31 .
32.
31.
30.
29.
30.
29.
28.
28.
27.
26.
26.
26.
31.
26.
29.
26.
25.
31.
21).
21).
25.
23.
2U.
21).
20.
27.
19.
19.
18.
19.
17.
18.
17.
17.
17.
HRS SCORE CALCULATED US 1
IE INDICATED WQ DEFAULT V
1" "!)" "5" "6"
01)
85
I|8
1 1
31
3D
MO
07
96
22
12
30
63
95
61
90
51)
95
54
09
04
29
29
D5
10
1)0
33
96
19
73
73
90
53
89
18
79
79
86
16.78
12.
12.
11).
21.
11.
10.
10.
9.
9.
9.
27.
8.
8.
75
28
03
16
1)0
35
18
34
64
15
76
D9
12
33.
33.
33.
3M.
33.
32.
31.
32.
31.
29.
29.
28.
27.
28.
28.
33.
28.
31.
28.
26.
32.
25.
25.
27.
21).
26.
26.
21.
29.
21.
21.
19.
21.
18.
19.
18.
18.
19.
17.
13.
12.
1U.
22.
12.
10.
10.
9.
10.
9.
29.
8.
8.
82
62
58
60
39
36
03
07
96
56
1)6
82
90
D5
39
67
31
9D
31
28
76
D5
D5
30
20
2<)
31)
96
21
05
05
80
02
71)
05
6D
6U
05
58
45
87
70
08
35
85
67
78
28
59
09
89
51
35.60
35.39
35.67
37.09
35. M8
3M. 38
32.66
31). 07
33.95
30.91
30.80
30.31)
29. 17
29.95
30.16
35.1)1)
30.08
33.91)
30.08
27. D8
3i). 1)8
26.60
26.60
29. 17
25.30
28.07
28.3>(
22.96
31.2M
22.37
22.37
20.70
22.52
19.60
19.91
19.1)9
19.49
20.21)
18.36
14.16
13.M5
15.37
23.01
13.30
11.35
11.16
10.23
10.92
10.02
30.41
9.29
8.89
37.38
37.16
37.78
39.57
37.57
36. MO
3M.30
36.08
35.96
32.25
32.11)
31.85
30. Ml)
31.M5
31.93
37.22
31.85
35.94
31.85
28.67
36.21
27.76
27.76
31.03
26.40
29.91
30.35
23.95
33.27
23.68
23.68
21.60
24.03
20.45
20.78
20.34
20.34
21.43
19.18
14.88
14.04
16.04
23.95
14.25
11.84
11.64
10.67
11.56
10.46
31.73
9.70
9.28
NC
"7"
39. 16
38.93
39.87
42.07
39.65
38.43
35.94
38.08
37.95
33.59
33.48
33.37
31.71
32.94
33.71
38.99
33.62
37.93
33.62
29.86
37.94
28.92
28.92
32.90
27.50
31.75
32.35
24.95
35.30
25.00
25.00
22.50
25.54
21.30
21.64
21.19
21.19
22.63
19.98
15.58
14.62
16.71
24.90
15.20
12.35
12. 13
11.12
12.20
10.90
33.05
10.10
9.66
"8"
MO . 9M
40.70
Ml .97
MM. 56
Ml .7M
40. M5
37.57
40.08
39.95
34.94
34.82
34.89
32.98
34. MM
35.48
40.76
35.39
39.92
35.39
31.06
39.66
30.07
30.07
34.77
28.60
33.60
34.36
25.95
37.34
26.32
26.32
23.40
27.06
22.15
22.51
22.03
32.03
23.82
20.78
16.29
15.21
17.38
25.85
16.15
12.85
12.61
11.56
12.84
11.34
34.37
10.51
10.05
These are cases 1n which the current HRS score (as shown In the data base) Is shown is 0.00 or Is anomalously Ion,
notwithstanding recorded site data from which a higher score can be calculated. Such sites have been excluded In
constructing Figure 4-1.
-------
TABLE 4-1 (Concluded)
SITE NAME
Cl iff/Oow Plant Site
Tallahassee Trash Dump
TAU Laboratory, Inc.
Federal Maine Term
Hastings Radiochemica 1 Site
Freeport
Schofield Barracks Landfill
Gas Works Park
Chem-Central (Romulus)
KDI Petroleum
Da 1 ton Township Dump
Gulf States Chemical Company
LOCAT 1 ON
Marquette
Leon County
Poughkeepsie
R i ve rv i ew
Pear land
Freeport
Schofield Barracks
Seattle
Romulus
Detroit
Da 1 ton
Lloyd
CURRENT
MRS
SCORE
7.30
6.94
6.77
6.70
6.55
5.35
5.35
5.20
3.90
2.37
1.62
*0.00
11
7
7
7
7
6
5
5
5
4
2
3
26
2"
.71
.64
.29
.02
.90
.53
.76
.48
.11
.50
.20
.60
11
a
8
7
7
7
5
6
5
4
2
4
27
MRS SCORE CALCULATED USI
HE INDICATED WQ DEFAULT V
3" "(jit 115" »6"
.09
.33
.81
.37
.24
.81
.17
.75
.38
.62
.80
.93
8.48
9.03
8.33
7.72
7.58
6.09
6.58
6.02
4.66
2.75
6.40
29.26
8.86
9.72
8.86
8.07
7.93
6.36
6.99
6.30
4.94
2.87
7.99
30.59
9.24
10.42
9.38
8.42
8.28
6.64
7.40
6.57
5.22
3.00
9.60
31.92
NG
tlyll
9.63
11.11
9.90
8.77
8.62
6.92
7.82
6.84
5.49
3.12
11.19
33.25
"8"
10.02
11.80
10.42
9.12
8.97
7.19
8.23
7.12
5.77
3.25
12.79
34.58
U>
These are cases in which the current HRS score (as shown in the data base) 1s shown as 0.00 or is anomalously low,
notwithstanding recorded site data from which a higher score can be calculated. Such sites have been excluded in
constructing Figure 4-1.
-------
may be deficient in sites with scores < 25.00 was made. The
results suggest that as many as 47 (rather than 28) sites could have
their scores increased to > 28.50 if a default value of 8 were used.
As noted in Section 2.0, the fact that not all of the data used in
this analysis has been subjected to quality assurance makes these
specific figures additionally uncertain. However, the conclusion
that the number of sites whose scores would be raised to > 28.50 is
small as compared to the number whose scores already exceed that
value is probably not unreasonable.
This was done by comparing the shape of the distribution of HRS
scores for the sites through Update #5 (which is deficient in sites
with HRS scores < 25.00) with the shape of the same distribution
for the set of HRS scores from which the original NPL was derived
(assumed not to have been as deficient in sites with HRS scores
< 25.00). The comparison gave an indication of the extent to which
the Update #5 distribution might be deficient in sites with HRS
scores < 25.00.
24
-------
5.0 DISCUSSION OF ALTERNATIVES
5.1 Continuing Present Use of a Default Value of 1
As noted in connection with Figure 3-2, the present policy
(i.e., use of a WQ default value of 1 for unknown waste quantity
sites) has resulted in a distribution of the HRS scores of unknown
waste quantity sites that is reasonably consistent with the
distribution of HRS scores for all sites. However, this may be more
a consequence of the relative insensitivity of the HRS score to waste
quantity than an indication that the present practice for scoring
unknown waste quantity sites leads to correct HRS scores for such
sites.
Figure 5-1 shows the distribution of WQ factor values for all of
the sites in the HRS data base through Update #5 for which the waste
quantity was known. If the sites with unknown waste quantity are, as
a set, not very different from those for which waste quantity is
known, it is clear that use of a default value of 1 results, in the
vast majority of cases, in a lower HRS score than would be correct.
Figure 5-2 does, in fact, show that the distributions of site
activity designations for the two categories of sites are reasonably
similar except that the unknown waste quantity sites are notably
higher in the well field designation. This similarity between the
two categories of sites would tend to support a hypothesis that the
present policy provides a bias against the listing of sites with
25
-------
500
400
300
Median-7
Mode -8
to
CO
•5
-------
40
30
CO
•5
Q.
20 •
ro
10 -
_
-
-
_
1
75
IT
1
I
M«
I
__
Q Sites With Known Waste Quantity
^ Sites With Unknown Waste Quantity
wmmmm,
i
^ 0)
^^ C c "P 51
i -E i 1 s I s
c 3 co "^Q. S
\
1 1
'5.
CO
1
wmmfm
CO
§
ffl " CO
c i -£
Q 2 TO
00 P
^n
1 1
1 f
E S
0 cg-5
•^B
^ 1 1
I
w
w\
^ c g .1
c = P '•§ =
)
i w "§ -i Q. a «
CD > .£j CO «J C. o JS'-
S -S-oER^'oX *= m^«
O c q>°C OCO O <1>C3
i. £ s. § £ 2 i Q: m 62 =
FIGURE 5-2
DISTRIBUTION OF SITES BY SELECTED SITE ACTIVITY
(THROUGH UPDATE #5)
-------
unknown waste quantity. However, the bias is not as strong as would
have been the case had EPA policy been to give waste quantity, if
unknown, a factor value of zero, as is done in the case of all other
factors for which the required information is lacking.
5.2 Using a Different (But Fixed) Default Value
An obvious alternative to using a default value of 1 is to use a
value that has significance in terms of the distribution of WQ factor
values for all sites for which waste quantity is known. Referring to
Figure 5-1, one sees that the median and mode of the distribution are
•ff
1 and 8, respectively. From Figure 4-1, one sees that using a
default value of 7 could result in 27 more sites scoring > 28.50
while a value of 8 could result in 28 reaching or exceeding that
score. Thus, there appears to be little to choose between them.
However, with reference to Figure 5-1 once again, one sees that the
choice of 8 as the default value would maximize the number of unknown
waste quantity sites scored correctly for waste quantity, but all
other unknown waste quantity sites would have their waste quantities
over-estimated. On the other hand, the choice of 7 as the default
value would provide for fewer correctly scored sites, but would, by
virtue of its being the median value, result in nearly equal numbers
of over- and under-estimates.
Since the WQ factor value is an ordinal number, the mean of the
distribution is of little significance.
28
-------
It is clear that using any single default value results in a
significant number of under-estimates and/or over-estimates of waste
quantity. Furthermore, the use of any of the larger possible default
values could discourage attempts to collect information about waste
quantity. In the extreme case, at sites where the waste quantity was
known to be less than that normally required to earn the default
value, it might encourage waste quantity to be reported as unknown,
if the scorer's bias was toward listing the site.
5.3 Using a Site-Activity-Dependent Default Value
A useful modification of the single default value approach
discussed in the previous section would be to select a WQ default
value that was appropriate for the type of site concerned. Such
default values could be obtained from the UQ factor value
distributions for sites of various activity types whose waste
quantities were known. (A default value for well fields would have
to be selected on a different basis since it follows from the
definition of well field (i.e., having an unknown source) that no
empirical data on the distribution of waste quantity for such sites
should be available.)
This is, in principle, an attractive approach since it is simple
and would be based on empirical data for the various types of sites.
Figures 5-3, 5-4, and 5-5 show distributions of waste quantity factor
values for known waste quantity sites with the same site activities
29
-------
30
20
10
Landfill (Municipal)
Landfill (Commercial/Industrial)
30
20
10
0
Waste Piles
Ground Water Route
345
Waste Quantity Factor Value
Surface Water Route
Air Route
FIGURE 5-3
DISTRIBUTIONS OF WASTE QUANTITY FACTOR VALUES FOR SITES
WITH KNOWN WASTE QUANTITY AND DESIGNATED SOLELY AS LANDFILL
(MUNICIPAL), LANDFILL (COMMERCIAL/INDUSTRIAL), SURFACE IMPOUNDMENT,
OR WASTE PILES - (THROUGH UPDATE #5)
30
-------
30
20
10
Spill
30
20
10
55 0
5
30
20
10
Containers And Drums
Tanks (Above Ground)
30
20
10
Tanks (Below Ground)
34
Waste Factor Value
Ground Water Route
Surface Water Route
Air Route
FIGURE 5-4
DISTRIBUTIONS OF WASTE QUANTITY FACTOR VALUES FOR SITES
WITH KNOWN WASTE QUANTITY AND DESIGNATED SOLELY AS SPILL,
CONTAINERS AND DRUMS, TANKS (ABOVE GROUND) OR TANKS
(BELOW GROUND) - (THROUGH UPDATE #5)
31
-------
30
20
10
Chemical Processing/Manufacturing
30
So 20
5
10
3
Wood Preserving
30
20
10
Other Manufacturing/Industrial
345
Waste Quantity Factor
Ground Water Route [J Surface Water Route
Air Route
FIGURE 5-5
DISTRIBUTIONS OF WASTE QUANTITY FACTOR VALUES FOR SITES
WITH KNOWN WASTE QUANTITY AND DESIGNATED SOLELY AS CHEMICAL
PROCESSING/MANUFACTURING, WOOD PRESERVING, OR OTHER
MANUFACTURING/INDUSTRIAL — (THROUGH UPDATE #5)
32
-------
as appear most frequently with the unknown waste quantity sites. For
each distribution, sites designated with only that site activity were
included; sites with multiple site activities were not used. The
median and mode for each distribution, along with the number of sites
in each distribution, are given in Table 5-1. The following features
are characteristic of the data in the table:
• The distributions for the ground water and surface water
routes are based on substantially more sites than are the
air route distributions.
• For each type of site, the data for the two water routes
are nearly identical; the data for the corresponding air
route may be quite different.
• In most cases, there is little difference between the
median and the mode for a particular route.
• As compared to the other types of sites, the land disposal
sites have higher medians and modes, which are, in most
cases, the maximum value, 8.
For the purpose of constructing a table from which a waste
quantity default value could be selected for a site, several other
considerations are relevant:
• In Section 5.2, it was noted that either the median or the
mode of a distribution could be used as a default value.
Use of the median would result in nearly equal numbers of
under- and over-estimates of waste quantity. Use of the
mode, on the other hand, would assign the correct value to
a larger number of sites, but could lead to a net over- or
under-estimating bias for all of the others. Using the
median would appear to be preferred in order to minimize
the likelihood of such bias.
• The relative lack of waste quantity factor values for the
air route, particularly for the non-land disposal sites,
decreases one's confidence that the apparent differences
between the air and water routes values for those sites are
33
-------
TABLE 5-1
PARAMETERS OF WASTE QUANTITY FACTOR DISTRIBUTIONS
FOR VARIOUS TYPES OF SITES (THROUGH UPDATE #5)
Land Disposal Sites
Landfill (Municipal)
Landfill (Commercial/
Industrial)
Surface Impoundment
Waste Piles
Non-land Disposal Sites
Spill
Containers and Drums
Tanks (Above Ground)
Tanks (Below Ground)
Chemical Processing/
Manufacturing
Wood Preserving
Other Manufacturing/
Ground
Number
of Sites
14
48
39
16
5
38
1
3
6
3
4
Water Route
Median
5
8
8
8
5
3
1
2
6
4
2
Mode
8
8
8
8
5
2
1
2
8
3
1
Surface
Number
of Sites
15
41
34
16
5
37
1
3
6
3
3
Water Route
Median
5
8
8
8
5
3
1
2
6
4
3
Mode
8
8
8
8
5
2
1
2
8
1
1
Air Route
Number
of Sites Median
3 8
11 8
5 8
3 8
0
5 4
0
0
1 5
0
1 1
Mode
8
8
8
8
—
4
—
—
5
—
1
Industrial
-------
significant. In light of this, plus the near equality of
the air and water routes values for the land disposal sites
(for which the statistics are significantly better), it
would appear that only in the case of the municipal
landfills is it useful to distinguish between the air route
and the water routes.
• Except for the site activity "Containers and drums," the
non-land disposal sites represented in Table 5-1 are few in
number. Combining them into appropriate groups and
calculating group medians would both improve their
statistical base and eliminate apparent differences among
them that may not be real. Two groupings could be:
containers and drums, tanks (above ground), and tanks
(below ground); and manufacturing or industrial sites.
Sites characterized as spills fit into neither of these
groups, and could stand alone, although any suggested
default value for spills would be based on little data.
• With regard to the land disposal sites, the differences
between the municipal landfills and the others appear to be
significant, justifying the retention of municipal
landfills as a separate category. The others have
identical values and could be combined.
Taking all of the above into account, Table 5-2, which gives a
set of default values by type of site, has been constructed. The
precise degree of discrimination observed may be spurious, arising
from the limited number of sites in the data base with only one site
activity designation. However, a difference between the land
disposal sites and all others is probably real. In any event, to the
extent that any real degree of discrimination may be provided by the
multiple default value approach, it is preferable to the single
default value approaches.
35
-------
TABLE 5-2
WASTE QUANTITY DEFAULT VALUES, BY TYPE OF SITE
Type of Site (Principal Designation) Routes Default Value
Landfill (Municipal) Ground Water 5
Surface Water 5
Air 8
Landfill (Commercial/Industrial)
Surface Impoundment f All
Waste Piles
Spills All
1
Containers and Drums
Tanks (Above Ground) J. All
Tanks (Below Ground)
Chemical Processing/Manufacturing
Wood Preserving
Electroplating ' Al1
Other Manufacturing/Industrial
36
-------
Since there are, as already mentioned, no comparable empirical
data for well fields, a different approach would be necessary for
them. One possibility would be to use the overall median for all
types of sites, on the basis that a well field could involve any type
of source. On this basis, the default value for well fields would
be 7. Alternatively, an overall, or weighted, median for the most
likely types of sources might be used (in a given case) if there were
sufficient basis for making the necessary judgements regarding the
types of sites to be included, and their relative importance.
5.4 Getting Sampling Data from Which Waste Quantity Could be
Estimated
EPA could adopt a policy of having field sampling done at
unknown waste quantity sites in order to estimate waste quantity.
This policy does not have much to commend it. First of all, at well
fields, which constitute a significant fraction of unknown waste
quantity sites, the source location is unknown. Second, the few
•!<
RIs that are available for unknown waste quantity sites indicate
that where the source location is known, it is extensive (e.g., a
trench 150' long x 40' wide x 10' deep, or a "5.2 acre site used for
dumping"). Thus, the number of samples needed for a useful estimate
of waste quantity could be large, leading to considerable expense.
RIs for Bog Creek Farm, Charlevoix Municipal Well, Le Hillier/
Mankato Site, Cecil Lindsey Site, and South Valley are the only
ones available, and were examined.
37
-------
In addition, obtaining the samples could be hazardous; and
definitional problems (e.g., what materials to analyze for?) could
limit the utility of the data. Cost/time/effectiveness
considerations would appear to make this approach questionable.
5.5 Dropping Waste Quantity as a Scoring Factor
Since the HRS score has a low sensitivity to the WQ factor
value, one could consider just dropping it. The consequences of
doing so are discussed below.
It is difficult to assess quantitatively the impact of dropping
waste quantity as a scoring factor; one would need to rescore all of
the original sites (omitting waste quantity from the scoring scheme),
establish a new cut-off score, recompute HRS scores for all of the
sites in the updates (once again, omitting waste quantity as a
scoring factor), and see how many sites would then equal or exceed
the new cut-off and to what extent they constituted a different set
than is now the case.
There are conceptual grounds, however, for retaining waste
quantity as a scoring factor. While application of the HRS does not
result in a quantitative risk assessment, it is intended to reflect
the relative threat posed by sites. Thus, each of the scoring
factors should be related to either, or both, the probability of an
adverse event (i.e., a release) and/or the seriousness of the event.
Waste quantity has no bearing on the probability of a release, but it
38
-------
can influence the seriousness of the release by its influence on both
the expected daily dose to a target population (or sensitive
environment) and the length of time exposure can be expected. Waste
quantity is not alone in this regard; net precipitation and
permeability of the unsaturated zone, for instance, also affect the
daily dose and the duration of exposure. However, in each case,
their effects on dose and duration are offsetting. A high factor
value for either implies a high rate of migration away from the site
and, accordingly, a high dose; however, there would also be a
correspondingly rapid depletion of the amount of hazardous substance
at the site and hence a reduction in the expected duration of
exposure. Only waste quantity acts in the same direction on both.
If the waste quantity factor value is raised, one expects both the
magnitude of the daily dose and the duration of exposure to
increase. Thus, while WQ is. not a factor of major importance in
terms of the sensitivity of the HRS score to a change its value, it
is the only factor whose value scales positively with both expected
daily dose and duration of exposure. This is its uniqueness, and,
without waste quantity, the scoring system could be regarded as
deficient.
There is yet another reason for retaining waste quantity as a
scoring factor. The scoring for toxicity in the HRS acts as a filter
to screen out sites that do not contain toxic substances. However,
39
-------
because a site can be scored on the basis of the most toxic substance
present, the toxicity/persistence factor value tends to be high for
many sites. Factoring in waste quantity potentially provides a
useful degree of discrimination among sites with similar toxicity/
persistence scores. This could be particularly valuable when chronic
toxicity is of concern.
40
-------
6.0 SURROGATES FOR WASTE QUANTITY
It follows from the previous section that a surrogate for waste
quantity should score positively for both magnitude of daily dose and
duration of exposure. It has not been possible to identify a waste
characteristic other than quantity that meets this criterion. This
section, therefore, only briefly considers several alternative
scoring schemes that could be used when the required information for
the present scoring scheme is unavailable.
The present scoring scheme requires an estimate of hazardous
waste quantity (the amount of waste contaminated with hazardous
substance) in terms of tons, cubic yards, or drums. In most cases,
this necessitates the availability of records or manifests for
deliveries made to the site or for wastes generated at the site.
Estimates based on visual observation of the site may be possible in
such cases as drum dumps, waste piles and surface impoundments, if
the materials are exposed. If such records are not available or such
visible estimates are not possible, it may be possible to score the
site on the basis of other quantities for which records or visual
estimates may be more easily obtained. Of the alternatives, the most
readily apparent is total waste quantity at the site. This quantity
could be estimated from records that did not necessarily address the
character of the waste delivered to, or generated at, the site. If
records were not available, visual inspection of the site might
-------
enable an estimate of the total amount of waste at the site. In
setting up a scoring scheme one would want to take into account,
where known, the general character of the waste (i.e., municipal,
industrial, mixed, etc.) in order to try to reflect the proportion of
the total that could be contaminated.
A second alternative is suggested by the RIs that have been
examined. They indicate that unknown waste quantity sites are likely
to be dumps at which the wastes have been mixed with soil and covered
over, and disposal records do not exist. In such cases, the
topography of the dump and the appearance of the ground could enable
an estimate of the area in which the waste is contained, and a
scoring scheme based on area could apply. The scoring scheme could
include a factor for the origins of the waste delivered to the dump
(municipal, industrial, or mixed) and a factor reflecting the
proportion of soil in the waste-containing volume.
In these alternatives, by introducing factors such as the origin
of the waste and the fraction of soil with which the waste is mixed,
one is implicitly trying to move toward an estimate of hazardous
waste quantity, albeit a highly uncertain one. Such factors need not
be introduced and one could score solely on the basis of estimated
total waste quantity, or area.
42
-------
Using data from RIs for sites with known waste quantity, one
might be able to establish empirical correlations among hazardous
waste quantity, total waste quantity, and area. If it were possible
to do so, then scoring tables could be established that would, in
principle, yield the same factor value for a site, regardless of
whether scoring was done on the basis of hazardous waste quantity,
total waste quantity, or area. It is questionable whether sufficient
data are available from the RIs to be able to generate such a scoring
scale for waste quantity, but data on the extent of the disposal area
ate usually provided in the site description sections. Because of the
imprecise character of this as well as the other possible surrogate
schemes, it would hardly seem justifiable to retain an 8-step scoring
scale. A 3- or 4-step scale should suffice.
Another alternative would be to simply use a gross character-
ization of the size of the site (small, medium or large), and the
likely origins of the wastes, assigning factor values drawn from the
waste quantity factor value distribution shown in Figure 5-1, or from
those discussed in Section 5-3.
A final alternative takes a fundamentally different approach. No
physical data from the site are necessary. One would use a scoring
scheme based on whatever information was available on the number of
years the site was used for dumping, the number of dumpers, the
frequency of dumping, and the nature of the dumpers. However,
A3
-------
developing and using this type of scheme requires data that could be
difficult to obtain, particularly for sites at which the waste
quantity is unknown.
Once again, well fields would require a separate approach since
the source is unknown, and the necessary information to apply any of
the approaches noted above would not be available. Perhaps the best
approach for well fields would be the use of a default value, as
discussed in Section 5.3.
It seems clear that surrogate schemes can be quite elaborate or
quite simple. It is not at all evident that the more elaborate ones
are more likely to be worth the additional effort to develop them. In
view of the need for a surrogate only in the 20-30 percent of sites
for which hazardous waste quantity cannot be credibly estimated and
the relative insensitivity of the HRS score to waste quantity, it
would appear that the simpler approaches are to be preferred.
-------
7.0 DISCUSSION AND RECOMMENDATIONS
It is clear that the present policy of assigning a default value
of 1 to all unknown waste quantity sites results in an underestimate
of waste quantity for the vast majority of them. Fortunately, the
number of sites possibly being kept off the NPL because of this is
small as compared to the number of sites already on the NPL.
However, it is possible that the fraction of unknown waste quantity
sites among those being considered in future updates will be somewhat
higher than it has been.
The easiest approach to a solution, simply selecting the median
or mode of the distribution of WQ factor values as a new default
value, would increase the number of correct WQ factor value
assignments and/or improve the balance between over- and under-
estimates. But, since the median and mode for this highly skewed
distribution are large, 7 and 8 respectively, such a choice has the
disadvantage that, where the scorer's bias is toward listing a site,
it could discourage attempts to collect information about waste
quantity. In the extreme case, it could encourage the reporting of
sites as having an unknown waste quantity in instances where the
waste quantity was known, but less than ordinarily required to earn
that higher factor value. On the other hand, the "threat" of a high
default value could, in some instances, result in the disclosure by
potentially responsible parties of waste quantity data that might not
otherwise come to light.
45
-------
Use of the multiple default value approach would appear to be an
improvement over the single default value approach, since the
available data show some discrimination among the different types of
sites. However, the degree of discrimination suggested by Tables 5-1
and 5-2 must be regarded as uncertain, since the number of sites on
which they are based is not large. The multiple default value
approach makes the designation of the type of site important, and
such designations can be subjective and imprecise, although using the
broad site activity categories, as in Table 5-2, should reduce such
concern. In addition, provision would have to be made for dealing
with sites that legitimately require designation with more than one
site activity, or do not fit into even the broad categories of
Table 5-2. The multiple default value approach also shares, with the
first approach, the disadvantage of possibly offering a disincentive
for data collection in certain cases.
A final disadvantage of the multiple default value approach is
that once a judgement is made as to the type of site concerned, there
is no further opportunity for the exercise of judgement as to whether
the median waste quantity factor value obtained from a large number
of sites of that type is applicable to that particular site, given
fairly obvious features of the site. For instance, should large and
small industrial/manufacturing sites be given the same waste quantity
factor value just because their actual waste quantitites are
unknown? If the multiple default value approach is selected, a body
46
-------
of guidance might be developed that would allow for increasing or
decreasing (by one unit, perhaps) the suggested default value for a
site, depending on whether it was a particularly large one or small
one of its type. Doing this for the land disposal sites, for
instance, might mean incorporating into the guidance definitions of
large, small and average, in terms of the area in which disposal
occurred. Analogous guidance inputs would have to be developed for
the other categories of sites as well.
Either of the default value approaches, single or multiple,
would result in a reduction of the number of incorrect WQ factor
value assignments to unknown waste quantity sites. In addition, the
over- and under-estimates would, in most cases, be more even in
number. It remains for EPA to decide whether these advantages
outweigh the disadvantages of possibly introducing an increased
amount of scorer bias for those sites identified as having an unknown
waste quantity.
The alternative of dropping waste quantity as a scoring factor
is not to be favored. First, it would degrade the risk-relatedness
of the HRS scoring system by giving up the only scoring factor that
scales positively for both expected daily dose and duration of
exposure. Second, one would forego a factor which helps discriminate
among the large number of sites that score high on toxicity/
persistence. These are compelling reasons not to consider this
alternative further.
47
-------
Among the alternatives considered in the discussion of
surrogates, one stands out as offering the most promise--scoring on
the basis of the size of the area in which dumping actually
occurred. It has the following positive features:
• in any instances in which the source location is known, it
should be possible to make a reasonable estimate of the
size of the dumping area;
• this means that it should be possible to generate a data
base from which a scale for scoring the dumping area could
be developed (i.e., from the remedial investigations);
• the assumption that the larger the dumping area, the
greater the waste quantity is intuitively reasonable
(although possible exceptions can easily be conceived);
• the assignment of a factor value would be site-specific,
taking into account whatever is known about the site (i.e.,
topography, history, overall appearance, etc.); and
• this approach avoids the problem, inherent in the default
value approaches, of treating all unknown waste quantity
sites, or even all sites of a given type, the same.
Its principal disadvantages would appear to be that it would involve
the exercise of considerable judgement by the site scorer, and it
might not be possible to find enough data to generate an acceptable
scoring scale. In addition, it is only applicable to the land
disposal sites.
It is recommended that, if the present WQ default policy is to
be replaced, EPA pursue the multiple default value approach as most
favored. It would provide a degree of discrimination among unknown
waste quantity sites that is unachievable if a single default value
48
-------
were to be used. The multiple default value approach would, with
appropriate guidance as discussed earlier in this section, be
relatively easy to implement. It has the desirable feature of being
based on empirical data.
As a final note, the highly skewed distribution of waste
quantity factor values shown in Figure 5-1 has implications beyond
the question of how to deal with unknown waste quantity sites. It
shows that the present scale for assigning WQ factor values provides
no discrimination among the largest 50 percent (approximately) of the
sites, notwithstanding a range of waste quantities of four or more
orders of magnitude among them. This is evident from Figure 7-1
which shows the distributions of actual waste quantity for the known
waste quantity sites in this analysis. Note that any waste quantity
in excess of 2500 cubic yards warrants a factor value of 8.
In setting the threshold for the maximum WQ factor value as low
as 2500 cubic yards (or equivalent), the waste quantity factor
discriminates primarily among the smaller sites, and treats all of
the larger sites as essentially equivalent. This reflects a policy
decision, made by EPA when the HRS was promulgated, to help screen
out small volume sites that might be addressed by state or local
governments. A consequence of structuring the waste quantity factor
scale in this way, however, is that it tends to compromise the
ability of the HRS to deal with chronic toxicants such as
49
-------
300
250
200
V)
•5
o
J3
150
100
50
2 23 34 4556 67 7
<10 10-10 10-10 10-10 10-10 10-10 10-10 >10
Waste Quantity (Cubic Yards)
FIGURE 7-1
DISTRIBUTION OF WASTE QUANTITY FOR SITES THROUGH UPDATE #5
(SITES WITH UNKNOWN WASTE QUANTITY EXCLUDED)
50
-------
carcinogens, mutagens and teratogens. As noted earlier, waste
quantity is the only scoring factor that scales positively with
duration of exposure, which is a major determinant of risk for
chronic toxicants. The present waste quantity scoring scheme
compresses a possible range of four or more orders of magnitude
difference in duration of exposure (assuming rough proportionality
between waste quantity and duration of exposure) into a single WQ
factor value. Such limited discrimination with respect to a factor
that is indicative of duration of exposure (as well as magnitude of
the daily dose) would not be consistent with possible modification of
the HRS to give more importance to chronic toxicity. If such
modification were to be undertaken, EPA should consider the
advantages and disadvantages of changing Table 1-1 to provide more
discrimination among the larger sites.
51
------- |