Hazard Ranking System (HRS) for Uncontrolled Hazardous Substance Releases Appendix A of the National Oil and Hazardous Substances Contingency Plan

6241
ENVIRONMENTAL PROTECTION AGENCY
" „ '•
40 CFR Part 300

^azard Ranking System (HRS) for Uncontrolled Hazardous Substance
Releases; Appendix A of the National Oil and Hazardous Substances
^ Contingency Plan
G—
$ AGENCY: Environmental Protection Agency
-------
-2-

A. Science Advisory Board
B. Review of Alternative Ranking Systems
C. Cutoff Score
D. SARA Conference Reports
E. Advance Notice of Proposed Rulemaking

IV. Background Documents
V. Discussion of the Proposed Rule

A. Overview
B. Major Revisions
C. Revisions Affecting Multiple Pathways

1. Structure
2. Toxicity
3. Hazardous Waste Quantity
4. Observed Release/Concentration of Hazardous
Substances in the Environment
5. Use of Health-Based Benchmarks in Evaluating
Target Populations
6. Consideration of Hazards to the Environment
7. Consideration of Effects on the Human Food Chain
8. Dilution/Distance Weighting of Targets
9. Population Risks and Risk to the Maximally
Exposed Individual
10. Scoring on the Basis of Current Conditions
11. Low Density Populations
12. Standby Wells and Surface Water Intakes

D. Individual Pathway Revisions

1. Ground Water Pathway
2. Surface Water Pathway
3. Air Pathway
4. Onsite Exposure Pathway
5. Fire and Explosion

E. CERCLA Section 125

VI. Required Analyses

A. Executive Order No. 12291
B. Regulatory Flexibility Act
C. Paperwork Reduction Act

I. BACKGROUND

In 1980, Congress enacted the Comprehensive Environmental
Response, Compensation, and Liability Act (CERCLA) (42 U.S.C.
9601 et seq.), commonly called the Superfund, in response to the
dangers posed by uncontrolled releases of hazardous substances,
pollutants, or contaminants into the environment.1 To implement
Section 105(8)(A) of CERCLA and Executive Order 12316 (46 FR
42237, August 20, 1981), the Environmental Protection Agency

1 For the purpose of this rule and preamble, the term
"hazardous substances, pollutants, or contaminants" will be
referred to simply as "hazardous substances".
-------
-3-

(EPA) revised the National Oil and Hazardous Substances Pollution
Contingency Plan (NCP), 40 CFR Part 300, on July 16, 1982 (47 FR
31180), with later revisions on September 16, 1985 (50 FR 37624),
and November 20, 1985 (50 FR 47912). The NCP sets forth guidelines
and procedures for responding to releases or potential releases
of hazardous substances, pollutants, or contaminants. Section
105(8)(A) of CERCLA (now Section 105(a)(8)(A)) required EPA to
establish:

criteria for determining priorities among releases or
threatened releases [of hazardous substances] throughout
the United States for the purpose of taking remedial
action and, to the extent practicable, taking into
account the potential urgency of such action, for the
purpose of taking removal action. Criteria and priorities
... shall be based upon relative risk or danger to public
health or welfare or the environment .... taking into
account to the extent possible the population at risk,
the hazard potential of hazardous substances at such
facilities, the potential for contamination of drinking
water supplies, the potential for direct human contact [and]
the potential for destruction of sensitive ecosystems...

To meet this requirement and help set priorities, EPA adopted
the Hazard Ranking System (HRS) as part of the revised NCP. The
HRS is a scoring system used to assess the relative threat associated
with actual or potential releases of hazardous substances from a
site. An HRS score is determined for a site by evaluating
several migration routes or "pathways", such as water and air.
The score for each pathway is obtained by evaluating a set of
"factors" that characterize the potential of the facility to cause
harm via that pathway. The factors, such as toxicity of the
substances at a site, waste quantity, and population, are each
assigned a numerical value according to instructions set out in
Appendix A to the final NCP (47 31180, July 16, 1982); this value
is multipled by a weighting factor yielding the factor score. The
factor scores are then combined within "factor categories"; the
total scores for the factor categories are multiplied together to
develop a score for the relevant pathway. Finally, the pathway
scores are combined according to a mathematical formula to produce
the HRS score for the site.

The HRS was designed to be applied uniformly to each site,
enabling sites to be evaluated relative to each other with respect
to actual or potential hazards. As EPA explained when it adopted
the HRS, "the HRS is a means for applying uniform technical judgment
regarding the potential hazards presented by a facility relative
to other facilities. It does not address the feasibility,
desirability, or degree of cleanup required" (47 FR 31220, July 16,
1982). Although the HRS was designed to assess relative risks,
it is not designed to be used as a quantitative risk assessment.

The HRS score is a crucial part of the Agency's program to
address the identification and cleanup of actual and potential
-------
— 4 —

releases of hazardous substances because the HRS is the primary
way of determining whether a site is to be included on the National ,
Priorities List (NPL). Each State can also designate a single
site to the NPL as a State top priority site regardless of its
HRS score and sites may be added in response to a health advisory
from the Agency for Toxic Substances and Disease Registry (see
NCP, 40 CFR Section 300.66(b)(4)). The NPL (Appendix B to 40
CFR Part 300) includes those sites that appear to pose the most
serious threats to public health and the environment and that
appear to warrant remedial investigation and possible cleanup
under CERCLA. Only sites on the NPL are eligible for Superfund-
financed remedial actions. Emergency removal and enforcement
actions can be conducted at any site whether or not it is on
the NPL.

Remedial action tends to be long-term in nature and involves
response actions that are consistent with a permanent remedy for
a release. Removal actions tend to be short-term or temporary in
nature and involve cleanup or other actions deemed necessary to
prevent or minimize damage to public health and the environment.

In 1986, Congress passed the Superfund Amendments and
Reauthorization Act (SARA) (P.L. 99-499), which added a Section
105(c)(1) to CERCLA requiring EPA to amend the HRS to assure "to
the maximum extent feasible, that the hazard ranking system
accurately assesses the relative degree of risk to human
health and the environment posed by sites and facilities
subject to review." Section 105(c)(2) as amended also requires
that the HRS appropriately assess the human health risks
associated with contamination or potential contamination of
surface waters, either directly or as a result of runoff of
any hazardous substance. This assessment should take into
account the use of these waters for recreation and the potential
migration of any hazardous substance through surface water to
downstream sources of drinking water.

The Amendments also add two criteria for evaluating sites
under Section 105(a)(8)(A):

o Evaluation of the damage to natural resources
which may affect the human food chain and which is
associated with any release or threatened release of
hazardous substances.

o The contamination or potential contamination of the
ambient air which is associated with a release or
threatened release.

Section 105(c)(1) states that the revised HRS shall be
applied to any site to be newly listed on the NPL after the
effective date of the revised HRS. Until the effective date,
sites will be scored with the current HRS. In addition,
Section 105(c)(3) specifies that EPA shall not be required to
-------
-5-

rescore any site evaluated with the current HRS before the
effective date.

CERCLA Section 118, added by SARA, requires EPA to give
a high priority to facilities where the release of hazardous
substances has resulted in the closing of drinking water
wells or has contaminated a principal drinking water supply.

CERCLA Section 125 (added by SARA) requires revisions to the
HRS to address facilities that contain substantial volumes of
wastes specified in Section 3001(b)(3)(A)(i) of the Solid
Waste Disposal Act (Resource Conservation and Recovery Act or
RCRA). These wastes include fly ash wastes, bottom ash wastes,
slag wastes, and flue gas emission control wastes generated
primarily from the combustion of coal or other fossil fuels.
Section 125 requires EPA to revise the HRS to assure the
appropriate consideration of each of the following site-specific
characteristics of such facilities:

1. The quantity, toxicity, and concentrations of
hazardous constituents which are present in such
waste and a comparison thereof with other wastes.

2. The extent of, and potential for, release of such
hazardous constituents into the environment.

3. The degree of risk to human health and the
environment posed by such constituents.

II. HAZARD RANKING SYSTEM

The current HRS serves as a screening device to evaluate
the relative potential of uncontrolled hazardous substances
to cause human health or safety problems or ecological or
environmental damage. The pre-remedial portion of the
Superfund program —- the portion prior to placing sites on the
NPL — is intended to identify those sites that represent the
highest priority for further investigation and possible cleanup
under CERCLA. During site discovery, the first step of the pre-
remedial process, possible releases of hazardous substances are
listed in the CERCLA Information System (CERCLIS). A preliminary
assessment is then conducted for all sites on CERCLIS; this low
cost, initial evaluation is meant to give as full and complete a
picture of the site as possible using existing information. EPA
is currently continuing to screen the approximately 30,000 sites
presently in CERCLIS.

If, based on the results of the preliminary assessment, EPA
determines that a site warrants further action, the Agency initiates
a site inspection as specified in the NCP (40 CFR Section 300.66).
The site inspection may include the collection of a limited number
of samples for chemical analysis. Such samples aid in ascertaining
what substances are present at the site and whether they are
being released. The purpose of the site inspection is to determine
-------
-6-

if there is a potential threat to public health or the environment,
to determine if there is an immediate threat to people in the
area, and to collect sufficient data to enable the site to be
scored using the HRS.

As required by CERCLA, EPA has designed the Superfund program
to focus its resources on the highest priority sites. Consequently,
the initial studies — the preliminary assessment and site inspection
— which are performed on a large number of sites, are relatively
modest in scope and cost compared to the remedial investigations
and feasibility studies subsequently performed on NPL sites.
Because of the need to carry out the initial studies expeditiously,
EPA has elected to place certain constraints on the complexity of
the HRS. The required HRS data should be information that, for
most sites, can be collected in a single site visit or that are
already available. Thus, the HRS cannot rely on data that require
extensive sampling or repeated sampling over a long period of
time. The HRS has also been designed so that it can be applied
consistently to a wide variety of sites.

The HRS provides a measure of relative rather than absolute
risk. Congress, in its Conference Report on SARA, confirmed the
appropriateness of this approach when it specified a substantive
standard against which HRS revisions could be assessed.

This standard is to be applied within the context of
the purpose for the National Priorities List; i.e.,
identifying for the States and the public those
facilities and sites which appear to warrant remedial
actions....This standard does not, however, require the
Hazard Ranking System to be equivalent to detailed risk
assessments, quantitative or qualitative, such as might
be performed as part of remedial actions. The standard
requires the Hazard Ranking System to rank sites as
accurately as the Agency believes is feasible using
information from preliminary assessments and site
inspections....Meeting this standard does not require
long-term monitoring or an accurate determination of the
full nature and extent of contamination at sites or the
projected levels of exposure such as might be done during
remedial investigations and feasibility studies. This
provision is intended to ensure that the Hazard Ranking
System performs with a degree of accuracy appropriate to
its role in expeditiously identifying candidates for
response actions (H.R. Rep. No. 962, 99th Cong., 2nd
Sess. at 199-200 (1986)) (emphasis added).

EPA wants to emphasize that the HRS was designed to assess
relative risk, and, thus, is not designed to be used as a
quantitative risk assessment.

Of the approximately 30,000 sites in CERCLIS, approximately
27,000 have received a preliminary assessment. About 9,000 of
those have had a site inspection and about 2,000 have been scored
using the HRS. To date, there are 1,175 sites on or proposed
-------
for the NPL (see 48 FR 40658, September 8, 1983; 49 FR 19480,
May 8, 1984; 49 FR 37070, September 21, 1984; 50 FR 6320,
February 14, 1985; 50 FR 37630, September 16, 1985; 51 FR
21054, June 10, 1986; 52 FR 2492, January 22, 1987; 52 FR
27620, July 22, 1987; 53 FR 23988, June 24, 1988 and 53 FR
33811, September 1, 1988). SARA provides EPA (see CERCLA Sections
116(a)(l) and (2), as amended) with goals for completing preliminary
assessments and site inspections. For all sites in CERCLIS as of
the date of the enactment of SARA, preliminary assessments should
be completed by January 1, 1988, and site inspections, where
needed, should be completed by January 1, 1989.

Although the NPL is ordered by HRS scores, EPA puts the
sites into groups to emphasize that minor differences in
scores do not necessarily indicate significantly different
levels of risk.

The revisions being proposed today reflect the Agency's
efforts to improve the accuracy of the HRS, addressing the
SARA mandate. While this proposed rule, Appendix A to 40 CFR
300 ("Proposed Rule") would add some new elements, the HRS
would still serve its intended purpose as a screening tool.
EPA has been careful to balance the potential increased costs
in time and resources to collect more data against the goal
of improving accuracy, so the revised HRS can list sites
correctly without impairing the Agency's ability to evaluate
releases quickly.

Current HRS

The current HRS evaluates the relative threat of a site
over five pathways — ground water, surface water, air,
direct contact, and fire and explosion. The first three of
these pathways reflect the risk from migration of hazardous
substances from the site. The scores for ground water, surface
water, and air are combined into an overall migration score
that is the primary consideration in placing a site on the
NPL. The last two pathways, direct contact and fire and
explosion, may be used to determine if the potential risk is so
acute that emergency action is required, but are not included in
the overall HRS migration score.

The current HRS uses a structured value analysis approach to
scoring sites. This approach assigns values to factors related
to or indicative of risk. The basic elements of the current HRS
are factors such as toxicity and containment. A scale of numerical
rating values is provided for each factor and a value is assigned
to each factor based on conditions at the site. Individual values
are then weighted. The factors are grouped into three factor
categories — observed release/route characteristics, waste
characteristics, and targets — and are combined to obtain
factor category scores. Each factor category has a maximum
value, as does each of the component factors within the category.

The relevant factor category scores are multiplied together
-------
-8-

within each pathway and normalized to obtain a pathway score.
Finally, the pathway scores for ground water, surface water,
and air are combined to obtain the HRS migration score.

The pathway scores are combined using a root-mean-square
approach to calculate the overall site score; that is, the final
HRS score is the square root of the sum of the squares of the
pathway scores divided by the square root of three. If all
pathway scores are low, the HRS score will be low. However, the
final score will be relatively high even if only one pathway score
is high. EPA considers this an important requirement for the HRS
scoring because some extremely dangerous sites pose threats through
only one migration mode. For example, leaking drums of hazardous
substances can contaminate drinking water wells, but if the drums
are buried deeply enough and the hazardous substances are not
very volatile, they may not release any hazardous substances
to the air or to surface water.

III. APPROACH TO HRS REVISIONS

EPA undertook a comprehensive review of the HRS in
developing this proposed rule. Based on comments to previous
rulemakings, and its own experience scoring sites, EPA prepared
or sponsored a series of issue analysis reports that formed
the basis for many of the options the Agency considered. These
reports covered issues such as methods to evaluate human food
chain exposure; methods to evaluate the potential for air
release; evaluation of appropriate air and surface water
target distance limits; factors to account for environmental
attenuation of hazardous substances in ground water and surface
water; methodologies for evaluating toxicity; methods for
determining direction of ground water flow and issues related
to the feasibility of using ground water flow direction measures;
and the feasibility of using hazardous substance concentration
data as well as evaluating waste quantity on a hazardous
constituent basis. These reports and other studies form the
basis of the Technical Support Document, available in the
Superfund docket for public review, which explains in detail the
basis for the options proposed in this revision.

To provide a broad spectrum of technical expertise in
developing these revisions, EPA sought information from a
number of sources. In 1986, EPA's Office of Emergency and
Remedial Response established an EPA work group to guide the
revision process. The work group's deliberations addressed
not only the broad structure and function of the HRS, but
also detailed technical issues.

A. Science Advisory Board

pathway scores are low, the HRS score will be low. However, the
final score will be relatively high even if only one pathway score
is high. EPA considers this an important requirement for the HRS
scoring because some extremely dangerous sites pose threats through
only one migration mode. For example, leaking drums of hazardous
substances can contaminate drinking water wells, but if the drums
are buried deeply enough and the hazardous substances are not
very volatile, they may not release any hazardous substances
to the air or to surface water.

III. APPROACH TO HRS REVISIONS

EPA undertook a comprehensive review of the HRS in
developing this proposed rule. Based on comments to previous
rulemakings, and its own experience scoring sites, EPA prepared
or sponsored a series of issue analysis reports that formed
the basis' for many of the options the Agency considered. These
reports covered issues such as methods to evaluate human food
chain exposure; methods to evaluate the potential for air
release; evaluation of appropriate air and surface water
target distance limits; factors to account for environmental
attenuation of hazardous substances in ground water and surface
water; methodologies for evaluating toxicity; methods for
determining direction of ground water flow and issues related
to the feasibility of using ground water flow direction measures;
and the feasibility of using hazardous substance concentration
data as well as evaluating waste quantity on a hazardous
constituent basis. These reports and other studies form the
basis of the Technical Support Document, available in the
Superfund docket for public review, which explains in detail the
basis for the options proposed in this revision.

A. Science Advisory Board

Several scientific questions were referred to the Agency's
Science Advisory Board for its review and recommendations.
The Science Advisory Board is a public advisory group providing
scientific information and advice to the Administrator and
other officials of EPA. The Board is structured to provide a
balanced expert assessment of scientific issues related to
problems facing the Agency. EPA referred three specific issues
to the Board: (1) options for revising the way toxicity of
hazardous substances is evaluated and scored in the HRS; (2)
the question of whether the HRS is biased against mining waste
and other high volume waste sites (including issues related to
the use of waste concentration data); and (3) the appropriate air
pathway target distance over which population exposure is assessed.
-------
-10-

EPA developed summary technical documents on these issues and
made several presentations to the Science Advisory Board, which
prepared a report addressing each issue. The technical documents
and Science Advisory Board recommendations are available in
the docket. The documents are: (l) "Discussion of Options For
Revising the Hazard Ranking System (HRS) Toxicity Factor," ICF,
Inc., May 1987; (2) "Analysis of the Air Target Distance Limit in
the Hazard Ranking System," EPA, 1987; (3) "The Superfund Hazard
Ranking System (HRS): Applicability to Mining Waste Sites," ICF
Inc., July 1987; and (4) "The Superfund Hazard Ranking System
(HRS): Feasibility of Using Concentration Data in a Revised HRS,"
ICF, Inc., July 1987. The Science Advisory Board's recommen-
dations are outlined in "Science Advisory Board Hazard Ranking
System Review Subcommittee: Review of the Hazard Ranking System,"
US EPA, 1988.

E. Review of Alternative Ranking Systems

One of the activities undertaken to develop a revised and
improved HRS involved examining alternative site evaluation
models chosen from a review of over 30 such systems. EPA
evaluated the models to determine: accuracy in predicting
potential risk; the extent to which the SARA requirements
were addressed by the model; the implementability of the model;
the amount of data that would have to be collected to evaluate
sites using the model; and the cost and time involved in gathering
these data.

In order to better understand the accuracy of these models,
EPA tested three of the site evaluation models, along with a
draft version of the revised HRS, on 20 sites. Through analysis
of these models, EPA hoped to better understand the factors
that affect the accuracy of site ranking methodologies, thereby
developing insights to guide revisions to the HRS.

The three alternative systems analyzed were the New York
State Human Exposure Potential Ranking System, the Air Force
Hazard Assessment Rating Methodology II, and the Department of
Energy Remedial Action Priority System. These systems were
chosen for further testing because they considered site-specific
conditions, were either fully developed and tested or in the
final stages of development, examined multiple media, and were
substantially different from the current HRS. None of them,
however, fully met the SARA requirements without some revisions.

EPA convened a site ranking panel of senior EPA staff and
managers selected to represent a cross-section of knowledge and
specialties. The panel members were to evaluate and rank the 20
sites according to the relative level of risk they perceived
the sites posed to human health and the environment. The
purpose of this exercise was to obtain expert judgments to
serve as a baseline for the comparative evaluations of the
site evaluation models and to gain a better understanding of
the relative weights of certain factors and exposure pathways
in evaluating the threat from a site.
-------
-11-

The 20 sites selected for the testing program were not
randomly selected and are too small a sample to be statistically
representative of the universe of potential Superfund sites.
These sites had been investigated under the Superfund program
and included NPL and non-NPL sites. The sites were selected to
represent a range of different types of sites (landfills, surface
impoundments, etc.) and a range of scores above and below 28.50,
the cutoff on the current HRS for placing sites on the NPL.
EPA also selected sites for which the necessary data were likely
to be available. In addition, EPA explicitly selected some
sites that contained features the current HRS migration routes
could not score (such as human food chain exposures, direct
contact exposures, and potential air releases). As a result,
caution should be exercised in generalizing the testing results.

The panel used the most complete data available to rank
the sites, more information than would normally be available at
the time a site is evaluated using the HRS. Remedial investiga-
tions or Public Health Evaluations had been performed at most
of the sites, providing quantitative risk estimates.

It is important to note that although the testing program
provided the Agency with some useful information, this approach
to assessing the site evaluation systems had some fundamental
limitations. First, only 20 sites were used for the evaluation,
and, as discussed above, these sites were not chosen randomly.
Second, the panel had access to data that, in some cases, could
not be used by (or were unavailable to) one or more of the models.
Third, the panel's assessments were based on the subjective con-
clusions of panel members; although some objective criteria were
applied, the same conclusions might not be reached by a second
panel. Finally, no firm consensus was reached by the panel at the
level of individual sites; rather, a fairly firm consensus on
"groupings" of sites was reached, and only after negotiations
and discussions among panel members. The limitations of the
testing program account for some of the differences between the
relative rankings of the expert panel and the rankings of the
various models tested.

Despite these limitations, the results of the comparative
evaluation of the models indicated that the draft revised HRS
best reflected the site ranking panel's consensus ranking.
This analysis.is discussed in more detail in "Analysis of Alter-
natives to the Superfund Hazard Ranking System" (Industrial
Economics, Inc., November 1988), as are each of the three
alternative models tested. Further discussion of the correlations
of the model rankings with the site ranking panel rankings
and an analysis of the reasons for these correlations are
provided in the above-referenced report, which is available
in the Superfund docket.

The model testing study suggested that the accuracy of the
HRS could be improved by ensuring that the model considers a
comprehensive set of exposure pathways and that the revised HRS
employ a weighting scheme that gives sufficient emphasis to
-------
-12-

hazardous sites dominated by risks along one or several exposure
pathways. In addition, the analysis suggests that the revised HRS -
should be flexible enough to take evidence of adverse health
effects and environmental damage into account in evaluating sites.
The analysis points out the sensitivity of site ranking models
to the data they employ, and the importance — subject to resource
constraints and the need to expeditiously evaluate sites — of
providing the revised HRS with the best available data. The re-
vised HRS attempts to implement these recommendations, as is dis-
cussed in greater detail later in this preamble.

Based on the evaluation of the HRS in the model testing study,
EPA determined that the draft version of the revised HRS met
all the statutory requirements, did not require more data than could
reasonably be collected within the limited scope of a site inspec-
tion, and was cost-effective and implementable.

C. Cutoff Score

The first NPL contained 418 sites. States had the
opportunity to designate one site as their single top
priority; the remaining sites were included because they had
HRS scores of 28.50 or higher. EPA chose the 28.50 cutoff
score as a management tool because it yielded an initial NPL
of at least 400 sites as required in CERCLA Section 105(8)(B)
(now CERCLA Section 105(a)(8)(B)), not because of any
determination that the cutoff represented a threshold of
unacceptable risks presented by sites.

In the Conference Report on SARA, Congress asked EPA to
address the relationship between risks at NPL sites and the
cutoff score. During its revision of the HRS, EPA addressed
this and other issues in a number of studies related to the
cutoff score. These studies were combined into one report,
which is available in the Superfund docket. ("SARA studies on
HRS Scores and Remedial Actions, HRS Scores and Potential
Dangers, and the Effect of the 28.50 Cutoff Score," CH2M
Hill, September 1988.) Although the study was limited in
scope and definitive conclusions are not possible, it did
indicate that some sites with scores below the cutoff can
also pose potential dangers to human health and the environment.
However, the cutoff score was not meant to set a no-risk
threshold, but rather to set a level above which a site
becomes a priority.2 Toward that end, the 28.50 cutoff has
been useful in identifying high priority sites for further
study and possible remedial action and has proven to
be an effective management tool. In general, NPL sites with
2 It should be noted that although sites scoring below 28.50
have not generally been placed on the NPL, they may be addressed
by CERCLA removal or enforcement authorities, or by State and
local governments, if response measures appear to be warranted.
-------
-13-
scores exceeding 28.50 present significant risks to public
health and the environment, necessitating some form of response,
On the other hand, three sites have been deleted from the
NPL after completion of a remedial investigation/feasibility
study (RI/FS). Based on the remedial investigations, EPA
concluded that none of the sites presents a significant threat
to public health or the environment and that no removal or
remedial action is necessary at those sites (51 FR 7935,
March 7, 1986; 53 FR 12680, April 18, 1988).

In the past, EPA considered the effects of both raising
and lowering the cutoff score for the NPL relative to the
current cutoff. Lowering the cutoff would add more sites to
the NPL and would tend to include more sites with lower risks
than sites currently on the NPL. EPA is concerned that
lowering the threshold might substantially increase the
number of NPL sites that are found to present no significant
threat to health or the environment after the RI/FS has been
completed. Raising the cutoff establishes a higher threshold
for newly scored sites and would tend to exclude from the NPL
sites that present significant risks to public health and the
environment. Since EPA believes that the current cutoff
score has been a useful management tool, the Agency is
proposing that the cutoff score for the revised HRS be
functionally equivalent to the current cutoff.

In light of the rather substantial revisions of the HRS
in this proposal, EPA concluded that it is necessary to
evaluate the practical effects of keeping the cutoff score at
28.50; that is, whether that score will continue to provide
an appropriate set of National Priorities for management
purposes. The Agency is examining several approaches for
defining "equivalent to 28.50." One alternative is to score
sites using both the current and revised systems; EPA would
then use statistical analyses to determine what revised HRS
score best corresponds to 28.50 on the current HRS. Another
alternative would be designed to yield an NPL of the same
size as would the current HRS and current cutoff score. That
is, EPA would estimate the size of the NPL if the current
approach were applied to the known inventory of sites and
then identify what cutoff score for the revised HRS would
result in the same number of NPL sites. A third alternative
involves identifying the quantitative risk levels that on the
average correspond to a current HRS score of 28.50 and then
determining what revised HRS score best corresponds to that
risk level.

EPA specifically requests comment on whether the cutoff
score for the revised HRS should be functionally equivalent
to the current HRS score of 28.50 and, if so, how to define
and determine functional equivalence. The Agency intends to
evaluate various cutoff score analyses based on the cost and
availability of data.

Although the Agency is proposing that the cutoff be
-------
functionally equivalent to the current score of 28.50, it is
premature to specify a numerical cutoff score for the revised
HRS at this time. As stated previously, every factor in the
current HRS has been revised, and new factors have been
included. An entirely new exposure pathway, the onsite
exposure pathway, has been included in the total site score.
Sites with certain characteristics (i.e., human food chain or
direct contact problems) may score higher on the revised HRS
than under the current HRS. Alternatively, certain sites may
score lower under the revised HRS because of factors that
allow target populations to be distance and dilution weighted,
and due to the addition of mobility factors. Thus, differences
in scores are anticipated between the current HRS and the
proposed revised HRS. While the Agency expects that the
changes will result in increased accuracy in assessing the
relative degree of risks to public health and the environment
for certain sites, it makes it difficult at this time to
identify an appropriate cutoff score. After performing further
analyses and reviewing public comments, EPA will select a
means of establishing a cutoff score for the revised HRS and
will announce that score in the preamble of the final rule.

D. SARA Conference Reports

In the Conference Report on SARA, Congress called on the
President to address a number of issues during the review of
the HRS (H.R. Rep. No. 962, 99th Cong., 2nd Sess. at 200
(1986)). Thus, in addition to the studies and reviews that
EPA performed based on its experience and on comments, the
Agency prepared the following:

1. An evaluation of the Preliminary Pollutant Limit Value
system used by the Department of Defense and comparison
with the HRS.

2. An explanation of how the HRS was developed and the
method of determining the relative hazards at different
facilities under the system.

3. A study determining the relationship of HRS scores
and the potential dangers to human health and the
environment.

4. An examination of the effect of establishing a
threshold value of 28.50 for facilities to be included
on the NPL.

5. A study determining the relationship between HRS
scores and the types of remedial actions that
are appropriate at such facilities.

These studies are available in the Superfund docket for this
proposal. (See "SARA Studies on HRS Scores and Remedial
Actions, HRS Scores and Potential Dangers and the Effect of
the 28.50 Cutoff Score," CH2M Hill, September 1988; "An Explana-
-------
-15-

tion of How the HRS was Developed and the Method of Determining
the Relative Hazards at Different Facilities under the System,"
US EPA, September 1988; and "Comparison of the Preliminary
Pollutant Limit Value (PPLV) System and the Hazard Ranking
System (HRS)," Versar, September 1988.) In addition, Congress
called for a determination of whether a new threshold value
should be established for inclusion of facilities on the NPL;
this subject is discussed in this preamble in Section III.C.

E. Advance Notice of Proposed Rulemaking

EPA published an advance notice of proposed rulemaking
(ANPRM) on April 9, 1987 (52 PR 11513), soliciting comments
on the revisions required by SARA as well as on the following
technical issues: existing scoring factors; other models
for ranking hazardous substance releases; mechanisms
for including direct contact in the HRS; and a mechanism for
incorporating human food chain exposures into.the HRS. on
May 7 and 8, 1987, EPA held a public meeting on the HRS.
The comments received during the public meeting and in response
to the ANPRM have been reviewed and considered in the development
of this Notice of Proposed Rulemaking. EPA will respond in
detail to all comments when the final rule is promulgated.

IV. BACKGROUND DOCUMENTS

The proposed revisions to the HRS are discussed in
three primary documents: (1) the proposed rule, (2) this
preamble, and (3) the "Technical Support Document: Revised
Hazard Ranking System," ("Technical Support Document"). The
proposed rule outlines the scoring system, emphasizing the
mechanics of scoring sites. This preamble provides an overview
of the scoring system, along with concise explanations of why
the changes were made.

The Technical Support Document contains a more detailed
explanation of the technical basis for the proposed revisions
to the HRS, along with descriptions of the options considered.
The Technical Support Document follows the same general outline
as the preamble, with one section describing revisions that
affect more than one pathway (e.g. toxicity), and the remaining
sections describing the four pathways of the revised HRS.
Each discussion in the Technical Support Document generally
contains a description of the current HRS, the options considered,
the revisions that are proposed for the revised HRS, and the
technical justifications for the option chosen. In addition,
the Technical Support Document references other background
documents that provide an even greater level of detail on the
proposed revisions. These documents, along with the Technical
Support Document, are available to the public in the Superfund
docket. To facilitate public review, EPA has prepared an
index to the proposed rule, the preamble to the proposed
rule, and the Technical Support Document with detailed cross
referencing of issues. This index will be available in the
Superfund docket. See the ADDRESSES section of this preamble
for further information on the Superfund docket.
-------
-16-

V. DISCUSSION OF THE PROPOSED RULE

A. Overview

As stated above, the current HRS is incorporated in the
NCP as Appendix A to 40 CFR Part 300. Appendix A, which is
essentially a user's manual for the HRS, includes the forms
as well as instructions for assigning values to each of the
factors.

The current Appendix A provides instructions for evaluating
five pathways. The surface water, ground water, and air
pathways comprise the migration pathways and are used in
determining the HRS site score. Direct contact and fire and
explosion may be calculated to determine if removal action is
warranted. EPA does not believe that the fire and explosion
calculation would provide a useful basis for scoring a site
for remedial action. The potential for fire and explosion is
considered in another part of the Superfund program, the
removal program, when determining if a removal action is
necessary. Therefore, the proposed HRS would delete the fire
and explosion calculations. For the reasons discussed in
Section V D 4, the current direct contact calculation would
also be deleted and replaced by an onsite exposure pathway.
This new pathway would be included in the HRS site score.

As can be seen from the diagrams preceding Sections V D
1 through 4, the essential structural features of the revised
HRS would generally remain the same as those of the current
HRS. However, every factor has been revised or is new. A few
factors would be eliminated, either because they do not
discriminate among sites or because they would be replaced by
more accurate measures.

The remainder of this preamble discusses the proposed
changes to Appendix A. After an overview of the four pathways
(both in their current and revised state), Section V B presents
a general discussion of the major proposed revisions. Following
that, Section V C describes in detail issues and factor
revisions that affect more than one of the pathways. Sections
V D 1 through 5 discuss all other changes to each specific
pathway. Finally, Sections V E discusses an issue that has
been the subject of special attention, wastes designated as
special study wastes under RCRA.

1. Ground Water

The ground water migration pathway in both the current
and revised HRS evaluates the likelihood that hazardous substances
at a site or facility will migrate through the ground beneath
them and contaminate aquifers. If the hazardous substances
reach an aquifer, the substances can potentially be transported
through the aquifer and contaminate drinking water wells that
draw from that aquifer.
-------
-17-

If hazardous substances have been released to an aquifer
or if the site characteristics make a release likely, the
principal questions the ground water pathway evaluates are the
impact of releases on the ground water resources and on the
people who draw their drinking water from potentially contaminated
wells. The revised HRS is designed to reflect the concept that
hazardous substances in ground water tend to become increasingly
diluted as distance from the site increases.

2. Surface Water

For the purposes of the HRS, surface'water is defined as
perennial streams, rivers, lakes, oceans, and intermittent
streams and ditches in arid and semiarid regions. If contaminated
runoff has reached surface water or if the site characteristics
make a release to surface water likely, the HRS evaluates the
potential for the release to affect people or the environment.
The revised HRS would include factors for evaluating flood
potential and address the effect of hazardous substances on
the human food chain. In addition, the revised HRS would
evaluate risks from recreational exposures.

3. Air

The current HRS air pathway is evaluated only if hazardous
substances at a site have escaped into the air either as
gases or as particulate matter. The revised HRS would also .
consider the site characteristics to assess the potential for
releases to occur even if no release has been documented.
Once the likelihood of release has been determined, the main
questions are how many people and sensitive environments
could be exposed to hazardous substances carried in the air
and the inherent hazard associated with potential exposures.

4. Onsite Exposure Pathway

The onsite exposure pathway deals with the possibility
that people or sensitive environments will have direct, physical
contact with hazardous wastes or contaminated soil. The
revised HRS would look at two populations to assess the risk.
The resident population consists of those people who live or
go to school or day care on land that is contaminated. The
nearby population consists of those people who live within a
one-mile travel distance of the site and might have access to
the site.

B. Major Revisions

The following is a summary discussion of the major
revisions being proposed. More detailed specifics of each
proposed revision as well as discussions of legal requirements,
of options EPA considered, and of the reasons for EPA's decisions
are provided in Sections C and D, which also give references
to the rule and to supporting documents available in the docket.
-------
1. Actual and Potential Contamination

The current HRS evaluates the potential for exposure of
populations and does not give additional consideration (i.e.,
higher scores) to situations where human exposures have been
documented. The current HRS is also primarily oriented toward
population risk rather than individual risk. EPA is proposing
changes in HRS calculations so that the HRS will more accurately
reflect the potential risk to individuals and to populations
exposed to documented contamination.

In assessing drinking water threats in the proposed
revisions, target populations in the ground water and surface
water pathways would be divided into four groups: people
exposed to documented contamination above health-based benchmarks;
people exposed to documented contamination not exceeding
health-based benchmarks, but significantly above background
concentrations (two groups are defined on the fraction of the
benchmark present); and people potentially exposed to contamination
from a site. The health-based benchmarks would be based on
the National Primary Drinking Water Standards (maximum contaminant
levels (MCLs)) (40 CFR Sections 141.11 through 141.16 and 52
FR 25690, July 8, 1987). If no drinking water standard has
been developed by the Agency for a substance, the health-based
benchmark would then be based on maximum contaminant level
goals (MCLGs) for noncarcinogens. For carcinogens and potential
carcinogens that have no MCLs, risk-specific concentrations
corresponding to an individual cancer risk of 10~4 would be
used as the benchmark instead of MCLGs. Where contamination
above a health-based benchmark occurs, people exposed to this
contamination would be weighted most heavily in determining
the factor score. (See Section V C 5.) The sensitive environments
subpathway in the surface water pathway would also be evaluated
on actual and potential contamination, using ambient water
quality criteria as ecologically-based benchmarks.

Likewise, the human food chain subpathway in the surface
water pathway assigns a higher value where a fishery has actually
been closed or shows contamination over an action level set
by the U.S. Food and Drug Administration (FDA) than where a
fishery has not been closed and is not known to be contaminated
above action levels. The recreation subpathway assigns a
higher value where a recreation area has actual contamination
father than potential contamination.

The potential risk to the maximally exposed individual
(MEI) would be represented by the distance to the nearest
drinking water well in the ground water pathway, the streamflow
at the nearest drinking water intake in the surface water
pathway, and the distance to the nearest occupied building in
the air pathway. When there is contamination above health-based
benchmarks in any well or water intake within the target
distance, the MEI factor would be assigned the maximum value
in the ground water or surface water pathways to ensure that
documented contamination is heavily weighted (See Section V c 9.)
-------
-19-

2. Dilution/Distance Weighting

In the current HRS, weighting targets based on dilution/
distance is explicitly included only in the air pathway,
although such weighting is implicit in some factors in the
ground water and surface water pathways. Because, under most
circumstances, the concentration of hazardous substances
declines as the substances migrate from a site, the revised
HRS would apply dilution/distance weighting directly to relevant
target category factors to better reflect the differential
exposures and risks to targets located at varying distances
from a site. The Agency believes such a revision would improve
the accuracy of the HRS.

For surface water, the weighting factor would be based on
dilution as reflected by the average annual flow; in air and
ground water, distance- would be used as a surrogate for dilution.
The weighting of the population at different distances is
based on the results of environmental transport models. In
the ground water and surface water pathways, dilution/distance
weighting of targets would be used for those populations who
do not have actual contamination in their drinking water
wells or drinking water intakes, but where the aquifer or
surface water body is contaminated or has the potential to be
contaminated. In the air pathway, all targets would be distance
weighted. In the onsite pathway, the nearby population would
be distance weighted, reflecting the likelihood of people
visiting the site, not the potential decrease in concentration
as hazardous substances migrate from a site. (See Section V C 8.)

3. Toxicity

Under the current HRS, the toxicity factor scoring is
based primarily on acute toxicity of hazardous substances.
However, EPA recognizes that adverse health effects at
hazardous waste sites may result from chronic exposures as
well as from acute exposures. To include the consideration of
such risks, EPA is proposing to change the basis of the toxicity
factor score. The current toxicity values are based on either
the Sax rating system or the rating system of the National
Fire Protection Association. The revised HRS would evaluate
hazardous substances and assign scores for three kinds of
toxicity: acute toxicity, carcinogenicity, and chronic
noncarcinogenic toxicity. The highest of the three scores for
a hazardous substance would become the toxicity factor value
assigned to that substance. EPA is proposing to score hazardous
substances using a system based on Reference Doses for chronic
noncarcinogenic toxicity; Cancer Potency Factors combined
with qualitative weight-of-evidence for carcinogencity, or,
when the Cancer Potency Factor is not available, the ED10;
and LD50 or LC50 values for acute toxicity. In addition,
EPA is proposing to include aquatic toxicity ratings to assess
potential risks to aquatic ecosystems. (See Section V C 2
and Section V D 2.) Toxicity values for a substance would be
combined in a matrix with mobility or persistence factors to
-------
-20-

calculate the final toxicity/mobility values for the ground
water and air pathways or toxicity/persistence values for the
surface water pathway.

4. Mobility

The current HRS does not directly consider the properties
of substances that affect their ability to be released and
migrate through environmental media; therefore, the current
HRS may not differentiate well between two highly toxic
substances with very different mobilities. The addition of a
mobility factor should better reflect the risks from a site.
Thus, EPA is proposing to incorporate mobility factors that
would combine with toxicity in a matrix calculation in the ground
water and air pathways to create a more accurate measure of
the likelihood that a particular substance will migrate- to
ground water or to the atmosphere and expose potential targets;
the air pathway would consider the mobility of both gases and
particulates. (See Sections V D 1 and 3.)

Although the surface water pathway has no mobility
factor per se, the persistence factor addresses this issue,
as would the proposed bioaccumulation factor in the human
food chain calculations and the proposed dose adjusting
factor in the recreation calculations. (See Section V 0 2.)

5. Hazardous Waste Quantity

Hazardous waste, in addition to including some proportion
of hazardous substances, almost always includes nontoxic
substances. When the current HRS was developed, EPA judged
that the cost of reliably determining the amount of hazardous
constituents within the hazardous wastes at a site was prohibitive
and, in some cases, technically impossible. Therefore, the
current HRS was designed to use the total quantity of waste
instead of the quantity of hazardous substances in the waste
for the calculation of the hazardous waste quantity factor.

EPA is proposing to modify the existing method of calculating
the hazardous waste quantity factor to make it a more accurate
reflection of relative risk. The proposed tiered approach
would allow for the use of calculated amounts of hazardous
substances for sites where the contaminant concentrations in
a given waste have been determined based on adequate sampling
and analytical methods. If these concentrations are not
available, waste quantity as deposited could be used, as
could source volume or source area. (See Section V C 3.)

6. Sensitive Environments

The sensitive environments considered under the current
HRS include wetlands and areas that are critical habitats for
plants and animals on the Federal endangered species list.
EPA is proposing a significant expansion of the sensitive
environments eligible to be scored in order to more fully
-------
-21-

address CERCLA's original mandate to consider threats to both
public health and the environment, as well as to respond to
the SARA requirement to accurately assess relative risk to
human health and the environment, to the maximum extent feasible,
In the revised HRS, sensitive environments would include
lands and waters that have been legally designated as protected
areas by either the Federal government or state governments,
as well as areas that have been identified by the Nature
Conservancy's National Heritage Program. (See Section V c 6.)

7. Onsite Exposure Pathway

The direct contact portion of the current HRS calculates
the potential for direct exposure to hazardous substances in
a way that essentially parallels the surface water, ground
water, and air pathways; however, the calculation is not
included in the score used to determine a site's eligibility
for the NPL. Currently, the direct contact pathway could be
used to determine whether removal action is required at a
site. An analysis of decisions on remedial actions indicated
that some significant risks from direct contact may not have
been completely addressed by removal actions and should be of
concern in determining priorities for remedial action. To
ensure that the potential for such contact is factored into
the HRS score, EPA is proposing to incorporate exposure to
onsite wastes and contaminated soils into the HRS migration
score by adding a separate pathway. The onsite exposure
pathway would be included in the HRS score to better respond
to CERCLA's original mandate to take into account the potential
for direct human contact in setting priorities, as well as to
respond to the requirement in SARA to accurately assess relative
risk to the maximum extent feasible.

The proposed onsite exposure pathway would separately
assess two populations, those people who live or attend
school or day care on the contaminated site and those who
live nearby and have access to the site. Resident children
under seven would be considered the high risk population and
would be weighted more heavily than adults because of their
greater likelihood of ingesting onsite contaminants. (See
Section V D 4.)

8. Surface Water

In the current HRS, the surface water pathway is primarily
concerned with the potential contamination of drinking water
and with the population that could be affected by this contami-
nation. A lesser weight is given to the impact of contaminants
on sensitive environments in surface water. As required by
SARA, EPA has considered other targets and is proposing to
evaluate separately the potential contamination of the human
food chain (based on fishery contamination) and recreational
exposures to contaminated surface water. Sensitive environments
would also be assessed separately. In addition, new factors
to assess flood potential would be incorporated into the
-------
-22-

likelihood of release factor category. The factors used to
assess route characteristics in the current HRS would be
replaced by a new set of factors that better assess overland
release potential. (See Section V D 2.)

9. Air

The HRS currently evaluates the air pathway solely on
the basis of site monitoring data; if no release of contaminants
has been documented at the site, the air pathway part of the
HRS is assigned a score of zero. In contrast, for the surface
water and ground water pathways, the potential for a release
to occur is considered where no observed release has been
documented.

CERCLA Section 105(a)(8)(A), as amended by SARA, specifies
that EPA should consider potential releases to the ambient
air in revising the HRS. Accordingly, the proposal described
in Section V D 3 includes a method to assess potential air
releases.

C. Revisions Affecting Multiple Pathways

This section discusses in detail those issues that affect
more than one pathway. Although some of these issues were
covered in the general discussion in the major revisions
section, this discussion reviews these cross-cutting issues
in more detail and describes the options EPA considered to
resolve these issues. Because the proposed changes are similar
in all the pathways affected, to prevent repetition, the
specific revisions for each pathway are discussed in this
section rather than in the individual pathway discussions in
Section V D. For most of these issues, detailed descriptions
of the options reviewed and the reasons for EPA's choice can
be found in the Technical Support Document, available in the
Superfund docket. Other related documents and tables for
determining some factor values are also available in the
Superfund docket.

1. Structure

The proposed HRS would retain the general structure of
the current HRS, as described in Section II; that is, the
structure of the revised HRS would continue to be based on
evaluating the relative risk of sites through the use of
factors, factor categories, and pathways to obtain a final
site migration score.

This structure was retained after a careful evaluation
of possible alternative structures to the existing HRS,
including alternatives to the entire structure, to factor
categories, and to individual factors. During this process,
EPA reviewed over 30 available site evaluation systems and
over 55 chemical ranking systems, including several systems
developed expressly for ranking hazardous waste sites.
-------
-23-

Evaluation of Overall Site Migration Score

During the course of this evaluation, EPA considered
several methods for revising the way in which pathway scores
are combined to calculate the overall HRS site score. EPA
evaluated the possibility of adding the pathway scores, but
found that with the proposed structure, this approach tends
to discount the effects of risks involving only one pathway.
The results of testing a system that sums pathway scores are
consistent with these findings. Other options considered
included selecting the single highest pathway score, using a
root-mean-square approach for the two highest pathways, and
using the root-mean-square approach for the three highest
pathways. The Agency believes, however, that using the root-
mean-square approach for all pathways in the revised HRS most
appropriately considers the risk through all four pathways,
without discounting risks involving only one pathway. Thus,
the structure for combining pathway scores into a single site
score is the same in both the current HRS and the proposed
revisions to the HRS.

Evaluation of Pathways

The major changes to the pathways of the existing HRS
are: (1) elimination of the direct contact route; (2) incor-
poration of an onsite exposure pathway in the calculation of the
total HRS site score; (3) elimination of. the fire and explosion
route; (4) modification of the surface water pathway; and (5)
addition of a potential to release calculation to the air
pathway.

For the onsite exposure pathway, EPA is proposing that
the maximum score for the entire pathway be equal to the
maximum score that could be assigned to either the resident
or nearby population (Proposed Rule, Section 5.0). EPA
concluded that exposures to either group can represent a
reasonable worst case risk and that this case should be reflected
in the final migration score.

Four threat categories have been incorporated in the
surface water pathway: drinking water, human food chain,
recreational, and environmental. Each of these threats is
evaluated, structurally, in a manner consistent with surface
water threats in the existing HRS. The four threat scores
are added to form a total pathway score.

Evaluation of Factor Categories

In the current HRS, the value for the targets factor
category in a pathway is the sum of individual factor values.
The maximum factor category value can be obtained only if
every factor is assigned its maximum factor value. This
means that under the current HRS, a site significantly
affecting 10,000 people would not receive a maximum score for
the surface water pathway unless sensitive environments are
-------
also given a maximum value. Similarly, a site would not
receive a maximum score for the air pathway if no one lives
within a quarter mile of the site and, for the ground water
pathway, a site would not receive a maximum score if there
were no intakes or wells within 2,000 feet of hazardous
substances.

The proposed rule would modify the method used to add
target factor values so that some sites that do not meet the
criteria for the maximum value for each target factor could
still receive the maximum score for the overall target
category> The sum of the available points for each type of
exposure would be greater than the number of points allowed
for the factor category. Under this approach it is possible
to allocate a significant number of points for sensitive
environments without reducing the importance of human health
risks. One effect of this revised scoring system would be to
condense scores at the upper end of the scale and more
accurately assess relative risks elsewhere.

As is the case in the current HRS, all factor categories
in the revised HRS have the same relative weight. That is,
the maximum value for the waste characteristics category has
the same weight as the maximum value for the targets category,
and each of those has the same weight in the air pathway as
in the ground water pathway. This is true despite different
factor category values because the values are multiplied and
normalized.

Evaluation of Factors

EPA is proposing structural changes in the way some
individual factors are evaluated and in the relative scoring of
factors within the targets factor category. To put this in
context, the relationship of factors within categories must
be understood. The relative importance of factors within
different factor categories cannot be compared by their
maximum factor values alone. Rather, their relative
importance depends on the maximum percentage of their category
they can account for; e.g., a factor with a maximum value of
50 contributes more to its category if the category maximum
is 100 than if it is 200. Within the same factor category in
a single pathway, the relative importance of factors is
comparable based on their maximum values. In other words, in
the surface water pathway the relative importance of drinking
water population versus surface water use can be compared
based on their maximum factor values. However, the relative
importance of drinking water population in surface water versus
ground water cannot be compared based on their factor values;
their relative importance can only be compared based on the
percentage their maximum value contributes to the maximum
value of their factor category.

In the current HRS, factors are weighted by assigning a
factor value and then applying a specified multiplier. The
-------
-25-
revised HRS would eliminate the explicit weighting by use of
a multiplier and would incorporate a measure of relative
importance of the factor to the factor category through the
factor values themselves. To assign relative values among
factors within factor categories for this proposed revision,
EPA combined the results of model testing programs, a review
of site ranking experience, and the results of analytical
models used to model fate and transport of hazardous
substances. In evaluating these data, EPA used two primary
considerations for assigning the appropriate value to a given
factor: (1) the relative importance of the factor as an
indicator of risk and (2) the expected accuracy with which
the factor can be measured or estimated based on site
inspection data.

Potential Revisions Under Consideration

The Agency is still considering a variety of revisions
to the algorithm and factor scales in the revised HRS.

The EPA Science Advisory Board, in addition to its
review of specific issues identified by EPA, offered ideas
regarding revisions to the algorithm used to calculate the
current HRS score and the factor scales. The Board suggested
that the Agency assess the current HRS algorithm to determine
if changes in the algorithm could provide increased accuracy
without increasing data collection costs.

Specifically, the Science Advisory Board suggested that
the current HRS algorithm be revised to more closely resemble
a guantititative risk assessment, with simplifications made
to account for the difficulties of data collection. The
Board also suggested that the factor category scales be
revised so that the logarithm of the actual number for a
factor becomes the factor value. Factor values might then be
summed to develop pathway scores.

Another change in the algorithm could include moving the
waste quantity factor from the waste characteristics category
to the likelihood of release category. Such a move would
increase the importance of toxicity and mobility/persistence
factors because these would be the only factors remaining in
the waste characteristics category. The change might also
provide better discrimination among sites based on likelihood
of release.

Another change in the structure might involve removing
the maximum values from some factors (e.g., waste guantity or
population) or from all factors. Scores would not be
normalized to a 100 point scale under this approach and there
would be no maximum possible score. In addition, EPA may
change specific values for certain factors (and/or factor
categories) to better reflect their relative importance.

EPA is planning to evaluate and possibly test such
-------
changes in the algorithm prior to promulgating a revised HRS.
Commenters should consider these and other possible revisions
to the algorithm.

2. Toxicity

Toxicity, a factor in the waste characteristics category
for all HRS pathways, is intended to represent the relative
potential of a substance to cause adverse health effects.
The toxicity factor does not provide an absolute assessment
of toxicity; each substance is assigned a value based on
its relative toxicity, and that value is used to determine
the relative toxic potential of substances at sites.

The current HRS approach to scoring toxicity is based on
the National Fire Protection Association (NFPA) rating scheme
and the toxicity ratings developed by N.I. Sax. Using one of
these ratings, both of which place primary emphasis on the
acute toxicity of a substance, the HRS assigns a toxicity
factor value from 0 to 3. Only the single highest scoring
substance available for release in a pathway is used in
assigning a value to the toxicity factor for that pathway.

Various rulemakings on the NCP and NPL, the subsequent
applications of the HRS to uncontrolled hazardous waste sites,
and EPA's request for comments in the ANPRM have raised a
number of issues concerning the current method of assessing
toxicity. In particular, commenters have questioned whether
chronic toxicity and carcinogenic effects are adequately
addressed.

As a first step to revising the HRS toxicity factor, EPA
reviewed over 55 chemical ranking systems, as well as the
toxicity components of over 30 site ranking systems. EPA
then evaluated in more detail a number of methods to characterize
and score toxicity, and presented several options to the
Science Advisory Board, including an option based on Reportable
Quantities (RQ); an option based on Reference Doses (RfDs)
and Cancer Potency Factors; and an option using modified
Acceptable Daily Intakes (ADI) and- a modified weight-of-evidence
approach. These options are discussed more fully in "Discussion
of Options for Revising the Hazard Ranking System (HRS) Toxicity
Factor" (ICF, Inc., May 1987), available in the Superfund docket.

EPA developed the RQ ranking system to aid in setting
reportable quantities for hazardous substances as required by
CERCLA; the system is described in detail in the following
Federal Register notices and their supporting material: 50 FR
13456, April 4, 1985; 51 FR 34534, September 29, 1986; and 52 FR
8140, March 16, 1987. In the RQ ranking scheme, each CERCLA
hazardous substance is assigned to one of five RQ categories.
Each category corresponds to a weight, in pounds, above which
releases must be reported. Under the option developed for using
the RQ approach in the HRS, three toxicity types would be
considered: chronic noncarcinogenic toxicity, carcinogenicity,
and acute toxicity.
-------
-27-

The RfD/Cancer Potency Factor option would use the two
quantitative toxicity parameters for chronic toxicity generally
used by EPA in site-specific risk assessments: the RfD for
noncarcinogenic effects and the Cancer Potency Factor combined
with the qualitative weight-of-evidence for carcinogenicity;
acute toxicity would not be considered under this option.

The Agency defines the RfD as an estimate (with uncertainty
spanning perhaps an order of magnitude) of a daily exposure
to the human population (including sensitive subgroups) that
is likely to be without an appreciable risk of deleterious
effects during a lifetime. The determination of the RfD
requires scientific judgments as to the appropriate NOAEL
(No Observed Adverse Effect Level), uncertainty factors, and
modifying factors. Uncertainty factors are reductions in
dose rate that are introduced to account for areas of scientific
uncertainty such as species extrapolation and variability
within the human population. General rules have evolved for
determining the overall uncertainty factor to use with various
data sets. However, the application of these rules in a
particular instance needs to be examined on a case-by-case
basis, exercising scientific judgment as to the quality and
quantity of the available data. As the magnitude of the
uncertainty factor increases, the estimate of the RfD becomes
less precise. The RfD is viewed by most toxicologists as a
"soft" estimate. While exposures higher than the RfD are
associated with increased probability of adverse effects,
that probability is not a certainty, since the calculation of
a RfD includes consideration of sensitive subgroups. Similarly,
while the RfD is seen as a level at which the probability of
adverse effects is low, the absence of risk to all people
cannot be assured at this level. Rfbs, which undergo a formalized
Agency-wide peer review and verification process, are derived
from available chronic and subchronic toxicity studies.

Cancer Potency Factors, which are developed by EPA's
Carcinogen Assessment Group and used to estimate potential
carcinogenic risk, are derived from studies on experimental
animals or from human epidemiologic data, if available. The
weight-of-evidence is defined as the overall strength of the
data indicating potential carcinogenicity, based on an evaluation
of all relevant studies and the nature and type of responses.
Methods for estimating Cancer Potency Factors and evaluating
weight-of-evidence are both described in more detail in EPA's
Guidelines for Carcinogenic Risk Assessment (51 FR 33992,
September 24, 1986).

The modified ADI/modified weight-of-evidence option
included three human toxicity types: carcinogenicity,
mutagenicity, and teratogenicity (CMT) considered as a group;
chronic, non-CMT toxicity; and acute toxicity. All substances
would be scored for all three types of toxicity. The modified
ADI method for assessing chronic noncarcinogenic toxicity is
similar to the RfD method, but the ADI values derived are not
subject to Agency-wide peer review and can be based on a
-------
-28-

wider range of toxicity data (e.g., acute data) than can RfDs.

The Science Advisory Board indicated that all three
options that EPA presented seem much better than the Sax
rating method and recommended that Sax be replaced. The
three options use essentially the same data bases for toxicity,
the Board stated, although the data are processed differently.
The Board preferred the RfD option for assessing chronic
noncancer toxicity.

Based on the Board review and input from the EPA work group,
EPA is proposing a scoring method that combines elements of
several of the options originally presented to the Board.
This combined approach would be based on three toxicity types:
carcinogenicity, chronic noncarcinogenic toxicity, and acute
toxicity. To evaluate the potential carcinogenicity of
substances, the revised HRS would use Cancer Potency Factors
combined with the qualitative weight-of-evidence. Where
Cancer Potency Factors are not available, ED10s, (i.e.,
dose at which a 10 percent response is observed) developed to
assess carcinogenicity for setting RQs, would be converted to
an equivalent scale and used for scoring.

For chronic noncarcinogenic toxicity, the revised HRS would
use a method based on verified RfDs because they represent
EPA's best scientific data and judgment regarding the
.potential noncarcinogenic effects of substances. RfDs are
currently the most widely used values for evaluating chronic
noncarcinogenic toxicity in EPA risk assessments. Acute
toxicity scoring would be based on the LDc0 or LC50 of a
substance, or the LD10 or LC10 if LDso an° LC50 are
not available.

In scoring toxicity using the revised HRS, separate
toxicity scores would be developed for each HRS pathway,
based on the substances relevant to that pathway. Within a
pathway, the relevant hazardous substances would be identified
— those associated with an actual or potential release to
ground water, surface water, or air, or, for the onsite pathway,
those associated with onsite exposure potential. For each substance
identified, the toxicity rating would be determined using the
methodology in the proposed rule (Section 2.2.1.1) or a
reference table based on that methodology. Toxicity ratings
for individual substances would be integer values on a scale
of 0 to 5, with 5 being the most toxic and 0 representing
insufficient information to score. If adequate toxicity data
are available, each substance would receive three subscores —
one for acute toxicity, one for chronic noncarcinogenic toxicity,
and one would be the highest subscore. The rating scales
developed for the revised HRS give less weight to acute toxicity
relative to chronic toxicity because remedial actions are
almost always in response to concerns over exposures associated
with potential carcinogencity or chronic toxicity (Proposed
Rule, Section 2.2.1.1).
-------
-29-

Asbestos and radionuclides are classified as human carcino-
gens (the highest weight-of-evidence category), but their
cancer potency values are not expressed in units directly
comparable to most other substances. Therefore, asbestos and
radionuclides cannot be evaluated and scored using the proposed
system. For purposes of HRS scoring, asbestos and radionuclides
would be assigned a 5, the highest toxicity value for carcinogens.

A default value of 3 would be used for a toxicity factor
value when appropriate toxicity data for scoring does not
exist for all hazardous substance relevant to that pathway or
threat, a situation that EPA anticipates would be very rare.
EPA solicits comment on this default value, which is the
midpoint of the scoring range and approximate median of the
substances scored to date by EPA.

The proposed rating scales for the three toxicity types
(carcinogenicity, chronic noncarcinogenic toxicity, and acute
toxicity) are provided in the proposed rule. A more thorough
discussion of the development of these rating scales and the
overall toxicity factor is provided in the Technical Support
Document, available in the Superfund docket. EPA solicits
comment on these rating scales.

The proposed toxicity scoring methodology would use
readily available, high quality toxicity data, rely on a data
base that is large enough to provide a score for all sites,
and be consistent with existing EPA procedures. For the
ground water, surface water, and air pathways, the overall
toxicity value for each hazardous substance would be combined
with its corresponding mobility or persistence value in a
matrix to assign a toxicity/mobility or toxicity/persistence
value. The highest toxicity/mobility or toxicity/persistence
value relevant to a pathway or threat would become the factor
value.

A hazardous substance's potential ecosystem toxicity, which
would be considered in the surface water pathway, would be evaluated
using the following hierarchy of data: EPA chronic water
quality criteria, EPA acute water quality criteria, or the
lowest LC50 value for the substance (Proposed Rule, Section
4.4.2.1.1). (See Section V D 2.)

Past commenters have questioned the use of the single
highest scoring hazardous substance in each pathway to score
toxicity in the current HRS. EPA reviewed a number of options
related to the number of hazardous substances scored for a
site including (1) retaining the current method of using
the highest scoring hazardous substance in each pathway to
score toxicity; (2) basing the toxicity value on all hazardous
substances known to be at the site; (3) basing the toxicity
value on a fixed percentage or number of hazardous substances
known to be present at the site; and (4) giving extra points
to sites with a large number of hazardous substances. These
options are more fully discussed in the Technical Support Document
-------
-30-

and in the paper presented to the Science Advisory Board:
"Discussion of Options for Revising the Hazard Ranking System
(HRS) Toxicity Factor" (ICF, Inc., May 1987). Both documents
are available in the Superfund docket.

EPA is proposing to retain the current method of scoring
the toxicity factor based on the single highest scoring
hazardous substance applicable in each pathway or threat
because the Agency believes that the single most hazardous
substance present in a pathway or threat generally provides
an adequate screening level evaluation of relative hazards of
the pathways, particularly when coupled with mobility or
persistence, as applicable. The sampling conducted during a
site inspection probably would not provide sufficient information
on the relative quantities or concentrations of hazardous
substances at a site. EPA believes that the proposed method
will ensure that any site where highly toxic substances are
present will receive a high toxicity score. EPA solicits
comments on the use of a single hazardous substance per pathway
or threat to score toxicity and on the alternatives discussed
above.

3. Hazardous Waste Quantity

In the current HRS, hazardous waste quantity is the
amount of waste containing hazardous substances (as defined
in CERCLA Section 101) present at a site, excluding any wastes
that are contained such that they cannot migrate. Values for
the hazardous waste quantity factor are combined with values
for the toxicity/persistence factor (in the ground water and
surface water pathways) or toxicity (in the air pathway) to
produce the final waste characteristics category score. If
it is not possible to make any determination of the hazardous
waste quantity at a site and it is known that hazardous
substances are in the waste, a default value of one for hazardous
waste quantity is assigned (the range is 0 to 8). EPA has
found that about 20 percent of all sites studied are assigned
the default value for waste quantity, signifying the absence
of data. This finding indicates that even hazardous waste
quantity, which is generally easier to estimate than hazardous
substance concentration, is still difficult to estimate.

In preparing revisions to the HRS, EPA considered a
number of alternatives to the current HRS method of calculating
hazardous waste quantity and presented two to the Science
Advisory Board: one would require use of hazardous substance
concentration data to estimate waste quantity at all sites
and the other would be a tiered approach that would use hazardous
constituent concentration data, waste quantity, volume, or
surface area of the source, in that order. These options are
discussed in detail in a paper prepared for review by the
Science Advisory Board: "The Superfund Hazard Ranking System
(HRS): Feasibility of Using Concentration Data in a Revised
HRS" (ICF, Inc., July 1987), and available in the Superfund
docket.
-------
-31-

In general, having adequate hazardous substance concentration
data might enable sites to be evaluated more accurately.
However, the cost of obtaining these data at all sites would
be substantial. If accurate records of wastes were available
and if the concentration of contaminants in the wastes were
known, it would be possible to calculate the amounts of hazardous
constituents. EPA's experience, however, has been that such
waste disposal records are frequently not available at Superfund
sites, and when they are available, they are often incomplete
and insufficiently detailed to estimate constituent quantities.

The comprehensive site sampling needed to estimate hazardous
substance concentrations and amounts with known accuracy
would not be feasible on a routine basis, given the resources
available and the statutory requirement to expeditiously
evaluate sites. Wastes at hazardous waste sites are typically
heterogeneous. Depending on the nature and history of the
site, very great differences in waste composition may occur
over just a few meters, with different sets of constituents
appearing and with levels varying by orders of magnitude.
Attempting to determine a single representative concentration
of a constituent or to estimate the total amount of all hazardous
substances would be very difficult at most sites. Temporal
variability also diminishes the relevance of any estimate of
concentrations. Data from a single sampling provide only a
"snapshot" of current conditions. Mobile substances may have
already moved into environmental media and only the least
mobile may remain at significant levels at the source. Thus,
substance concentration data taken only from waste source
materials may result in underestimates of waste mass available
for transport. ,

In evaluating alternatives for developing a hazardous
waste quantity value, the Agency recognized that, at some
sites, sufficient data may be available to determine the
concentration of hazardous constituents. At most sites,
however, obtaining these data would be difficult and costly.
Thus, the Agency is proposing the tiered approach to scoring
hazardous waste quantity in the revised HRS (Proposed Rule,
Section 2.2.2). As the Science Advisory Board stated, the
tiered system "would encourage the use of concentration data,
but would also provide the flexibility to use indirect estimates
of a constituent's mass when direct measurements of concentration
are not available."

The tiered approach involves the development of a single
hazardous waste quantity value for each pathway at a site.
This factor would be based, in order of preference, on three
factors: (1) hazardous constitutent quantity, (2) site
wastestream quantity, (3) site disposal capacity. The
hazardous constituent quantity factor represents the actual
quantity of hazardous substances deposited on the site. The
wastestream quantity factor represents the quantity of
hazardous substances potentially deposited on the site based
on available information about the nature and quantity of
-------
wastes that were deposited on the site. The site disposal
capacity factor represents the quantity of hazardous
substances potentially deposited on the site based on the
available information about the quantity of hazardous
substances the site could have received as indicated by the
sizes of the sources identified on the site. These three
factors are in turn evaluated based on some or all of the
following waste quantity measures depending on the quality
and completeness of supporting data: (1) the quantity of
hazardous substances deposited, (2) the quantity of waste
deposited that contain hazardous substances, (3) source
volume, (4) source area. The actual approach used in
evaluating the hazardous waste quantity factor in terms of
the three subsidiary factors and the four waste quantity
measures is complex. The approach reflects judgments as to
the appropriateness of employing each measure and factor
under varying site-specific conditions of data completeness
and quality. The hazardous waste quantity factor evaluation
is presented in Section 2.2.2 of the Proposed Rule; the basis
for the approach is presented in Section 2.4 of the Technical
Support Document.

EPA has concluded that this tiered approach will make
the hazardous waste quantity factor more accurate by using
the best available data without imposing significant new
costs or demands on resources. The tiered approach would
also allow the scoring of the hazardous waste quantity factor
at many more sites. In addition, the flexibility of the
approach would accommodate a wide variety of data gathering
strategies; efforts could be varied so that more resources
were devoted to complex sites or to sites suspected of presenting
severe health risks.

The proposed revision to the hazardous waste quantity
factor is also responsive to the SARA requirement (CERCLA
Section 125, added by SARA) to consider the quantity, toxicity,
concentration of hazardous substances at facilities that
contain substantial volumes of waste described in Section
3001 (b)(3)(A)(i) of RCRA. The wastes include fly ash
wastes and other wastes generated from combustion of coal or
other fossil fuels. For a further discussion of these wastes,
see Section V E.

The hazardous waste quantity factor in the onsite pathway
would be calculated differently (Proposed Rule, Section 5.2.1.1).
In the other pathways, the hazardous waste quantity factor
reflects the magnitude and duration of potential releases.
In the onsite pathway, the question is not the release and
migration of the hazardous substances, but rather the potential
for direct contact with the contaminated area. For this
reason, the hazardous quantity factor for the onsite pathway
would be based primarily on the total surface area of the
known sources at the site. If the original source of the
contamination is unknown, the waste quantity factor would be
based on the area of soil contaminated at levels significantly
-------
-33-

above background.

4. Observed Release/Concentration of Hazardous substances
in the Environment

The current HRS scores an observed release if the
measured concentration of the hazardous substance is
significantly above the background level and if that
concentration can reasonably be attributed to the site.
The current HRS gives little consideration to the specific
concentrations of hazardous substances in the environment nor
does it consider the relationship between the concentrations
and health standards.

Some commenters to previous rulemakings have stated that
only concentrations exceeding health-based benchmarks
should be used in determining whether an observed release has
occurred. Other commenters have stated that if a substance
is detected at a level above a health-based benchmark it
should receive a higher score than if the benchmark is not
exceeded. Incorporating an assessment of environmental
concentrations into the HRS would, in the opinion of some
commenters, more accurately define the nature and degree of
potential risk.

EPA evaluated several approaches for directly using
environmental concentration data to assess potential risk as
part of the HRS. One approach would have based the observed
release value on the highest measured concentrations of
hazardous substances in the environment, using this level as
a measure of potential exposure. The Agency, however, concluded
that site inspection personnel would generally be unable to
identify areas where maximum contamination could be found.
The temporal and spatial variance of contaminants makes it
difficult to identify the most contaminated location on a
site during a limited investigation. Furthermore, the maximum
concentrations may not yet have occurred at the time of the
sampling.

Another approach studied was the use of simple fate and
transport models to predict concentrations at the receptor.
Fate and transport models combine movement in the transport
medium (e.g., water) with the fate of the substance in the
medium. For example, one substance may immediately dissolve
in water and persist indefinitely while another may precipitate
out of water and biodegrade within hours. However, using fate
and transport models in this manner would require the accurate
characterization of the source of contaminants and the media
through which the contaminants must travel. Given the
limited scope of the site inspection, it would be difficult
to characterize the source and the media adequately enough to
predict concentration levels accurately.

Knowledge of the concentrations and release rates of
constituents at the source is essential to predicting levels
-------
at more distant points. As discussed in the section on waste
quantity, EPA has concluded that at most sites it cannot
dependably characterize the constituents of a waste source
during the site inspection. Therefore, the Agency is not
proposing to include a factor that would be scaled according
to measured concentrations in releases.

EPA is retaining the current approach to scoring observed
releases, but is proposing to specify better, more precisely
defined criteria for determining when a release is significantly
above background (Proposed Rule, Sections 2.1.1, 3.1.1, 4.1.1.1,
and 5.0.1). The proposed criteria for significant releases
are as follows:

o If no background concentration is detected, a
significant release is three or more times the
detection limit.

o If the background concentration is greater than or
equal to the detection limit but less than two times
that limit, a significant release is greater than or
equal to three times the applicable background concentration
or greater than or equal to four times the detection
limit, whichever is less.

o If the background concentration is greater than or equal
to twice the detection limit, a significant release is
greater than or equal to twice the applicable background
concentration.

The detection limit could be the minimum of the EPA contract-
required quantitation or detection limit specified in EPA's
Contract Laboratory Program, the method detection limit for a
given analytical procedure or instrument (or in the case of
real time field instruments, the detection limit of the instrument
used in the field), or the actual detection limit achieved by
the laboratory for the set of samples in question. Negative
sampling results would not necessarily form the basis for
refuting an observed release that is based on a separate valid
sampling and analysis because releases may be episodic in nature.

As mentioned above, commenters have suggested that a
release at a concentration below a known health-based benchmark
should not be considered significantly above background. When
the current HRS was proposed, EPA explained that finding an
observed release indicates that the likelihood of a release
is 100 percent. The release of some substances into the
environment is a good indication that substances from the
site can escape and increases the likelihood of subsequent
releases. Data on frequency and quantity of actual releases
would require long-term monitoring, which is not feasible at
the site inspection stage. In addition, the results of limited
sampling may not be representative; higher concentrations
than those detected when the sampling was done may exist or
may occur at other times. EPA has concluded that the proposed
-------
-35-
criteria provide a reasonable definition of an observed release.

EPA has also concluded that limited environmental concen-
tration data cannot be used to demonstrate that concentrations
will remain below health-based benchmarks. Such an approach
would most likely lead to the omission of some high risk
sites from the NPL.

EPA is, however, proposing to use environmental concentra-
tion data in evaluating and scoring target populations as
discussed below in Section V C 5. This scoring method would
weight any population actually exposed to documented contamination
more heavily than those potentially exposed.

5. Use of Health-Based Benchmarks in Evaluating Target
Populations

In assessing target populations in the current HRS,
people actually exposed to contamination do not count more
than people potentially exposed, nor is the level of exposure
considered. Under Section 118, added by SARA, EPA is required
to give high priority to sites that have led to the closing
of drinking water wells or the contamination of principal
drinking water supplies. To respond to this mandate, EPA
considered an option of weighting closed wells higher than
operating wells. Such a factor, however, would create other
problems, such as how to weight contaminated wells that should
be closed, but are not, and wells that may be closed in the
future. Instead of including closed wells, EPA decided to
give greater weight to known exposures using two other mechanisms.
First, as discussed in Section V C 9, factors reflecting
risks to the MEI would be added to the revised HRS. Second,
populations whose wells or intakes show documented drinking
water contamination would receive higher weightings than
those of populations only potentially exposed.

To improve the accuracy of the scoring system by giving
increased weighting to populations based on their actual
exposure, the Agency is proposing to expand the evaluation of
exposed populations in both the ground water and surface
water pathways to include weighting factors based on health-based
benchmarks.

However, even though the Agency is attempting to consider con-
centrations of contaminants in drinking water in the proposed
revisions to the HRS, it is important to remember that these data
are from limited site investigations, and are used, in the HRS,
simply to make initial screening decisions. Health-based
benchmarks and cancer risk numbers are not used in the HRS to
identify levels of risk from drinking contaminated water, but
rather to provide added weight to populations actually exposed
to site contaminants in determining pathway scores.

For the ground water and surface water pathways, the
health-based benchmarks that EPA is proposing to use would be
-------
-jo —
the Federal primary drinking water standards (maximum contaminant
levels or MCLs) proposed or promulgated under the Safe Drinking
Water Act (SDWA). Where no MCL has been proposed or finalized
for a substance, the health-based benchmark for noncarcinogens
would be the proposed maximum contaminant level goal (MCLG)3;
for carcinogens or potential carcinogens, where the proposed
MCLG has been set at zero, a concentration corresponding to
specified individual lifetime cancer risks would be used as
the health-based benchmarks.

MCLGs are health-based levels at which no adverse health
effects would arise with a margin of safety. They are not
enforceable under the SDWA. MCLs are enforceable limits
under the SDWA and are set as close to the MCLGs as possible,
taking several factors into account, including the effective-
ness of treatment by the best available technologies, detect-
ability, and practical quantitation limits. For known or
probable human carcinogens, MCLGs are set at zero. MCLs for
carcinogens will never be set at zero, but are expected to be
set so that the risk of drinking water at the MCL falls within
the range of 10"4 to 10~7 individual lifetime cancer risk.
These cancer risk numbers assume that the individual is a
70 kg (150 Ib) person consuming 2 liters of water a day for
70 years.

Using these health-based benchmarks in assessing drinking
water threats, the target population factor in the surface
water and ground water pathways would be divided into four
population groups:

(1) Level l: The population drinking from wells or
intakes that are contaminated with hazardous sub-
stances at concentrations greater than:

- proposed or final MCLs;

- proposed MCLGs (for noncarcinogens with no
proposed or final MCL); or

- a 10"4 individual lifetime cancer risk (for
carcinogens and potential carcinogens with no
proposed or final MCL).

The population drinking water at these concentrations
would be weighted 100 times as much as the popula-
tion drinking water at Level 3 concentrations.

(2) Level 2: The population drinking water from wells or
intakes that are contaminated with hazardous sub-
stances at concentrations siginificantly above back-
ground, but within the following ranges:

3 This proposed rule discusses "proposed" MCLGs only because
MCLGs and MCLs will be finalized concurrently and, for the purposes
of the proposed HRS, the final MCL will supercede the final MCLG.
-------
-37-
- greater than 1/1000 of the proposed or final MCL
but less than or equal to the proposed or final
MCL;

- greater than 1/1000 of the proposed MCLG but
less than or equal to the proposed MCLG (for
noncarcinogens with no proposed or final MCL); or

- greater than 10~7 but less than or equal to
10~4 individual lifetime cancer risk (for
carcinogens or potential carcinogens with no
proposed or final MCL).

The population drinking water at these concentra-
tions would be weighted 10 times as much as the
population drinking water at Level 3 concentrations.
Level 2 would also include any hazardous substance
that shows up in a drinking water well or intake
at concentrations that are significantly above
background, but has no proposed or final MCL or
MCLG, or cancer risk number.

(3) Level 3: The population drinking from wells or
intakes that are contaminated with hazardous sub-
stances at concentrations significantly above
background, but less than or equal to:

- 1/1000 of the proposed or final MCL;

- 1/1000 of the proposed MCLG (for non-
carcinogens with no proposed or final MCL); or

- 10~7 individual lifetime cancer risk (for
carcinogens with no proposed or final MCL).

These populations would not be given any additional
weight.

(4) Potential contamination: The population whose wells
or intakes are not known to be contaminated, but the
ground water or surface water is already contaminated
or has the potential to be contaminated. Only
populations who use drinking water from within the
target distance limit would be counted as potentially
exposed. This potentially exposed population would
be distance-weighted in the ground water pathway and
dilution-weighted in the surface water pathway.

In all groups, the population counted would be the people
whose drinking water is drawn from wells or intakes within the
target distance limit; the populations would not have to live
or work within the target distance limit.
-------
-JB-

The weighting of these groups was chosen to give high
priority to sites where exposures to contamination attributable
to the site were known to be occurring. Where Level I concen-
trations exist, only 200 people would need to be exposed to
the contamination for the population factor to be assigned
the maximum score.

The specific health-based benchmarks used in determining
this factor were chosen to be consistent with other Agency
programs. The range of individual lifetime cancer risks of
between 10"4 and 10~7 was chosen to be consistent with the
approach currently taken by the Superfund program in determining
cleanup levels. For noncarcinogens, a three order of magnitude
range of concentrations would be used in the Level 2 concen-
trations group to be more consistent with the way carcinogens
are treated in Level 2. EPA has chosen to count populations
drinking from wells or intakes contaminated at concentrations
significantly above background but at concentrations less
than or equal to the MCL or MCLG (or with an individual lifetime
cancer risk of less than or equal to 10"') because some
contamination, albeit low, has been detected, and may be the
leading edge of a contaminant plume. Moreover, the concentration
found in a well or intake during a site inspection is only a
one-time picture of the contamination. During a remedial
investigation/feasibility study, the Agency does extensive
sampling to determine the extent of the contamination at a site.

MCLs have been developed for relatively few hazardous
substances. More hazardous substances have MCLGs and cancer
risk numbers. However, of the hundreds of hazardous substances
found at Superfund sites, most currently have no health-based
benchmarks. To be protective, the Agency has decided that
populations exposed to hazardous substances with no health-based
benchmarks should be included in Level 2 if the substances
are found in the drinking water at concentrations significantly
above background and are attributable to the site. The Agency
solicits comments on this approach.

If more than one substance is present in the drinking water
at levels significantly above background but not above the MCL,
MCLG, or 10"* individual lifetime cancer risk, for each such
substance, the percentage of its health-based benchmark at
which it is present would be calculated. If the total sum of
the percentages exceeds 100, the concentration in the drinking
water would be considered Level 1 concentrations (Proposed
Rule, Sections 3.3.2.1. and 4.1.3.2.1) and the population
using that contaminated water would be weighted as Level 1
concentrations. If the sum of the percentages is greater than
0.1 and less than or equal to 100, the population using that
contaminated water would be weighted as Level 2 concentrations.
Finally, if the sum of the percentages is less than or equal
to 0.1, the population using that contaminated water would be
considered as a Level III population.

This proposal for summing benchmark ratios would give
-------
-39-

higher scores to those sites where several hazardous substances
are found in the drinking water at concentrations near their
benchmarks. The Agency analyzed other methods for achieving
this, including evaluating hazardous substances individually
against their benchmarks and basing the determination of target
population weighting levels on the single substance found in
the drinking water with the highest benchmark ratio. The Agency
would like comment on the appropriateness of both these approaches
to scoring multiple substances detected in drinking water, as well
as any alternative approaches.

The Agency solicits comment on the proposed population
groups, on the weightings assigned to these groups, and on
the health-based benchmarks used to define these groups. In
addition, the Agency solicits comment on the risk range (10~4
to 10"') used to define the proposed levels. Specifically,
EPA would like comment on whether a risk range should be used
in the context of the HRS and, if so, whether it should be
10"4 to 10~7 (as is used in the proposed rule) or 10~4 to
10"6.

EPA is proposing the system described above for considering
health-based benchmarks in the HRS,* however, the Agency has
evaluated a variety of alternatives. Three are discussed in
this preamble. (See the Technical Support Document, available
in the Superfund docket, for further detail.)

Alternative I is a three-tiered system that would use
only proposed or final MCLs as the health-based benchmarks.
The populations drinking water contaminated with hazardous
substances at concentrations that exceed a proposed or final
MCL would be included in the first tier, which would be
weighted 100 times as much as the potentially exposed pop-
ulation. The population in the second tier would include
those people drinking water at concentrations at or below a
proposed or final MCL, but significantly above background,
and those people drinking water containing hazardous substances
for which proposed or final MCLs do not exist; this population
would be weighted 10 times as much as the potentially exposed
population. The third tier would include those people potentially
exposed to contamination; this population would be distance- or
dilution-weighted.

Alternative I would be considerably simpler than the one
proposed and would be based solely on proposed or final MCLs,
health-based benchmarks developed by the Agency for use in
evaluating drinking water contamination, rather than MCLGs or
cancer risk numbers. However, since MCLs have been developed
for only a handful of chemicals, the Agency believes this
approach may not accurately assess the relative degree of hazard
and therefore does not discriminate among sites as well as
the proposed approach.

Alternative 2 would weight target populations using bench-
marks similar to those used in the proposed option — MCLs,
-------
MCLGs, and cancer risk numbers. However, Alternative 2 con-
tains three tiers instead of four, and uses a single cancer
risk number of 10~6 rather than a risk range to delineate
the different tiers. The target population would be divided
into the following three tiers:

(1) Level 1: The population exposed to concentrations
greater than the proposed or final MCLs, the proposed
MCLGs (for noncarcinogens with no MCLs), or a 10~6
individual lifetime cancer risk (for carcinogens with
no MCLs) would be weighted 100 times as much as the
people potentially exposed.

(2) Level 2: The population exposed to concentrations less
than or equal to the proposed or final MCLs, the
proposed MCLGs (for noncarcinogens with no proposed
or final MCL), or a 10~6 individual lifetime cancer
risk (for carcinogens with no MCLs), but significantly
above background, would be weighted 10 times as much
as the people potentially exposed.

(3) Potential Contamination: The population whose wells
or intakes are not known to be contaminated, but the
ground water or surface water from which these wells
draw is already contaminated or has the potential to
be contaminated. This potentially exposed population
would be distance-weighted in the ground water pathway
and dilution-weighted in the surface water pathway.

In all groups, the people counted would be those whose
drinking water is drawn from wells or intakes within the
target distance limit; the people would not have to live or
work within the target distance limit.

The evaluation of multiple substances under this alterna-
tive would be similar to the proposed approach, in that,
for each substance the percent of its health-based benchmark
at which it is present would be calculated. As in the proposed
option, if the total sum of the percentages is greater than 100,
the concentration would be considered to be above the benchmark
and the populations exposed to the contaminated water would
be weighted as Level 1 contamination. If the concentration
is significantly above background but not above the health-based
benchmarks, the people exposed to the contaminated water
would be weighted as Level 2 contamination.

The weighting of the three groups was considered so that
sites where exposures to contamination above health-based
benchmarks were known to be occurring would be given high
priority. This approach is simpler than the proposed approach,
and may be appropriate given the limited data available at
the site inspection stage, and the purpose of the HRS as a
screening tool (rather than a risk assessment).

The third alternative would be identical to Alternative
-------
2, except a 10~4 individual lifetime cancer risk level,
instead of a 10~6 risk level, would be used to differentiate
between Level 1 and Level 2 contamination.

The Agency requests comment on these alternative approaches
to assessing actual drinking water exposures in the HRS.

EPA is not proposing at this time to incorporate health-based
benchmarks for the air and onsite pathways due to a number of
unresolved technical issues. These include selection of the
point at which samples would be taken to demonstrate exposure,
the length of the monitoring period, and interpretation of
data in light of the extreme temporal and spatial variability
of air releases. EPA is considering ways to evaluate observed
human exposure in the air pathway and solicits comments on
how these technical issues could be addressed. It should be
noted that the onsite pathway would be evaluated only if
there is documented contamination at the site. In addition,
the varying exposure to the non-resident population would
depend more on the frequency of contact than on the level of
contamination. For this reason, the nearby population would
be distance-weighted.

The surface water pathway would apply ecologically-based
benchmarks to the sensitive environment targets factor as well
(Proposed Rule, Sections 4.4.3.1.1, 4.4.3.1.2, and 4.4.3.1.3).
This would be similar to the health-based benchmark system in
terms of the relative weights or groups, except that there
would be three levels instead of four. The ecological benchmarks
are based on ambient water quality criteria. In the human
food chain subpathway and the recreation subpathway, actual
observed contamination would also be weighted more heavily
than potential contamination.

6. Consideration of Hazards to the Environment

CERCLA Section 105 required EPA to create an NPL of at least
400 sites. At the time of enactment, EPA realized that
thousands of sites posed potential public health and environmental
threats. The Agency believed that, given the need to set
priorities for the expenditure of limited monies, the HRS
should place greater weight on sites that posed threats to
public health rather than those that posed risks to the
environment.

CERCLA Section 105 mandated that EPA consider both threats
to public health and to the environment when it developed the
NPL. Although SARA did not specifically require EPA to take
any particular action regarding threats to the environment,
consideration of the impact that waste sites have on sensitive
environments is emphasized in SARA's addition of Section
105(c)(l), which requires the HRS "to the maximum extent
feasible to accurately assess the relative risk to human
health and the environment." EPA's experience with many
potential Superfund sites suggests that a number of sites
-------
posing a serious threat to the environment are not scoring
high enough to be placed on the NPL and addressed under CERCLA.
Therefore, the Agency has determined that overall accuracy
would be improved, in certain cases, by placing on the NPL
sites that have significant impact on the environment, even
when the sites pose less of a threat to human health.

Although the proposed HRS was designed to give greater
weight to environmental impacts than the current HRS,
the scores related to the relative risks to human health
would still be weighted more heavily than sensitive environments.
If sensitive environments, but no people, are affected by a
site, the score would not reach the maximum for the targets
category. EPA does, however, intend to assign a sufficiently
high value to sensitive environments so that the most serious
environmental impacts in the absence of any public health
risks would have scores above the NPL cutoff. EPA would like
comments on the relative weightings of the two types of impacts.

EPA is proposing to modify several features of the current
sensitive environment factors. In the current HRS, if more
than one sensitive ecosystem exists within the target distance,
only the one with the highest score is included. The proposed
HRS would base the sensitive environment factor in the surface
water and air pathways on the sum of the values for all appropriate
ecosystems within the target distance, with each ecosystem
generally weighted for distance or dilution (Proposed Rule,
Sections 2.3.4 and 4.4.3). This proposed change is intended
to provide a more accurate assessment of the potential risk
to sensitive environments.

The primary ecosystems considered in the current HRS are
wetlands of greater than five acres and habitats of Federally
designated endangered species. The land use factor in the
air pathway also evaluates other ecosystems, such as wildlife
reserves. The revised HRS expands significantly the list of
sensitive environments to include those environments that are
protected under Federal or State designations. The full list
of the proposed sensitive environments is Table 2-18 of the
Proposed Rule. These environments include marine sanctuaries,
National Parks, designated Wilderness Areas, National Monuments,
National Seashore Recreational Areas, and National or State
Wildlife Refuges. The final item on the list covers particular
areas, often relatively small in size, that are important to
the maintenance of unique biotic communities 1 EPA wants to
specify the particular communities in the final rule and
requests comments on which such communities should be listed.

In revising the sensitive environments factor, the Agency
evaluated several ecological ranking models that were either
in use or were in the latter stages of development. From this
work, and from comments received in response to the ANPRM,
the Nature Conservancy's National Heritage Program was identified
as having the potential to supplement the sensitive environments
list in the revised HRS. When the National Heritage Program
-------
-43-

is used in conjunction with the expanded list, EPA will be
able to identify not only sensitive environments that have
been formally designated by State or Federal agencies, but
also those environments that score high on the National Heritage
Program's database, which ranks a site according to its rarity
and vulnerability.

The proposed sensitive environments factors could be
assigned values based on their inclusion on the list of
protected areas or on their rating under the National Heritage
Program (see Table 2-19 of the Proposed Rule). If the values
assigned for the type of area and for its National Heritage
Program ranking differ, the higher of the values would be used.

Sensitive environments considered in the air pathway
would be distance-weighted so that those closest to the site
would have higher values than those at increasing distances up
to the target distance limit. In the surface water pathway,
sensitive environments that have been subject to actual
contamination would receive a score based on whether the
contamination was above or not exceeding ecological benchmarks.
Where no actual contamination has been documented, the scores
assigned would be dilution-weighted. In the onsite pathway,
only terrestrial sensitive environments with observed contamination
would be considered since the exposure is presumed to be from
direct contact with the hazardous substances, not from the
migration of hazardous substances.

7. Consideration of Effects on the Human Food Chain

SARA (see CERCLA Section 105(a)(8)(A), as amended)
requires EPA to consider, in revising the HRS, the effects of
hazardous waste sites on the human food chain. When EPA
developed the current HRS, the Agency decided that it could
not apportion human food chain effects according to population
risks, but it did deal with the effects qualitatively in the
land and water use factors. In addition, the risks to the
human food chain from using potentially contaminated water
for irrigation are reflected in the current procedure for
estimating the target population; an additional 1.5 people
per acre are added to the population total for each acre of
irrigated food or forage crops.

In revising the HRS, EPA has determined that the most
significant, measurable human food chain risks are those
associated with the aquatic food chain. Therefore, in the
HRS surface water pathway, EPA is proposing to evaluate the
potential risk to the human food chain based on potential or
observed contamination of aquatic food chain organisms. Details
of the proposed method the Agency would use to incorporate
human food chain effects in the scoring of surface water are
discussed in the section on the surface water pathway (Section
V D 2) .

The potential exists for some sites to adversely affect
-------
the human food chain via other pathways. These pathways
would primarily affect the terrestrial food chain (deposited
air pollutants migrating to the edible portion of plants,
ground water or surface water used to water animals or irrigate
crops). EPA considered ways to account for terrestrial food
chain effects in the revised HRS. However, the Agency decided
that because the terrestrial food chain was more complex and
not as well understood as the aquatic food chain, it was
impractical to include a detailed assessment in the HRS.
Rather, points are assigned to the ground water and surface
water use factors if the water is used for irrigation of
commercial food or forage crops, for commercial livestock
watering, or for commercial food preparation. In the ANPRM,
EPA asked for comments on methods of incorporating human food
chain effects into the HRS and is continuing to seek comments
on whether food chain contamination by hazardous substances
in air and soil should be included in scoring either the air
or ground water pathway and if so, the basis for estimating
human health risks from such food chain exposure.

8. Dilution/Distance Weighting of Targets

The current HRS directly weights the population factor
by dilution/distance only in the air pathway. The ground
water pathway combines the total population using water drawn
from the area within the target distance limit in a matrix
with distance to the nearest drinking water well. The surface
water pathway uses a matrix to combine the distance to an
intake and the population using that intake. The greater the
distance, the lower the HRS value for any population category.

In reviewing ways to account for the greater risks to
populations close to sites, EPA considered using analytical
models that would require data such as wind speed and
temperature in the air pathway to calculate the rate of
dispersion for a particular substance at a specific site and
data such as ground water flow direction and gradients and
dispersion to calculate dilution for the ground water pathway.
EPA decided, however, that the reliability of the results,
given the limited data available from site inspections, was
not great enough to consider using such models and equations
for each site.

The revised HRS would use distance- and dilution-weighting
factors in calculating the scores for certain populations and
environments that are potentially exposed to contamination
from sites. The weighting factors would reflect the diminishing
risk as substances disperse or dilute and were generally
developed using analytical models. For each prescribed distance
(e.g., a quarter to a half mile from the site) or for the
appropriate flow characteristics of the surface water, the
potentially affected population in the area would be multiplied
by a weighting factor. The target population would be the
sum (subject to the maximum) of the distance or dilution
weighted groups plus any populations exposed to documented
contamination.
-------
Although EPA did not design the revised HRS to be an
analytical model, the Agency did use models to help develop
the scales of values and weights for distance weighting.
These models are more fully explained in the Technical Support
Document, available in the Superfund docket. EPA concluded that
this distance and dilution weighting approach uses the best
elements of analytical models without requiring site-specific
data and thus represents a significant increase in accuracy
without a major increase in data collection costs. In reviewing
the effects of air emissions on populations surrounding Superfund
sites, the Science Advisory Board supported the weighting of
population according to distance from a site.

The proposed distance weighting factors for ground water
(Proposed Rule, Sections 3.3.1 and 3.3.2) are derived from a
three-dimensional fate and transport model that determines
relative concentrations as a function of distance from a
site. Those relative concentrations provide the basis for
the weighting factors. An exception to the distance weighting
would occur when the aquifer is a karst aquifer. (See Section
V D 1 and Proposed Rule, Section 3.3.2 for a discussion of
karst aquifers.)

The air pathway distance weighting factors are based on
the effects of atmospheric diffusion and were calculated
using a simple Gaussian plume model (Proposed Rule, Sections
2.3.1 and 2.3.2). The surface water pathway would not use
distance weighting, but would instead employ dilution weighting.
The extent of the dilution would be considered a function of
the flow characteristics of the water available for dilution
(Proposed Rule, Sections 4.1.3.1 and 4.1.3.2).

For the onsite pathway, EPA is proposing a distance
weighting factor for the nearby population that lives within
one mile travel distance of the area of contamination (Proposed
Rule, Section 5.2.3), but does not live where contamination
is present. This factor would be set based on the relative
frequency with which an individual could travel to the site,
which, in turn, is assumed to be based on the distance between
the person's residence and the site. EPA has not identified
any studies that provide estimates of incursion rates into
contaminated land areas or the relationships between frequency
of incursions and distance from a site. EPA is proposing a
factor based on distance and solicits comments on how frequency
of incursion might be taken into account in the onsite pathway.

9. Population Risks and Risk to the Maximally Exposed Individual

Maximally exposed individuals (MEIs) are those individuals
in the exposed population that are expected to be exposed to
the highest ambient concentration (and thus receive the highest
dose) of the hazardous substance in question. Population
risk would be the effect on the exposed population over an
extended period of time, usually assumed to be 70 years.
-------
-46-

The current HRS incorporates both concepts in developing
target population scores. For example, the total population
is evaluated at distances around a site, and the population
scores are either distance weighted (air) or combined in a
matrix with distance to the nearest well (ground water) or
distance to the intake (surface water).

Although the HRS was not designed to be a risk assessment,
the Agency believed that an explicit factor based on potential
MEI risks should be added to improve the overall assessment
of potential risks to human health within the HRS and to make
the revised HRS more consistent with general Agency risk
assessment approaches. Usually/ EPA evaluates both MEI risks
and exposed populations as part of its risk assessments to
provide a better overall indication of potential threats.
Consequently, several proposed-changes related to MEI and
population risks are included in the revised HRS. Population
scores would be weighted based on known or potential exposure
to contaminants (see health-based benchmarks in Section V C
5). Factors reflecting the risks to the MEI via the ground
water, surface water, and air pathway would be included in
the revised HRS.

For ground water, the MEI risk would be assessed through a
factor based on the distance to the nearest well (Proposed
Rule, Section 3.3.1). This measure was chosen because it is
likely that, all other things being equal, the well closest
to the site would have the highest level of contamination.
Since contamination usually decreases with distance, the
farther the nearest well is from the site, the lower the
assigned value would be, with three exceptions. First, if any
well has documented contamination above health-based benchmarks
attributable to the site, the MEI factor would be assigned
the maximum value. Second, if the site overlies a karst
aquifer, the MEI factor would be assigned the maximum value if
any well draws drinking water from the karst aquifer within the
target distance limite. This reflects the potentially shorter
travel time within such aquifers. Third, different distance
weighting factors would be applied to wells in karst and
wells not in karst to reflect differences in dilution.

For the surface water pathway, the risk to the MEI from
drinking water would be represented by a value based on the
flow characteristics of the body of water at the nearest
intake (Proposed Rule, Section 4.1.3.1). This method was
selected because the flow characteristics of surface water
are a major factor in determining the concentration of contaminants;
i.e., the greater the volume of water, the greater the dilution,
and therefore the lower the potential risk. The assigned
value for the MEI factor would be a multiple of the dilution
weighting factors for the different flow characteristics of
surface water. If any drinking water intake has documented
contamination above health-based benchmarks attributable to
the site, the factor would be assigned the maximum value.
The human food chain, recreation and environmental subpathways
do not contain an MEI factor.
-------
-47-

For the air pathway, the risk to the MEI would be based
on the distance from the emission sources on the site to the
nearest individual, using the distance to the closest residence
or regularly occupied building or area (Proposed Rule, Section
2.3.1). Values would be assigned for distances of up to two
miles from the site. Values beyond two miles would be zero due
to the distance weights beyond two miles.

Because the onsite pathway is not a migration pathway
and because of the nature of the exposure, the onsite pathway
does not lend itself to an MEI factor. EPA is seeking
comments on whether such a factor should be incorporated and
if so, how to do so.

10. Scoring on the Basis of Current Conditions

Under the current HRS, EPA generally scores the air,
ground water, and surface water pathways based on the present
condition of the site excluding any response action that has
been taken ("initial conditions"), rather than on the present
condition of the site, taking into account response action
("current conditions"). The Agency has used this approach
for a number of technical and programmatic reasons explained
at 47 FR 31187 (July 16, 1982), and discussed below. In
conjunction with revising the HRS, the Agency decided to review
this policy. The Agency believes it may, in some situations,
be appropriate to evaluate the site based on current conditions
and to consider prior responses in calculating a HRS score.
The Agency intends to determine under what conditions prior
response actions should or should not be considered, to ensure
that this results in a more feasible and accurate assessment
of potential risk to human health and environment. EPA is,
therefore, requesting comment on the following issues and
approaches under consideration by the Agency.

CERCLA Section 105 (c) , as added by SARA, requires EPA to
amend the HRS, and states that, "Such amendments shall assure,
to the maximum extent feasible, that the HRS accurately assesses
the relative degree of risk to human health and the environment
posed by sites and facilities subject to review." The Agency
believes that, to the extent that risks at a site are reduced
due to response actions, it may be appropriate to base the
HRS score on that reduced risk. In addition, EPA believes
that if properly devised, such a policy may encourage a bias
toward action that protects human health and environment
without distorting the HRS1 ability to assess the relative
risk of sites. Furthermore, the Agency believes that the
current, rather than initial, conditions generally may represent
a more accurate basis of scoring. Nonetheless, EPA recognizes
that certain situations exist where implementing this policy
may not be practical or technically feasible.
existing policy of evaluating sites based on initial
conditions was based on three principal concerns associated
with considering current conditions. The first concern was
-------
-48-

that including consideration of current conditions would
create undesirable incentives at hazardous waste sites that
may be eligible for the NPL. For instance, some private
parties may only take action sufficient to lower the score so
the site would not be listed on the NPL, but the site could
still pose a potential threat to public health or the environment.
Those types of score changes could be accomplished by such
actions as removing wells from service to lower target scores,
or by removing wastes from a site to lower waste quantity
scores; however, in both cases contaminated ground water
would still exist at the site.

Another undesirable incentive may be to cause public
agencies to be reluctant to perform removals if such actions
could lower the score and thereby prevent the site from being
included on the NPL. Only sites listed on the NPL are eligible
for remedial action using Superfund monies. These early
response actions are important to address immediate problems
posed by the site. The Agency is concerned that if prior
removal actions were considered in the calculation of the HRS
score, public agencies may delay responding to threats to public
health and the environment in order to ensure listing on the
NPL, and the resulting availability of long term remedial
response funding under Superfund.

The second issue of concern was that the ability of the
HRS to approximate risk at a given site is based on a number
of presumed relationships between various factors considered
in calculating the HRS score. When partial response actions
are taken into account in site scoring, the validity of these
relationships for the purpose of approximating risk posed by
the site might be affected. For example, the hazardous waste
quantity factor, in combination with toxicity and likelihood
of release, helps predict the relative risk of a given release.
For a site that has been in existence for some time, hazardous
substances may have migrated to the ground water or surface
water. If the hazardous materials on the surface are removed
and the site is scored according to conditions existing after
the removal ("current11 conditions), the site could be assigned
a low value for waste quantity, even though an unknown quantity
of the hazardous material may be in the soil on the site and
remain a potential threat to public health or the environment
via the ground water, surface water, or air pathways. Thus,
EPA was concerned that if a site were scored to reflect conditions
after a response, the expected reduction in the HRS score
based on current conditions might not reflect a commensurate
reduction in the level of risk presented by the site, because
the Agency may not be able to determine, at the time of site
inspection, the extent of contamination that has occurred.

Finally, the Agency considered the programmatic issue of
how to define "current conditions" when conditions may be
changing between the time of initial data collection and final
listing. Response actions often are ongoing at sites during
-------
-49-

the evaluation process, and it would be unduly burdensome to
continually recalculate scores to reflect such actions.

The Agency is considering two approaches to incorporate
current site conditions in the HRS score while minimizing the
concerns discussed above. Under either approach, EPA would
only consider removals prior to a site inspection, as
EPA cannot continuously update the score of a site to
reflect ongoing cleanup activity. The first approach involves
consideration of removal actions for certain pathways or
specific factors where appropriate. The second approach is
to consider current conditions routinely, but to identify and
exempt situations where current conditions will not lead to
a more accurate assessment of risks.

Under the first approach, EPA would identify for each
pathway (i.e., ground water, surface water, air and onsite)
and for non-target factors (e.g., likelihood of release and
waste characteristics) and target factors (e.g., population
and distance to nearest well) whether scoring based on current
conditions is appropriate. In scoring sites using the current
HRS, the migration pathways have generally been scored on the
basis of initial conditions for non-target factors and current
site conditions for target factors. However, there are some
exceptions. For example, targets have not been scored on
current conditions where a site has contaminated soils of a
residential area such that it is advisable for people to
relocate.

Under this first approach, the Agency would score all factors
for the onsite pathway based on current conditions at the time
of the site inspection or equivalent because potential
exposure in that pathway is based on direct contact with
contaminated materials (Proposed rule, Section 5.0.1). The
Agency believes that this is a better approach for measuring
relative risk under those circumstances, and generally it is
more feasible to determine whether a threat to health in the
onsite pathway has been addressed than in the other pathways.
For example, if the contaminated soils have been removed and
permanently disposed, the risks through direct contact are
probably no longer significant.

The Agency believes there may also be certain non-target
factors for which scoring on the basis of current conditions
could give a more accurate indication of risks at the site
e.g., flood potential and potential air releases. The flood
potential factor in the surface water pathway is evaluated
based on the site's location in a flood plain and on the
source's containment. If the site has not been flooded since
the disposal and if all wastes have been removed by an interim
action, the Agency believes this factor could be evaluated
based on current site conditions. In the potential air release
factor, for example, if a site had a surface impoundment with
volatile toxic substances, releases to air may occur. If,
however, prior to the site inspection, the impoundment is
-------
-50-

drained and any remaining sludge is adequately capped, the
threat of a release of volatiles is mitigated by the response
action, and the Agency believes this factor could also be
evaluated on current site conditions.

Unlike the situation in the air pathway, it may be much more
difficult to determine whether response actions have mitigated
the potential for a site to release to ground water or surface
water. For example, removal of a waste pile or draining a
surface impoundment stops adding to the source of contamination,
but does not ensure that other potential sources of contamination
(e.g., contaminated soils) or impacts (e.g., contaminated
ground water) have been addressed.

Generally, target factors (e.g., population and distance to
receptors), have been scored on initial conditions. For example,
when a temporary drinking water supply has been provided, the
initial target conditions may better represent the adverse
impacts caused by the site. The same would be true when
people have been relocated due to contamination, in which
case scoring on the basis of initial conditions may better
reflect the seriousness of the problem.

The Agency believes that the approach of taking response
actions into account for more HRS pathways and factors than in
the current HRS would provide a more accurate assessment of the
risks at sites. Such an approach would also provide incentives
for both public and private parties to perform responsible
response actions. For both target and non-target factors,
the Agency requests comment on additional factors that may be
appropriately scored on the basis of current conditions.

The second approach the Agency is considering is to
score all factors based on current conditions at the time of
the site inspection, except for situations where this is not
appropriate or feasible. For example, as just discussed,
if a temporary drinking water supply has been provided or
residents relocated, initial site conditions will be more
appropriate. Also, under this approach, removals performed by
non-Federal public agencies should not be considered, as
their actions are a recognition of the site's threat to public
health or the environment, and the public agency should not be
discouraged from taking early action when appropriate.

In consideration of the concern about accurately assessing
hazardous waste quantity when the vast majority has been
removed, EPA has identified two alternatives under this approach.
One method for making such determinations would be to require
additional soil and ground water samples if a removal has
occurred. This could add significantly to the cost of a
site inspection, as sampling and analysis tend to be the
highest cost components of performing the site inspection.
Alternatively, EPA could develop a factor to modify hazardous
waste quantity based on quantity removed and storage time at
the site. This would be less accurate, but because it would
-------
-si-

be an objective model, it could be more simply and consistently
applied, and would cost significantly less and be significantly
faster than site specific sampling. The Agency specifically
requests comment on these two approaches to assessing hazardous
waste quantity after a removal has occurred.

The Agency believes that this second approach to incor-
porating current conditions in the HRS score could be more
successful in assessing risk at the site than the first approach,
if the hazardous waste quantity issue is satisfactorily resolved.
This approach also maintains incentives for public and private
sector removals to the same extent as the first alternative.

EPA specifically seeks comment on the two approaches to
scoring sites, and on specific factors and situations that
should be evaluated on initial conditions. The Agency is
also interested in recommendations on other ways to consider
removal actions that would allow recognition of the site's
current conditions without encouraging incomplete solutions
that reduce the HRS score below the cutoff and possibly leave
significant health threats unaddressed, or significantly
affect the cost of performing a site inspection. Finally,
the Agency requests comment on how to evaluate the effectiveness
of a response action to account for wastes that may have
migrated to the soil, ground water, or surface water and may
be posing a potential threat to public health and the environment.

11. Low Density Populations

In the current HRS, the population close to a site must
be quite large (1000 to 10,000 people generally within one to
two miles) before the score for the targets category will
approach the maximum. Commenters on previous NPL rulemakings
have pointed out that this requirement for large populations
prevents some dangerous sites from being listed and dispro-
portionately affects certain groups. They have stated that
hazardous waste sites on or near Indian tribal lands or in
isolated, rural areas usually do not obtain high scores because
the population density is low.

EPA is proposing to establish high values in the revised
HRS for MEIs and for populations actually exposed to contamination,
especially where health-based benchmarks are exceeded in
drinking water. EPA believes that this will place greater
weight on dangerous sites in isolated or rural areas, and allow
such sites to be listed. EPA is seeking comments on this issue
and suggestions for other ways to consider it in the HRS.

12. Standby Wells and Surface Water Intakes

In the current HRS, EPA generally does not differentiate
between wells or intakes used as primary water sources and
those which are u-ad as standby water sources. Such an
approach tends to emphasize the value of the drinking water
resources, based on the rationale that standby water
-------
-52-

resources are often indispensable during periods of peak
demand or drought. However, treating standby intakes and
wells the same as primary water sources does not recognize
the significant difference in the use of standby and primary
intakes and wells, and the difference in potential risk based
on this difference in exposure. To improve the ability of
the HRS to distinguish between sites, EPA is proposing to
differentiate between primary and standby water sources.
EPA is considering three alternative approaches. One alternative
involves assigning values to a standby water source based on
the percentage of the system it supplies or the percentage of
the year the source is used. Such an approach could more
accurately assess the relative risks to public health from
chronic exposure by better evaluating the relative degree of
exposure through drinking water. A second alternative would
involve considering as targets, only wells or intakes that
are regularly maintained and are used more than some specified
annual amount. Such an approach would be simpler than the
above option, but it would not be as accurate. A third alter-
native approach is to give standby wells and intakes a fraction
of the value of a primary water source. This alternative
would not reflect exposure from a site, but would distinguish
between primary and standby sources and be easier to implement.
Aspects of the second and third alternative are incorporated
in the proposed rule in the MEI, population, and drinking water
use factors. EPA specifically requests comment on how standby
wells and intakes should be evaluated under the targets cate-
gory, and would like specific comments on the three alternatives
and what the particular cutoff levels or fraction should be.

D. Individual Pathway Revisions

Sections V D 1 through 4 detail the specific proposed
revisions to the ground water, surface water, air, and onsite
exposure pathways. A diagram that compares the current and
the proposed structure precedes each section, except for the
onsite exposure pathway where there is no current pathway for
comparison. Each of the pathway discussions is organized in
the same way: after a brief summary of the most significant
proposed revisions and a discussion of any general considerations
such as the distance over which risk is evaluated, the specific
proposed revisions are grouped within the three factor categories
— likelihood of release (likelihood of exposure for onsite
exposure), waste characteristics, and targets. Where the
revisions have already been specified in Section V C, those
discussions are referenced. Section 5 discusses the fire and
explosion pathway. As with the issues covered in Section V
C, a more detailed discussion of the options EPA considered
and the reasons for the proposed revisions can be found in
the Technical Support Document, available in the Superfund
docket.

1. Ground Water Pathway

As can be seen in Figure l, the revised ground water
-------
Ground Water Migration Pathway
Current HRS
Release -

Observed Release
or
Waste Characteristics
Route Characteristics
Depth to Aquifer of Concern
Net Precipitation
Permeability of Unsaturated Zone
Physical State
Containment
Hazardous Waste Quantity
ToxJcity/Pers/sfe/ice
Targets

3 Ground Water Use
3 Distance to Nearest
Wetffopulation Served
3
Q
3
Revised HRS
Likelihood of X
Release
Observed Release
or
Potential to Release
3 Depth to Aquifer/
HYDRAULIC
CONDUCTIVITY
3 Net Precipitation
Q SORPTIVE CAPACITY
3 Containment
Waste Characteristics X Targets
3 Hazardous Waste Quantity* 3
3 Toxicity/MOBILlTY 3
Ground Water Use*
Population*
MAXIMALLY EXPOSED
INDIVIDUAL
WELLHEAD
PROTECTION AREA
Items in italic under Current HRS have been dropped or replaced.
Items in caps under Revised HRS are new. Most items not in caps have been
revised significantly.
'Factor based on several sub-factors.
-------
-54-

migration pathway would retain the same structure as in the
current HRS. In both versions, the likelihood of ground water
contamination is evaluated by assessing the actual or potential
release of hazardous substances to aquifers. The likelihood
of release is then combined with the characteristics of the
hazardous wastes and with the targets to obtain a pathway
score. EPA is proposing revisions to every factor of the
ground water pathway, the most significant of which are in
the targets category. As already discussed, population would
be assessed by how far drinking water wells subject to potential
contamination are from a site (distance weighted) and by
whether people are drinking from a well with contaminants
above or below health-based standards. In addition, the
proposed revisions would change the distance (target distance
limit) within which drinking water wells are considered. In
the waste characteristics category, EPA is proposing to combine
toxicity with mobility rather than with persistence as is
done in the current HRS. A new factor, sorptive capacity,
would be added to the potential to release calculations.

General Considerations

Target Distance. EPA is proposing to extend the target
distance limit within which the target factors are evaluated
from three to four miles; where there is documented contamination
attributable to the site beyond the four-mile target distance
limit, wells with such contamination would be considered to
be within the target distance for the evaluation of all target
factors. In deciding on the change from three to four miles,
the Agency used a combination of empirical data and modeling
estimates. The empirical data consist of documented plume
distances of greater than three miles in length. The modeling
studies provided estimates of contaminant travel times for a
variety of hydrogeological conditions. The Technical Support
Document, available in the Superfund docket, contains further
information on these studies.

In addition to increasing the target distance limit, the
proposed HRS would change the locations from which the target
distance would be measured (Proposed Rule, Section 3.0.1.1)
and would distance weight the population drinking from wells
within the target distance except where there is documented
contamination in .these wells. In the revised HRS, the locations
from which the distance is measured would be the sources of
contamination at the site, not the extent of contamination as
is done in the current HRS. The Agency would also incorporate
a distance weighting factor for evaluating the potentially
exposed population. (See Section V C 8.) Thus, the population
drawing from wells located between three and four miles from
a site would be counted least heavily unless the wells they
are drinking from show documented contamination. Extending
this target distance limit to four miles is not expected to
result in overestimating the target population because the
distance weighting factors will be used to adjust the weights
of populations as a function of distance from the site, unless
-------
those populations use drinking water from wells with documented
contamination.

Aquifers Considered. The current MRS designates a single
aquifer as the aquifer of concern; this aquifer is the aquifer
that produces the highest ground water pathway score. The
targets counted are only those that use that aquifer. In the
revised HRS, if more than one aquifer is present, a migration
score would be calculated for each of the aquifers and the
aquifer with the highest score would be used to evaluate
the site. In calculating the targets value for an aquifer,
both the population using water from that aquifer and the
population using water from all overlying aquifers must be
considered, except when the hazardous substances were placed
directly in the aquifer (Proposed Rule, Section 3.0). The
inclusion of targets from overlying aquifers is based on the
assumption that contaminants must migrate through the shallower
aquifers before reaching a deeper aquifer.

Aquifer Interconnections. In the current HRS, multiple
aquifers can be considered a single aquifer if they function
as a single hydrological unit within a three-mile radius of
the site. Because specific determination of the degree of
aquifer interconnection may require professional judgment to
evaluate the site, EPA has developed guidance for applying
such judgments in determining that multiple aquifers
constitute a single hydrologic unit (Proposed Rule, Sections
3.0.1.2 and 3.0.1.2.1).

At present, both the target distance limit and the distance
over which the aquifer interconnections are determined extend
for a three-mile radius from the extent of known contamination,
except where a lateral discontinuity exists. EPA would reduce
both the distance over which geologic conditions are evaluated
for the potential to release and the distance over which aquifer
interconnections are evaluated from three miles to two miles,
except where contamination attributable to the site extends
beyond two miles; areas underlying this contamination are
included in the evaluation. This reflects EPA's belief that
the geologic conditions near the site will, in most cases, be
of primary importance in affecting the release of contaminants
to an aquifer and from upper aquifers to lower aquifers. EPA
requests comments on the two-mile radius.

Aquifer Discontinuities. Aquifer discontinuities
(Proposed Rule, Section 3.0.1.2.2) result when a geologic,
topographic, or other structure or feature completely transects
an aquifer, significantly disrupting water from flowing out
of or into the aquifer. If the discontinuity exists within
the target area, any part of an aquifer beyond the discontinuity
is not counted in the aquifer unless the discontinuity does
not entirely transect the target area. If multiple aquifers
are considered as a single unit, any discontinuity must entirely
transect the boundaries of the single unit before the area
beyond the discontinuity is discounted. EPA is proposing to
-------
-56-

extend the distance for considering aquifer discontinuity
from the current three-mile radius to a four-mile radius to
keep it consistent with the distance over which population
factors are evaluated.

Karst Aquifers. Karst aquifers are those associated with
karst terrain, which refers to a type of topography formed in
limestone, dolomite, or gypsum by dissolution by rain and
ground water. Karst aquifers often have very high ground
water flow velocities. Currently, sites in karst terrain are
not given any special consideration in the HRS. (See Proposed
Rule, Section 3.0.1.3 for a further definition of karst.) In
the several factors in the proposed revisions, karst aquifers
would be scored differently than other types of aquifers
(Proposed Rule, Sections 3.1.2 and 3.3.1). The proposed
revisions would reflect the high potential for contaminants
to migrate through karst aquifers with little reduction in
the concentration of the hazardous substance through dispersion,
dilution, or attenuation. Karst aquifers would be treated
differently than other aquifers in the depth to the aquifer/
hydraulic conductivity factor, the sorptive capacity factor,
the maximally exposed individual factor, and the population
potential contamination factor. For more detail, see the
Technical Support Document, available in the Superfund docket.
EPA also considered scoring other types of aquifers, such as
fractured bedrock, differently for similar reasons. The
Agency solicits comments on how the HRS could reflect the
special characteristics of those types of aquifers.

Ground Water Flow Direction. The current HRS does not
consider the direction of ground water flow in determining
which populations or environments may be affected by the
migration of hazardous substances at the site. The target
factors give equal weight to the entire population within a
three-mile radius from the site. In adopting the current HRS,
EPA decided that the time and level of effort required to
obtain sufficient geohydrologic information to determine
ground water direction accurately over the entire three-mile
radius would be inconsistent with the goal of expeditiously
scoring sites. A reasonably accurate determination of the
flow direction requires extensive geohydrological investigation
because the direction of flow may be altered by seasonal
variations, long-term historical changes, and the effects of
pumping wells. In addition, when considering the migration
of hazardous substances in ground water, other problems arise;
for example, immiscible liquids may migrate in a direction
other than the primary direction of the ground water flow,
making ground water flow direction an inaccurate surrogate
for the direction of contaminant flow.

In the ANPRM, EPA sought comments on the question of
using ground water flow direction measures. Most of the commenters
supported using ground water flow direction, at least under
certain conditions. EPA investigated several options for
considering ground water or contaminant flow direction, including
-------
the following.

o Estimating the direction of ground water flow from
piezometric measurements and dividing the target
population into upgradient and downgradient categories.
The two population categories would be evaluated
differently.

o Determining population categories based on the
direction(s) of observed contaminant migration rather
than on direction(s) of ground water flow. The direction
of ground water flow is a surrogate measure for evaluating
the most probable direction of contaminant migration
and is a less direct means of identifying the population
at risk. As in the above option, the population would
be divided based on location with respect to the direction
of contaminant migration. Various alternatives were
considered for how the two population categories might
be evaluated.

o Retaining the current system that does not consider
either the direction of ground water flow or contaminant
migration flow in determining the target population.

To evaluate these options, EPA considered both the
technical feasibility and the cost of obtaining reliable
information. See the Technical Support Document (available in
the Superfund docket) for a more detailed discussion of the
options considered.

EPA is proposing to retain the current system, which
does not directly consider ground water flow direction, in
evaluating the population potentially exposed to contaminants.
However, where there is known contamination in wells, the
populations normally using those wells would be weighted
higher than those only potentially exposed. Based on its
review of technical feasibility, EPA determined that even if
the general direction of the flow around the site could be
defined, the localized direction of the flow may not be consistent
with the general flow. Accurately determining the local flow
within the target distance would require extensive geohydrologic
investigations. EPA concluded that the considerable expenditure
of time and public funds that would be required for geohydro-
logical investigations is justified only at the nation's
highest priority sites, i.e., those on the NPL. The revised
HRS would indirectly take substance migration direction into
account by using the MEI factor and by assigning weights to
people drinking contaminated water either above or below
health-based benchmarks.

Likelihood of Release

The proposed revisions provide the same general structure
as the current HRS for assessing the likelihood of contaminants
to migrate from a site to an aquifer — observed release and
-------
-58-

potential for release (formerly route characteristics/contain-
ment) (Proposed Rule, section 3.1).

Observed Release. As discussed in Section V C 4, the
proposed HRS would include general criteria to define when a
release can be considered significantly above background
levels (Proposed Rule, Section 3.1.1).

Potential to Release. The proposed potential to release
factor category (Proposed Rule, Section 3.1.2) is comparable in
intent to the route characteristics/containment portion of
the current HRS. The name would be changed to clarify that
the factor represents the potential of the site to release
contaminants to an aquifer rather than the potential for the
contaminants to migrate once they enter the aquifer.

EPA is proposing a number of changes in how potential
releases are scored. The current HRS has four route
characteristics factors — depth to the aquifer, net precipita-
tion, permeability, and physical state. The values for
these factors are added together, then multiplied by the
containment factor value. The proposed HRS would use four
factors — depth to aquifer/hydraulic conductivity, net
precipitation, sorptive capacity, and containment. The release
potential would be calculated as the sum of the values of the
first three factors multiplied by the value for containment
(Proposed Rule, Section 3.1.2.5).

In the proposed HRS, the depth to aquifer would be
combined in a matrix with hydraulic conductivity (Proposed
Rule, Section 3.1.2.3). Considered together, the two factors^
provide an indication of the relative travel time required
for hazardous substances to reach the underlying aquifer. In
the current HRS depth of aquifer factor, aquifers deeper than
150 feet are assigned a value of zero. The depth to aquifer
factor would be modified in the revised HRS to include aquifers
with depths up to and exceeding 800 feet. Values would be
assigned using a matrix that combines depth with hydraulic
conductivity. This change in the depth reflects both the fact
that aquifers are known to be used at depths exceeding 800 feet
and that documented contamination has been found in deeper
aquifers. See the Technical Support Document for further
detail.

For HRS scoring purposes, the hydraulic conductivity
factor would be calculated by deriving a thickness-weighted
average hydraulic conductivity, a measure that combines the
hydraulic conductivity of each layer of geologic material between
the contaminant source and the aquifer with the thickness of
that layer (Proposed Rule, Section 3.1.2.3.2). The one exception
is when the layer consists of karst; karst aquifers would
always be assigned a thickness of zero feet regardless of
their actual thickness. The hydraulic conductivity factor is
a renaming of the permeability factor; the proposed addition
of the thickness component would make the new hydraulic
-------
-ay-

conductivity factor a more accurate measure of travel time
and, therefore, a more accurate reflection of the potential
for migration. As with aquifer interconnections, hydraulic
conductivity would be examined within a two-mile radius of
the site, except as noted in the Proposed Rule (Section 3.1.2.3).
EPA requests comments on this distance.

The net precipitation factor, which indicates the amount
of water potentially available for infiltration to ground
water, would be revised. Under the proposed HRS, the net
precipitation factor value (Proposed Rule, Section 3.1.2.2) would
be based on a new method of estimating annual net precipitation
rather than seasonal or annual net precipitation as determined
in the current HRS. In addition, the factor value would be
based on the sum of the months in which there is a positive
net precipitation. This will better reflect the potential of
hazardous substances to migrate to aquifers. A map providing
net precipitation values for specific areas will be included
in the final rule.

When the HRS was adopted, some commenters objected
because it did not consider geochemical removal mechanisms.
At the time, EPA did not believe that the data regarding
these mechanisms were sufficiently broad to warrant inclusion
in the HRS. In response to these comments, and to similar
comments on the ANPRM, EPA is proposing to add a new factor
to the ground water potential to release category, sorptive
capacity. Sorptive capacity measures the potential of geologic
materials to sorb contaminants and thereby retard their
migration to aquifers (Proposed Rule, Section 3.1.2.4). The
sorptive*capacity factor is intended to reflect relative
differences in the ability of various types of geologic materials
to inhibit the migration of contaminants. A table of sorptive
capacity values is provided in the proposed rule. Sorptive
capacity would be evaluated based on the clay and organic
carbon content of the geologic materials that occur between
the hazardous substances and the aquifer within a two-mile
radius of the site (except as noted in the Proposed Rule),
consistent with the way other geologic factors are evaluated.
EPA requests comments on the sorptive capacity factor and the
distance over which it is evaluated.

The containment factor (Proposed Rule, Section 3.1.2.1) is a
measure of the means taken to minimize or prevent the release
of hazardous substances from a site to ground water. The
containment factor would be revised in the proposed HRS to
provide a greater range of assigned values and more detailed
descriptions of each type of containment. These changes
would make the determination of conditions and of the adequacy
of containment more objective.

Certain ground water containment elements (e.g., liners,
cover thickness, and permeability) cannot be examined visually
during a site inspection. In the past, the Agency has relied
principally on the records of site owners and operators to
-------
determine the adequacy of such containment, measures, in the
absence of such records, EPA has assumed that no such containment
measures were undertaken. EPA proposes to retain this same
approach when determining containment.

The physical state affects the potential of the waste to
migrate from a site or, alternatively, for it to be contained
at a site. Physical state can, therefore, be used as a
component of either waste containment or waste migration
potential. Physical state is used in the current HRS as a
measure of waste migration potential in the ground water and
surface water pathways. EPA is proposing to eliminate the
physical state factor used in the current HRS because experience
has shown that it seldom provides meaningful discrimination
among sites. Most sites scored with the current HRS contained
at least some liquids, therefore receiving the maximum value
for physical state. In the proposed revisions, physical
state has been integrated into the containment factor to
better reflect the interrelationship of the two in the release
of hazardous substances from a source area. EPA seeks comment
on eliminating the physical state as a separate factor.

Waste Characteristics

The current waste characteristics factor category (Proposed
Rule, Section 3.2) includes hazardous waste quantity and
toxicity/persistence factors. EPA is proposing a number of
changes in the calculation of waste characteristics for all
pathways. The toxicity factor and hazardous waste quantity
factor would be revised as discussed in Sections V c 2 and 3.
Jn addition, for the ground water pathway, the toxicity factor
would be combined in a matrix with mobility rather than with
persistence as is done in the current HRS (Proposed Rule,
Section 3.2.1). EPA decided to eliminate the persistence
factor because the method currently used to evaluate persistence
is based on biodegradability and is generally not applicable
to ground water.

The Science Advisory Board, as part of its review of the
applicability of the HRS to mining waste sites, supported
the incorporation of a mobility factor in the HRS. The Board
indicated that mobility would more accurately reflect the
potential for a substance to migrate through the ground
water to a target population than does persistence.

The Board suggested that speciation of metals was an
important consideration in evaluating mobility. However,
to accurately assess the mobility of a specific metal,
the various metal species present must be determined, both in
the waste and in the subsurface environment. This, in turn,
requires knowledge of the concentrations of anions, cations
and dissolved organic materials; pH; redox potential; and
adsorption characteristics of the geologic material. In
evaluating options for mobility, EPA believed that it was not
feasible to obtain reliable measures of these parameters given
-------
-61-

the temporal and spatial variations and the difficulty in
sampling. Furthermore, the Agency concluded that the mobility
factors added to the proposed HRS will increase the accuracy
of the waste characteristics assessment.

The proposed mobility factor (Proposed Rule, Section 3.2.1.2)
would be a measure of the tendency of a hazardous substance to
become mobile in the aqueous phase. Mobility would be evaluated
for all hazardous substances that are available to migrate to
ground water. Any substance documented in an observed release
at a facility would be assigned the maximum value because its
presence is an indication that it is sufficiently mobile at
that facility to pose a hazard. For other substances, a
mobility tendency value would be assigned to specific organic
and inorganic contaminants based on water solubility, and to
inorganic cations and anions based on each ion's coefficient
of aqueous migration value. The coefficient of aqueous migration
reflects the mobility of uncombined or free inorganic substances
under geochemical conditions that maximize their mobility.
For a more detailed discussion see the Technical Support
Document in the Superfund docket.

The purpose of this new mobility measure is to increase
the accuracy of the waste characteristics factor category by
taking into account the differing abilities of substances to
migrate and, therefore, increasing the accuracy of the scoring
system. Mobility would be considered in a matrix with toxicity
and thus would play a role in the selection of the substance
used to assign the toxicity/mobility value (Proposed Rule,
Section 3.2.1.3). Combining mobility and toxicity would lead
to selecting the contaminant that poses the most significant
threat, thus increasing the accuracy of the HRS.

The toxicity/mobility factor value would be added to the
hazardous waste quantity value to obtain a waste characteristics
score.

Ground Water Targets

The ground water targets factor category (Proposed Rule,
Section 3.3) reflects the human population and resources
potentially at risk from an actual or potential release of
hazardous substances from the site to an aquifer. Currently,
the ground water targets factor category includes two factors
— a use factor and a factor derived from a matrix that combines
distance to the nearest well with the population served by
ground water. These factors are evaluated for drinking water
and irrigation wells drawing from the aquifer of concern
within the target distance of the site.

Four factors would be added together to derive a value
for the targets category under the proposed rule: ground
water use, the presence of high priority ground water areas,
the MEI, and population.
-------
-62-

Ground Water Use. Currently, the ground water use
factor takes into account four possible conditions and uses
of the ground water drawn from the aquifer within the three-
mile radius: drinking water with no unthreatened alternative
source; drinking water with alternative sources or commercial,
industrial or irrigation uses without alternatives; commercial,
industrial or irrigation uses with alternatives, or not used but
usable water; and unusable water. Only the ground water use
with the highest value is used to assign a value to the factor.
The proposed revisions (Proposed Rule, Section 3.3.3) to the
HRS would divide this factor into two subfactors — drinking
water use and other water use. The drinking water use factor
reflects the use and value of the ground water and would consider
whether the drinking water wells are private, public, or
standby, parameters the current HRS does not consider (Proposed
Rule, Section 3.3.3.1). (Private wells are defined as wells
that have less than 15 connections and serve less than 25
people.) The other water use factor would assign values if
the wells are used for specified agricultural, commercial, or
industrial purposes (Proposed Rule, Section 3.3.3.2). The
ground water use factor value would be the sum of the highest
value assigned for drinking water use and for other water
use, subject to a maximum (Proposed Rule, Section 3.3.3.3).
Expanding the ground water use factor to consider these
additional uses would provide increased discrimination.

The ANPRM specifically asked for comments on modifying
the ground water use factor to account for future use. The
Agency received comments for and against this concept. Those
favoring the addition of future use stated that unused
aquifers should be considered resources to be protected.
These commenters suggested mechanisms to predict future use,
such as town planning documents. Those opposed to considering
future use stated that the factor would be subjective and
conjectural and that communities would develop unrealistic
plans so their sites would receive higher HRS scores.

The proposed HRS would continue to place a high priority
on current use. EPA has increased the relative weight of the
revised ground water use factor among the factors in the
target category. EPA concluded that the size of the target
distance area, the consideration of alternative water
supplies, and the high value placed on the resources give
appropriate consideration to future use because it is likely
that resources being heavily used at present will continue to
be heavily used. In addition, the drinking water use factor assigns
points for aquifers that are not used, but usable. However,
the Agency recognizes that this approach may not account for
a drastic increase in future use. The Agency has not identified
a method for accurately and uniformly predicting such future
changes in land and water resource use. Therefore, the Agency
requests comment on two issues: (1) can local population
changes, land use changes, or changes in ground water use patterns
be reliably predicted within the context of the HRS? and (2)
how should the Agency weight sites where ground water resources
-------
have been degraded, regardless of whether these resources are
presently used?

Population. The population factor (Proposed Rule, Section
3'.3.2) is an indicator of the number of people actually or
potentially at risk from exposure to hazardous substances in
drinking water as well as a measure of the value of the poten-
tially affected resources. In the current HRS, all the people
who drink water drawn from wells within three miles of the
site are counted equally. The total population is then combined
in a matrix with distance to the nearest well to assign a
single factor value. In the proposed HRS, these factors
would be separated to more clearly reflect MEI risks and
resource value/population risk.

As discussed in Sections V C 5 and 8, the population
served by ground water factor would be divided into four
possible groups with the first three groups based on how the
concentration in the drinking water well compares to the
health-based benchmarks (MCLs, MCLGs, or unit cancer risk
numbers). The last group represents the population whose
wells may not be contaminated, but the aquifer itself is
contaminated, or has the potential to be contaminated. This
last group would be distance-weighted. The population factor
value would be the sum of the four population factor values,
subject to a maximum (Proposed Rule, Section 3.3.2.4).

Several other changes would affect the population factor.
EPA would clarify its definition of which wells may be considered
in determining target populations. In evaluating each aquifer,
EPA would consider the population drawing drinking water from
the aquifer being evaluated and those drawing from overlying
aquifers except when the hazardous substances have been introduced
directly into that aquifer (Proposed Rule, Section 3.3). EPA
would also consider populations drawing from a well beyond
four miles if that well has contamination attributable to the
site. In addition, EPA would use county census data when there
is no actual population count available rather than using a
conversion factor of 3.8 people per residence as in the current
HRS (Proposed Rule, Section 3.3.2). As explained in Section V C 7,
EPA has not been able to identify a reliable way of consistently
evaluating terrestrial contamination effects on the human
food chain. Therefore, the revised HRS would delete the
conversion of agricultural acreage to equivalent population
in the current HRS.

Maximally Exposed Individual. The current distance to
the nearest well factor would be treated as a separate factor
in the proposed revisions and would be used to indicate the
risk to the MEI, as discussed in Section V C 9. If the concen-
tration of a substance (or substances) at any drinking water
well exceeds health-based benchmarks and that contamination
is attributable to the site under evaluation, then the MEI
factor would be assigned the maximum factor value (Proposed
Rule, Section 3.3.1). In addition, any well drawing drinking
-------
water from a karst aquifer that underlies the site would also
be assigned the maximum value. This factor addresses the
concerns of Section 118 of SARA, which requires EPA to give
high priority to wells closed because of contamination.

High Priority Ground Water Areas. CERCLA Section 118(a),
as amended by SARA, requires the Agency to give a high
priority to sites that have contaminated a principal drinking
water supply. The use of health-based benchmarks in weighting
target populations (see Section T.C.5) is one way the revised
HRS gives greater weight to sites where actual contamination
has occurred in a drinking water well. For example, a site
responsible for contaminating wells or a water supply has the
population factor increased, depending on how the level of
contamination compares with health-based benchmarks. The use
of the health-based benchmarks ensures that sites where actual
contamination has occurred are given greater weight.

The Agency proposes adding a new factor to the target
score to take into account the presence of a Wellhead Protection
Area (WHPA) designated under Section 1428 of the Safe Drinking
Water Act (SDWA). This factor would address CERCLA Section
118(a), as amended by SARA. This factor, which receives a
maximum value if the source or hazardous substance released
from the source is located within a WHPA or zero if the source
and its hazardous substances are not located within a WHPA,
would increase the target score when a hazardous waste site
could endanger a WHPA. The Agency specifically requests
comment on the weighting that should be given to a source
located in a WHPA.

Section 1428 of the SDWA, which sets out the requirements
for the WHPAs, requires each State to develop and submit for
EPA approval, a program to protect wellhead areas supplying
public water systems from contaminants that may have an
adverse effect on human health. WHPAs are further defined
as "the surface and subsurface area surrounding a water well
or wellfield, supplying a public water system, through which
contaminants are reasonably likely to move toward and reach
such water well or wellfield." It is estimated that less than
six percent of the land area of the U.S. will likely be included
within such areas. Therefore, the WHPA designation would
likely discriminate among sites where a principal drinking
water supply is threatened. One disadvantage to the approach
of using WHPAs is that currently no States have established
WHPA programs. Section 1428(a) of the amendments requires
each State to adopt and submit to EPA by June 19, 1989, a
program designed to protect wellhead areas. Section 1428(g)
requires each State to make every reasonable effort to implement
the State wellhead protection program within two years of
submitting the program to EPA.

The Agency considered using the presence of a sole source
aquifer (SSA) rather than a WHPA to fulfill Section 118(a).
A SSA is established pursuant to Section 1424(e) of the SDWA.
-------
-03-

The Agency decided against using the SSA designation because
the criteria for SSA designation are not very selective,
i.e., the primary test is whether 50 percent or more of the
current population are served by ground water. Although there
are currently relatively few designated sole source aquifers in
the country, the criteria could potentially allow much more
of the land area of the U.S. to be designated as SSAs than
would the WHPA designations. If this were the case, assigning
a maximum value to the target score if a SSA exists within
the target distance limit would result in little discrimination
among sites.

EPA's Office of Ground Water Protection has available
the following guidance documents dealing with wellhead protection
and sole source aquifers:

(1) Sole Source Aquifer Designation Petitioner
Guidance, US EPA, February 1987.

(2) Guidelines for Delineation of Wellhead Protection
Areas, US EPA, June 1987.

(3) Guidance for Applicants for State Wellhead
Protection Program Assistance Funds Under the
Safe Drinking Water Act, US EPA, June 1987.

Comment is requested on whether the Agency should use
the WHPA designation in meeting the requirements of SARA
Section 118. Comment is also requested on the desirability
and mechanisms for incorporating other high priority areas in
a State, such as those formally recognized within a State
ground water classification system.

Ground Water Migration Score

The ground water migration score is the product of the
likelihood of release value, the waste characteristics value,
and the targets value, divided by a normalizing factor. A
ground water migration score would be calculated for each
aquifer underlying a site (Proposed Rule, Section 3.4). The
highest ground water score for an aquifer would be used as
the ground water pathway score (Proposed Rule, Section 3.5).

2. Surface Water Pathway

As can be seen in Figure 2, EPA is proposing major
changes in the surface water pathway. As required by CERCLA
Section 105 (as amended), EPA has assessed several potential
effects of surface water contamination and is proposing
to revise the surface water pathway to better consider the
threats to human health through drinking water, the human
food chain, and recreational water use. EPA is also proposing
to revise the environmental component of the surface water
pathway. Each of these threats would be evaluated separately
for likelihood of release, waste characteristics, and targets.
-------
burrace water migration ratnway
Current HRS
Release

Observed Release
or
Route Characteristics
Waste Characteristics X

3 Hazardous Waste Quantity
3 Toxicity/Persistence
3
3
3
Facility Slope/
Intervening Terrain
One year, 24 Hour
Rainfall
Physical State
Distance to Nearest Surface Water
Containment
Targets

3 Surface Water Use
3 Population Served/
Distance to Nearest
Intake Downstream
3 Distance to a Sensitive
Environment
Revised HRS
Likelihood of
Release
Drinking Water Threat

Waste Characteristics X
Observed Release 3
or 3
Potential to Release
OVERLAND FLOW
3 Containment
3 RUNOFF*
Q Distance to Surface Water
POTENTIAL TO RELEASE BY FLOOD
3 CONTAINMENT (FLOOD)
3 FLOOD FREQUENCY
Hazardous Waste Quantity*
Toxicity/Persistence
Targets

3 Surface Water Use'
3 Population*
3 MAXIMALLY EXPOSED
INDIVIDUAL
Likelihood of
Release

(same as above)
Human Food Chain Threat

Waste Characteristics X

3 Hazardous Waste Quantity*
3 Toxicity/Persistence/
BIOACCUMULATION
Targets

3 FISHERY USE
3 POPULATION*
Likelihood of
Release

(same as above)
Recreational Threat

Waste Characteristics X

3 Hazardous Waste Quantity*
3 Toxicrty/Persistence/DOSE
ADJUSTING FACTOR
Targets

3 POPULATION'
Likelihood of
Release

(same as above)
Environmental Threat

Waste Characteristics X

3 Hazardous Waste Quantity*
3 ECOSYSTEM
TOXIC ITY/Persistence
Targets
Sensitive
Environments
/ferns in italic under Current HRS have been dropped or replaced.
Items in caps under Revised HRS are new. Most items not in caps have been
revised significantly.
'Factor based on several sub-factors.
-------
The pathway score would be the sum of the scores of the threats
or subpathways. This revised structure provides a relatively
simple way to account for the different substances and targets
that may b« important for the different types of potential
exposure in the surface water pathway. The structure allows
the HRS to take into account aquatic toxicity for sensitive
environments, mammalian toxicity in drinking water, mammalian
toxicity and bioaccumulation in the food chain, and mammalian
toxicity, dermal permeability, and mass flux dilution for
recreation.

If hazardous substances could migrate from the site to
surface water in more than one watershed, each of the threats
would be evaluated over each watershed and the final pathway
score would be the sum of the scores of the watersheds (Proposed
Rule, Section 4.0.1). In evaluating a threat to a watershed,
only the information (waste quantity, observed release, etc.)
appropriate for that watershed would be used. The same
wastes, for example, would not be used to score more than one
watershed except where it is not feasible to determine the
locations of those wastes relative to the watershed boundaries.

Target Distance

In the current HRS, the target distance is measured from
the probable point where the contaminated water enters the
surface water to a point three miles downstream of the
farthest observed contamination (one mile in static water
such as a lake). EPA is proposing to extend the target
distance limit to 15-stream-miles from the probable point of
entry (a 15-mile arc for lakes and oceans). If an observed
release is based on sediment samples, as opposed to water or
benthic samples, the target distance would extend 15 miles
beyond the farthest sediment sample showing contamination
attributable to the site. This distinction is made because
contaminated sediments may serve as a continuing source of
contamination as particles become resuspended. In addition,
if there is observed release based on water or benthic samples
beyond the 15-mile limit, drinking water intakes, fisheries,
or sensitive environments up to the point of observed contamination
would be used to evaluate targets (Proposed Rule, Section
4.0.2).

To derive the 15-mile limit, EPA analyzed how far contaminants
could travel before being attenuated to the point where they
were no longer considered important in characterizing risk.
More information concerning this analysis is provided in the
Technical Support Document, available in the Superfund docket.
Based on this analysis, the Agency concluded that 15 miles
provides a reasonable balance between ensuring that all potential
receptors are evaluated and limiting the data collection
effort to a reasonable level. The current and proposed target
distance limits should not be directly compared. The current
HRS indirectly includes distance and dilution weighting factors
by assigning a value based on the distance between the probable
-------
-68-

point of entry of the contaminants into the surface water and
the intake or sensitive environment. The proposed revisions
explicitly include dilution weighting factors that are dependent
upon the volume of flow. In the ANPRM, EPA requested comments
on the target distance limit for surface water. The Agency
would like comments on this proposed revision to the target
distance.

Likelihood of Release

The current HRS evaluates the likelihood of release
as an observed release or as a potential to release. This
basic structure has been retained in the revised HRS. Because
the likelihood of release factor category value would be
calculated once and used in evaluating each of the applicable
threats (drinking water, human food chain, recreation, and
sensitive environments) at a site, it is discussed here
without reference to the threat calculations.

Observed Release. An observed release would be scored
when it can be demonstrated that a site has released hazardous
substances to surface water. Either aquatic, benthic, or
sediment sampling or direct observation of the release could
be used to demonstrate that an observed release to surface
water has occurred (Proposed Rule, Section 4.1.1.1). (See
Section V c 4 for a discussion of the observed release criteria.)

Potential to Release. The current HRS calculates the
potential to release factor category by multiplying the route
characteristics by the containment. Route characteristic
factors are the facility slope and the slope of intervening
terrain; the one-year, 24-hour rainfall; the distance to the
nearest surface water; and the physical state of the waste.
These factors are added together.

The proposed HRS would replace the potential to release
factors with two new groups of factors, overland flow and
potential to release by flood, which would be added together
to obtain the potential to release category score (Proposed
Rule, Section 4.1.1.2). The proposed overland flow factors
are comparable to the route characteristics times containment
portion of the current HRS. The name has been changed to
reflect that the factors represent the potential of the site
to release hazardous substances to surface water rather than the
the potential for hazardous substances to migrate once they
enter the surface water body. Although both the current and
revised factors measure the potential for a site to release
hazardous substances, the revised factors would emphasize
total releases rather than just peak releases.

In the current HRS, if the distance to the surface water
is more than two miles, the distance to surface water factor
is assigned a value of zero but the rest of the potential to
release factors are evaluated and scored. Under the revised
HRS, if the distance to surface water is greater than two
-------
-69-

miles, overland flow would be assigned a value of zero.
In addition, if no overland segment can be defined (e.g., the
site is in a topographic depression) the overland flow is
assigned a value of zero. Flood potential would still be evaluated,
however, as would the rest of the factors in the pathway.

overland Flow. Overland flow is the sum of the runoff
and the distance to surface water factors, multiplied by the
overland containment factor (Proposed Rule, Section 4.1.1.2.1.4).
As stated above, the maximum distance from the site to the
surface water would remain at two miles, but the scale of
assigned rating values for interim distances' would be modified
to include more distance ranges, thus expanding the scale
of possible values to better reflect the threat posed by the
overland flow pathway. (Proposed Rule, Section 4.1.1.2.1.3).

The runoff factor, the measure of runoff available for
carrying hazardous materials from a site to surface water,
would include three components: (1) rainfall, (2) runoff
curve number, and (3) drainage area. The rainfall factor
considers the potential for storms to cause surface water
contamination as a result of runoff—as reflected by the
two-year, 24-hour rainfall. The runoff curve number reflects
the ability of the soil types present and of the predominant
land surface to facilitate or impede runoff. The runoff
curve number would be obtained from a matrix of hydrologic
soil groups and the predominant land use (e.g., cultivated
land, forests, streets) within the drainage area. The drainage
area of interest is a new factor that considers the size of the
drainage area and provides an additional measure of the amount
of runoff available for hazardous substance migration. The
area of interest refers only to the area contributing runoff
from the site into the overland migration pathway. It includes
the site and any area upgradient of the site that sends water
through the site.

EPA chose this method of calculating runoff because it
more accurately assesses runoff from the site. The precipitation
value would be based on two-year, 24-hour rainfall data rather
than the current one-year, 24-hour rainfall because of data
availability. For more detail, see the Technical Support
Document, available in the Superfund docket.

The sum of the runoff factor value and distance to surface
water factor value would be multiplied by the overland containment
factor value. The containment factor is a measure of the
means taken to minimize or prevent the release of hazardous
substances from a site to surface water. The containment
factor would be revised in the proposed HRS to provide a
greater range of assigned values and more detailed descriptions
of each type of containment (Proposed Rule, Section 4.1.1.2.1.1).
These changes would make the determination of conditions and
of the adequacy of containment more objective.

EPA proposes to eliminate the current physical state and
-------
-70-

facility slope and intervening terrain factors. Most sites
scored with the current HRS contained at least some liquids,
thus scoring the maximum for physical state. Consequently/
the physical state of the waste does not provide meaningful
discrimination among sites. In the proposed revisions, physical
state has been integrated into the containment factor to better
reflect the interrelationship of the two in the release of
hazardous substances from a source area. EPA seeks comment on
eliminating the physical state as a separate factor.

Facility slope and slope of intervening terrain would be
eliminated in part because of the difficulty in estimating
these factors with sites commonly having irregular slopes or,
in the case of a lagoon, having no slope but still posing a
threat of release from a dike failure, from leakage, or from
overtopping due to an inadequate freeboard.

Potential to Release by Flood. The current HRS accounts
for flooding at hazardous waste sites only by assigning an
observed release if a site has been inundated by a flood.
The Agency is concerned that certain sites, such as
those near the banks of a river, could release hazardous
substances during a flood and that the threat of such a
release is not adequately accounted for in the current HRS.
The proposed potential to release by flood factor would better
reflect the potential for a site to release hazardous substances
if a site is flooded.

In determining the value for potential to release by
flood, each source on a site is evaluated separately for the
flood plain in which it lies, and the highest value calculated
for any source would be the value for the factor. For each
source, a value would be calculated by multiplying the flood
frequency value for each flood plain in which the source lies
by the containment (flood) value for that source for each
specific flood plain (e.g., 10-year, 100-year). (Proposed
Rule, Section 4.1.1.2.2.3).

Flood frequency (Proposed Rule, Section 4.1.1.2.2.2)
would be based on available flood plain information. A value
greater than zero would be assigned to a source for each flood
plain in which the source is located; that is, if a source is
in a 10-year and 100-year flood plain, a value would be assigned
to the source for each. Flood containment for the source
(Proposed Rule, Section 4.1.1.2.2.1) would be scored on an
all or nothing basis for each flood plain in which the source
is located. If a source is in a flood plain (e.g., a 10-year
flood plain) and a professional engineer certifies that the
containment will prevent a release of hazardous substances
from that source under such a flood, the containment factor
for that flood plain would be assigned a value of zero for that
source. If the containment would not prevent a release, the
containment flood factor would be assigned the maximum value.
Containment wou"i be evaluated for each flood plain in which
a source is locked. EPA requests comments on the inclusion
-------
-71-

of flood potential and criteria to be used for determining
flood containment.

In addition to the overland and flood mechanisms, EPA
considered adding a mechanism to evaluate a site's potential
to contaminate surface water through ground water discharges.
EPA was not able to develop a system for reliably predicting
such releases based on site inspection data. The Agency is
concerned that discharges of contaminated ground water can be
significant sources of hazardous substances in surface waters
and would consider including a mechanism for evaluating such
potential releases if reasonable. EPA solicits comments on
how such a potential could be evaluated within the context of
the HRS. Releases of contaminated ground water to surface
water are addressed as observed releases where they can be
documented.

prinking Water Threat

Waste Characteristics. The current HRS evaluates the
characteristics of the hazardous substances actually or
potentially released to the surface water pathway by adding a
value from a matrix of toxicity and persistence to a value
based on hazardous waste quantity. These factors are retained
in the proposed revisions, but evaluated differently. Scoring
of toxicity and hazardous waste quantity would be revised as
discussed in Sections V C 2 and 3.

The persistence factor in the current HRS is based on
biodegradation. To better account for actual, substance-specific
attenuation processes, the persistence factor would be revised
to include five decay processes: biodegradation, hydrolysis,
photolysis, volatilization, and free-radical oxidation. In
evaluating how the HRS handles the mobility of hazardous
substances, the Science Advisory Board supported using attenuation
measures for surface water because they most closely approximate
what actually happens in the environment. In the proposed
rule, four levels of persistence would be defined by a measure
of the half-life of the substance, that is, the time it takes
the concentration of the substance to be reduced by half
(Proposed Rule, Section 4.1.2.1.2). The persistence value
would then be assigned based on the half-life of the hazardous
substances and the type of the surface water, which together
represent the time the hazardous substance will take to travel
through the water. Of substances studied by the Agency for
purposes of revising this factor, about 90 percent receive
the maximum value for persistence in rivers, oceans, and the
Great Lakes.

In cases where persistence data do not exist, the revised
HRS would assign default values specific to the types of hazardous
substances and to the types of surface water affected by a
release. For example, hazardous substances that are metals
would be assigned a default value of three for all surface
water bodies; all other hazardous substances released to a
-------
-72-

river or stream would be assigned a default value of two,
while a default value of one would be assigned for releases to
lakes based on the longer travel time. A more detailed discussion
of the persistence factor can be found in the Technical Support
Document.

For each substance actually or potentially releasable
from the site, persistence would be combined with toxicity in
a matrix. The substance with the highest toxicity/persistence
value would be used in calculating the drinking water waste
characteristics factor category value (Proposed Rule, Sections
4.1.2.1.3 and 4.1.2.3).

The value for toxicity/persistence would be added to the
value for hazardous waste quantity to derive a waste charac-
teristics score.

Targets. The drinking water target category reflects
the humans and resources potentially at risk from exposure to
hazardous substances in drinking water obtained from surface
water sources (intakes). In the current HRS, two factors are
used to evaluate the population potentially affected: surface
water use and population served by drinking water intakes
that are within the target distance from the probable point
of entry of releases from the site to the surface water.
Currently, the population factor is combined in a matrix with
distance to an intake to produce a single assigned value.
The drinking water targets category in the revised HRS retains
the surface water use and population factors, but substantially
modifies them. The distance to an intake would be replaced
with an ME! factor that would be evaluated separately. These
three factors (surface water use, population, and MEI) would
be added together to obtain a value for the drinking water
targets.

The surface water use factor takes into account the
value of the resource and the use of the water taken from
surface water intakes within the target distance limit. In
the current HRS, the use is evaluated based on whether the
water is used for drinking; for irrigation, commercial food
preparation, or recreation; for commercial/industrial purposes
or is not used. Only the surface water use with the highest
value is used to assign a value to the factor. The proposed
revisions (Proposed Rule, Section 4.1.3..3) to the HRS would
divide this factor into two subfactors — drinking water use
and other water use. The drinking water use factor would
take into account whether the drinking water is a public or
private water supply, whether reasonable alternative supplies
exist, whether available alternatives are unthreatened by the
site, whether the water is a standby source, whether the
water has been designated for water use but is not used, and
whether it is not used or is not usable for some reason unconnected
with the site (Proposed Rule, Section 4.1.3.3.1). These new
considerations would provide a better method for evaluating the
threat posed by the site to the surface water resource. As
-------
-73-

in the current HRS, other surface water uses (such as commercial
food preparation, commercial livestock watering, or commercial
crop irrigation) would also be assigned values (Proposed
Rule, Section 4.1.3.3.2). The surface water use value would
be the sum of the highest values assigned for drinking water
use and for other water use, subject to a maximum (Proposed
Rule, Section 4.1.3.3.3), allowing both types of use to be
reflected in the score.

The population factor is an indicator of the number of
people actually or potentially at risk from exposure to
hazardous substances in drinking water. In the current HRS,
population is combined in a matrix with distance downstream
to the surface water intake to obtain a single factor value.
No distinction is made between actual and potential contami-
nation. As explained in Section V C 5, EPA has decided that
those people actually exposed to contaminated drinking water
should be weighted more heavily than those potentially exposed.
For this reason, in the revised HRS, the population factor
would be determined using four population groups. The first
three groups are based on how the concentrations at the drinking
water intakes compare with health-based benchmarks (MCLs,
MCLGs, or unit cancer risk numbers). The last of these four
population groups would represent the population whose intakes
are not known to be contaminated, but have the potential to
be contaminated. This last group would be dilution-weighted.
Where actual population counts are not available, population
figures would be derived from county census data instead of
being based on an assumption of 3.8 people per residence as
in the current HRS (Proposed Rule, Section 4.1.3.2). The
emphasis on the risk to individuals exposed to actual as
opposed to potential contamination is consistent with the
ground water approach.

EPA is also proposing to use the dilution weighting factor
at the nearest drinking water intake in assigning a value to
the maximally exposed individual factor, as discussed in Section V
C 9. The dilution weighting factor would be assigned based
on the average flow at the intake, and would be multiplied by
50 to obtain the value for the MEI factor, subject to a maximum
of 50. If the concentration at any drinking water intake
within the target distance limit exceeds a health-based benchmark
and the hazardous substances can be attributed to the site,
then the maximum value would be assigned to the MEI factor
(Proposed Rule, Section 4.1.3.1). Because mixing of hazardous
substances depends on the characteristics of the body of
water, EPA is proposing a three-mile zone of mixing for quiet
flowing rivers with an average annual flow of greater than 50
cubic feet per second. Any intake within the mixing zone
would be assigned a higher value than intakes on a similar size
river that are not in the mixing zone.

Human Food Chain Threat

CERCLA Section 105 (as amended by SARA), required EPA to
consider the possible effects of hazardous substance releases
-------
-74-

on the human food chain in revising the HRS. In the ANPRM,
EPA specifically sought comments on the addition of human food
chain factors; most commenters supported the inclusion of
factors in the HRS to assess the impact on the human food chain.

To develop the human food chain threat, EPA evaluated
other ranking systems that considered human food chain effects.
All the systems essentially used the same factors for determining
exposure through the human food chain: a bioconcentration-type
factor coupled with an estimate of the amount of food ingested.
EPA has included these factors in the proposed human food
chain threat.

The likehood of release would be calculated as explained
earlier. In evaluating exposure via the human food chain, a
single hazardous substance would be selected on the basis of
its bioaccumulation potential and its toxicity and persistence.
The same hazardous substance would be used to evaluate all
the waste characteristics and target factors for the human
food chain exposure calculations, but would not necessarily
be the same hazardous substances used in evaluating drinking
water or recreational uses or sensitive environments. All
hazardous substances known to be at the site and not contained
in such a way as to prevent migration to the surface water
would be eligible to be assessed for bioaccumulation. Each
eligible hazardous substance would be assigned a bioaccumulation
potential value and the. hazardous substance with the highest
value would be used in assessing human food chain exposure;
if more than one hazardous substance has the highest value,
the one with the highest toxicity/persistence value would be
chosen (Proposed Rule, Section 4.2.2.1.4). EPA specifically
requests comments on the use of a single hazardous substance
to score the human food chain threat.

The data that would be used to determine bioaccumulation
potential are, in order of preference, bioconcentration, the
logarithm of the octanol-water partition coefficient, and
water solubility. Because, for the purpose of the HRS, EPA
considers that bioconcentration provides the best measure of
bioaccumulation (see the Technical Support Document for further
detail), bioconcentration values are proposed as the principal
means of evaluating the potential for hazardous substances to
increase in concentration in an organism (Proposed Rule,
Section 4.2.2.1.1). Bioconcentration values would be assigned
based on either EPA Water Quality Criteria Documents or on
peer-reviewed literature.

If bioconcentration data are not available, the logarithm
of the octanol-water partition coefficient data could be used
as a surrogate. The logarithm of the octanol-water partition
coefficient has been found to have a statistically significant
linear correlation with the logarithm of the bioconcentration
factor of organic chemical compounds. If bioconcentration
and log octanol-water partition coefficient data are not
available or if the log octanol-water partition coefficient
-------
-75-

exceeds 6.0, water solubility data could be used because they
also have a statistically significant correlation with the
bioconcantration factor for organic compounds.

If a hazardous substance biomagnifies — that is, if the
tissue concentration of the bioaccumulated hazardous substance
increases at each step in the human food chain — the assigned
value would increase by one in all cases, subject to a maximum.
The bioaccumulation value would be employed to select the
hazardous substance used in evaluating the toxicity/persistence
factor, except as mentioned above when two substances have
the highest bioaccumulation value. In addition, the bioaccumu-
lation value would be used to evaluate targets. Therefore,
the same hazardous substance has to be used for both calculations.

Waste Characteristics. Hazardous waste guantity and the
toxicity/persistence factors would be calculated in the same
way as in the drinking water waste characteristics factor,
except that the predominant water category used to assign the
persistence factor would be based on the type of water (e.g.,
lake, river) between the probable point of entry and the
nearest fishery (Proposed Rule, Section 4.2.2.1.3) as opposed
to the predominant water category between the probable point
of entry and the nearest water intake. The waste characteristics
score is the sum of the hazardous waste guantity factor and
the toxicity/persistence factor values (Proposed Rule, Section
4.2.2.3).

Targets. This category would reflect the threat to people
from consumption of aquatic food chain organisms taken from
the surface water migration path. Human food chain organisms
are not limited to finfish, but could include other species
used as human food. The human population exposed to hazardous
substances through the aquatic food chain may be distinctly
different from the local population, particularly if contaminated
fish are caught for nonlocal commercial distribution. The
potentially wide distribution of contaminated fish makes
direct counting of the people who consume the fish infeasible.
Furthermore, the direct counting or estimation of the population
involved in local recreational or subsistence fishing is also
not feasible. Thus, EPA is proposing a surrogate approach:
the target population would be estimated based on the amount
of food chain products harvested from the contaminated surface
water body and the bioaccumulation of the hazardous substance.
Two factors would be summed to obtain the human food chain
targets value: population and fishery use.

The population factor value would be the sum of two
factors: potential human food chain contamination and actual
human food chain contamination (Proposed Rule, Section 4.2.3.1.3).
Actual contamination would be used to score a fishery only
if, within the limits of the observed release, there is a
closed fishery (or a portion of a closed fishery) and the
hazardous substance(s) that caused the closing have been
documented in an observed release from the site; or, a tissue
-------
-76-

sample from a fishery exceeds an FDA action level and the
hazardous substance(s) that exceeds the action level has been
documented in an observed release from the site. If either
of these conditions apply, the actual human food chain contamina-
tion (population) score for the fishery would be based on the
human food chain population value, which is derived from a
matrix of the bioaccumulation potential value and the human
food chain production values (Proposed Rule, Sections 4.2.3.1
and 4.2.3.1.2).

The human food chain production is the annual production
(in pounds) of human food chain organisms from within the
fishery under evaluation. Human food chain production would
be estimated from actual data on yield, where available, if
actual data are not available, actual data on productivity
should toe used, and, if these are not available, default
values for the standing crop of the water body should be
used. If the standing crop data are used, they would be
converted to pounds and then multiplied by 0.2 to convert the
standing crop data to human food chain production yield.
This conversion factor represents an assumed ratio between
the amount of aquatic organisms caught and the amount of
aquatic organisms within the surface water body.

The value for the actual human food chain contamination
would be the sum of the human food chain population values
for each fishery, subject to a maximum. EPA is soliciting
comments on the use of 0.2 to convert standing crop data to
catch in estimating human food chain production.

If the conditions for actual human food chain contamination
are not met, scoring of the fishery would be based on potential
contamination (Proposed Rule, Section 4.2.3.1.1). The potential
human food chain contamination (population) score would be
calculated in the same manner as actual human food chain
population, except that for each fishery, it would be
multiplied by a dilution weighting factor based on the flow,
and would be divided by 100, similar to the approach taken to
evaluate potential contamination for the drinking water threat.

The fishery use factor would reflect the nature and
utility of the fishery area. The surface water in question
would be assigned values according to whether it is used for
commercial fishing for human consumption, subsistence fishing,
or recreation or sport fishing (Proposed Rule, Section 4.2.3.2).
This factor would be a means of putting a high value on the
resource and therefore protecting both the resource and the
human users.

Either the drinking water use factor or the fishery use
factor would be assigned a value of zero so as to assign
only the highest overall use value to the surface water and
prevent double counting of surface water use. The method
used to determine which factor would be assigned the nonzero
value is specified in the Proposed Rule, Section 4.2.3.2.
-------
-77-
Reereation Threat

The current HRS does not consider the significance of
possible recreational exposures to hazardous substances
occurring at or near Superfund sites except as a subfactor of
surface water use. No mechanism is included to estimate
relative risks from recreational exposure to hazardous
substances. In response to the SARA requirement (see CERCLA
Section 105(c)(2), as amended) that risks from recreation in
contaminated surface water be appropriately assessed, the
Agency performed an analysis to estimate potential risks to
swimmers and fishermen (exclusive of any food chain risks)
who might use surface water near selected current NPL sites.
EPA concluded that health risks from recreational surface
water exposures may be potentially significant at some sites.
In addition, EPA has confirmed that some surface waters near
NPL sites are used for recreation. Consequently, EPA is
proposing a method of evaluating such risks as part of the
surface water pathway.

EPA's efforts have focused on methods of evaluating waste
characteristics and target category factors; the likelihood
of release category would be the same as in the drinking
water threat.

Waste Characteristics. In the waste characteristics
factor category, two factors would be included: toxicity/
persistence and hazardous waste quantity. In evaluating
toxicity/persistence, a dose adjusting factor would be assigned.
The dose adjusting factor represents the ratio of the dose to
an individual that would be obtained via recreation to that
dose that would be obtained via consumption of the same water
(Proposed Rule, Section 4.3.2.1.1). The dose adjusting factor
makes use of a dermal permeability constant that accounts for
dermal exposures and a mass flux dilution factor that accounts
for inhalation exposures, and is explained in more detail in
the Technical Support Document.

From the set of hazardous substances with the highest
dose adjusting factor, the substance with the highest
toxicity/persistence value would be selected. The toxicity
and persistence values would be determined by the same
procedures used in the drinking water subpathway, as would
hazardous waste quantity (Proposed Rule, Sections 4.3.2.1.4
and 4.3.2.2.)* The toxicity/persistence value would be added
to the hazardous waste quantity factor value to obtain the
recreation threat waste characteristics value.

Targets. The targets factor category reflects the
population potentially at risk from an actual or potential
release of hazardous substances from the site to surface waters
used for recreation (swimming or fishing). The targets category
has one factor, population, which would be evaluated for each
recreation area within the target distance limit based on
whether the recreation area is subject to actual contamination
or potential contamination.
-------
Only those people who use recreational areas within the
limits of an observed release would be considered in the
actual contamination factor. The recreation population value
would be determined for each recreation area using appropriate
distance categories and distance multipliers, an accessibility/
attractiveness factor, and a recreational dose adjusting
factor. Actual counts of the number of people who live within
set distances (0-5 miles, 5-10 miles, etc.) from the recreational
area would be multiplied by the distance category multipliers
for each distance. Where actual population counts are not
available, census data would be used. The values for the
accessibility/attractiveness factor would be assigned based
on the presence of specific improvements such as waterfront
parks, boat ramps, designed swimming beaches, etc. (Proposed
Rule, Section 4.3.3.1.1.1.) The dose adjusting factor value
for the substance used to assign the toxicity/persistence
value would be used to express the recreational population
exposure in terms of an equivalent drinking water population
exposure (Proposed Rule, Section 4.3.3.1.1.2). The actual
recreation population value for a recreation area would be
divided by 10, and the highest value for any recreation area
would be used as the value for this factor.

The potential contamination factor would be evaluated
for recreational areas within the target distance limit that
do not have documented contamination attributable to the
site. A human recreation population value would be determined
for each recreation area .using the same method as for the actual
contamination factor. The potential contamination value for a
recreation area would be obtained by multiplying the recreation
population value for that recreation area by the appropriate
dilution weighting factor used for drinking water populations
and dividing by 100. The potential contamination factor
value would be the highest of the potential contamination
values assigned to individual recreation areas within the
target distance limit, subject to a maximum (Proposed Rules,
Section 4.3.3.1.2).

More detailed discussion of the recreation subpathway
can be found in the Technical Support Document, along with
other options EPA considered. EPA invites comments on refining
these approaches.

The higher of the values for actual contamination and
potential contamination would be assigned as the population
factor value for the watershed (Proposed Rule, Sections 4.3.3.1.3
and 4.3.4).

Environmental Threat

Sensitive environments are included in the current HRS
surface water pathway as a factor in the targets category.
The factor is assigned a value based on the distance to the
particular type of sensitive environment involved. The
revised HRS would place more emphasis on environmental damage
-------
-79-

and expand the types of environments considered, as discussed
in Section V C 6.

The likelihood of release would be determined in the
same manner as it is in the drinking water subpathway.

Waste Characteristics. The hazardous waste quantity
factor would be revised as discussed in Section V C 3. Ecosystem
toxicity would be combined in a matrix with persistence and
would be evaluated for all hazardous substances at the site
that are available to migrate to surface water. The final
ecosystem toxicity/persistence score would be determined by
the substance with the highest assigned value. The Agency
requests comments on the appropriateness of using a single
substance to evaluate toxicity in sensitive environments.

Because exposure of sensitive environments is more
likely to be chronic than acute, the ecosystem toxicity value
would be determined by using EPA chronic water quality criteria
for the protection of aquatic life, if available. If these
data are not available, EPA acute water quality criteria
would be used and divided by 100. If EPA acute water quality
criteria are not available, the lowest LC$Q value (median
lethal dose value from animal studies) for the hazardous
substance would be used and again divided by 100 (Proposed
Rule, Section 4.4.2.1.1). The divisors are safety factors
used to account for uncertainty.

Ecosystem persistence would be evaluated as described
for drinking water, except that the predominant water category
between the probable point of entry and the nearest sensitive
environment would be used (Proposed Rule, Section 4.4.2.1.2).
The final ecosystem toxicity/persistence value would be derived
from a matrix to reflect the relationship of these two factors
in determining the relative threat posed by hazardous substances
(Proposed Rule, Section 4.4.2.1.3).

Hazardous waste quantity would be added to the toxicity/
persistence value to obtain a score for the waste characteristics
factor category.

Targets. This category reflects the sensitive environments
potentially at risk from an actual or potential release of
hazardous substances into surface water. The targets category
consists of one factor, sensitive environments. Each sensitive
environment would be given a value based on an expanded list of
sensitive environments or the Natural Heritage Program infor-
mation (see Section V C 6). Each sensitive environment would
be placed into three groups: (1) those with contamination
above ecologically-based benchmarks (Level I concentrations);
(2) those with contamination not above ecologically-based
benchmarks but significantly above background levels (Level
II concentrations); and (3) those that could potentially be
contaminated (Proposed Rule, Section 4.4.3.1). Weighting
factors would be applied to give the greatest weight to those
-------
-80-

sensitive environments with levels of contamination above the
ecologically-based benchmarks; potentially exposed sensitive
environments would be dilution weighted. (See discussion of
weighting factors and benchmarks in Sections V C 5, 6, and 8.)

The values assigned to the sensitive environments within
each of these three groups would be added together to determine
the environmental threat target factor score (Proposed Rule/
Section 4.4.3.1.4).

Surface Water Migration Pathway Score

The score for each threat (drinking water, human food chain,
recreational, and environmental) would be the product of the
likelihood of release value, the waste characteristics value,
and the targets value (Proposed Rule, Sections 4.1.4, 4.2.4,
4.3.4, and 4.4.4). The surface water migration score would
be the sum of the scores for the four types of threats, subject
to a maximum and normalized. As stated in the introduction
to this section, a surface water migration score would be
calculated for each watershed at a site (Proposed Rule, Section
4.5). The surface water migration pathway score would be the
sum of the watershed scores, subject to a maximum (Proposed
Rule, Section 4.6).

3. Air Pathway

As Figure 3 indicates, the proposed air pathway has the
same general structure based on the three factor categories
as in the current HRS air pathway. However, as stated
before, EPA is proposing to revise every factor. The
current HRS scores the air pathway using only observed
releases; if no release can be documented, the pathway score
is zero. The revised HRS would have a factor category to
evaluate a site's potential to release substances to the air.
In the waste characteristics category, a new mobility factor
would be added. The targets category would have a new factor
to assess the risk to the MEI.

Likelihood of Release

Observed Release. EPA studied different approaches to
monitoring observed releases to the air. As the Science
Advisory Board stated in its review of the air pathway: "Because
air emissions are often episodic or narrowly focused along a
particular wind direction, they are difficult to observe." If
weather conditions are unfavorable when sampling occurs
(e.g., if there is a high wind), the sampling may result in a
false negative. Improving the accuracy of air release observations
would have required either substantially more monitoring or
monitoring during specific meteorological conditions. Because,
for most site inspections, sampling must be conducted during
a single visit, which cannot always be scheduled at the optimum
time, EPA has decided to retain the current system for scoring
observed releases (Proposed Rule, Section 2.1.1). To make
-------
figure
Air Migration Pathway
Current HRS
Release X

Observed Release
Waste Characteristics X

3 Hazardous Waste Quantity
Q Toxicity
Q Reactivity and
Incompatibility
Targets
Land Use
Population Within
4-Mile Radius
Distance to Sensitive
Environment
Revised HRS
Likelihood of X
Release
Observed Release
or
POTENTIAL TO RELEASE
Q SOURCE TYPE
Q SOURCE MOBILITY*
Q SOURCE CONTAINMENT
Waste Characteristics X

Q Hazardous Waste Quantity*
a Toxicity/MOBILITY*
Targets
Land Use
Population
MAXIMALLY
EXPOSED INDIVIDUAL
Sensitive
Environments
Items in italic under Current HRS have been dropped or replaced.
Items in caps under Revised HRS are new. Most items not in caps have been
revised significantly.
'Factor based on several sub-factors.
-------
the scoring of this factor more consistent, the Agency is
proposing more specific criteria discussed in Section V c 4.

Potential to Release. SARA (see CERCLA Section 105(a)(8)(A)
as amended), required EPA to consider potential releases of
hazardous substances to the air. Because of the problems in
sampling for the air pathway, air releases are often difficult
to detect. Furthermore, sites that are not emitting hazardous
substances into the atmosphere at the time of sampling may
begin to do so at some later date. For these reasons, the
Science Advisory Board encouraged the development of a potential
to release factor for air. EPA is proposing to modify the
HRS to include the potential for release when no observed
release can be documented.

The proposed potential to release measure is intended to
provide a reliable method for evaluating, in the absence of
an observed release, the likelihood that a site will release
a potentially significant amount of hazardous substances to
the atmosphere (Proposed Rule, Section 2.1.2). The potential
for a site to release contaminants to the air is dependent on
the physical characteristics of the site, the physical and
chemical characteristics of the hazardous substances located
at the site, and the ways in which the hazardous substances
are contained. In the proposed revisions, three factors that
correspond to these characteristics would be used to evaluate
a site's potential to release hazardous substances — source
type, source mobility, and source containment. Further
information on air releases from Superfund sites and the
options EPA evaluated in developing the potential to release
calculation are provided in the Technical Support Document
and in "HRS Issue Analysis: Options for Revising the Air
Pathway,1* (Mitre, 1987) available in the Superfund docket.

A source type value would be assigned to each source at
the site that meets a minimum size requirement as specified
in the Proposed Rule, Section 2.1.2.2; only sources that
contain hazardous substances could be used to calculate size.
The six types of sources that would be assigned values are:
containers (including tanks); contaminated soil (including
land treatment); fire sites; landfills; surface impoundments;
and waste piles. The source type values reflect the likelihood
that an uncontained source of that type would release a
potentially significant amount of relatively immobile hazardous
substances to the air.

Source mobility (Proposed Rule, Section 2.1.2.3) reflects
the relative propensity of hazardous substances contained in
a source to migrate from a source as a gas or as particulates.
For a gaseous hazardous substance, the mobility factor would
be based on three physical-chemical characteristics of the
hazardous substance: its vapor pressure, Henry's constant,
and dry relative soil volatility. Gas mobility would be
scored as specified in the Proposed Rule, Section 2.1.2.3.1.
-------
-83-

Particulate mobility represents the ability of particles
contaminated with hazardous substances to escape into the
air. Since the moisture content of the soil is a relative
measure of particulate mobility, EPA is proposing to use the
Thomthwaite Precipitation-Effectiveness (PE) index, a surrogate
measure of the relative moisture content of the soil, as the
basis for this factor (Proposed Rule, Section, 2.1.2.3.2).
The gas mobility factor and the particulate mobility factors
would be combined in a matrix to obtain the mobility value
for the source (Proposed Rule, Section 2.1.2.3.3). A more
detailed discussion of mobility is included in the Technical
Support Document, available in the Superfund docket.

The third factor that would be considered to calculate
the potential for release to air is the ability of the containment
of hazardous substances to inhibit their escape. Containment
includes natural and constructed barriers to escape. EPA would
assign factor values for both gas containment and particulate
containment; the higher of the two values would be used for the
source (Proposed Rule, Section 2.1.2.1).

Each source would be assigned a value calculated by
adding its source type value and its source mobility value,
and multiplying the sum by the containment value (Proposed
Rule, Section 2.1.2.4). The release potential value would be
the highest of the values assigned to the sources at the site.

This factor approach to assessing potential for release
was chosen because the principal alternative approaches,
based on emission equations developed for sites regulated
under RCRA, were only applicable to certain types of Superfund
sites and would have required a substantial expansion of the
site inspections. As described above, EPA believes the proposed
scoring system will reflect the likelihood that the overall
site will release contaminants to the air. EPA is seeking
comments on whether these are the most appropriate factors to
assess potential to release.

Waste Characteristics

In the waste characteristics category of the current
HRS air pathway, the reactivity, incompatibility, and toxicity
of the hazardous substances are evaluated, as is the hazardous
waste quantity. The proposed HRS includes several revisions
to the evaluation of the waste characteristics factor. The
changes to toxicity and hazardous waste quantity are discussed
in Sections V C 2 and 3. The reactivity and incompatibility
factors would be deleted because these factors primarily
predict the likelihood of sudden releases. While these releases
could be important in rare cases, they may not be applicable
to the vast majority of Superfund sites. These events are
more appropriately assessed when determining the need for
removal actions that respond to imminent danger. The waste
characteristics score in the proposed HRS would be the sum of
the toxicity/mobility factor value and the hazardous waste
quantity factor value.
-------
EPA is proposing to add mobility to the waste characteristics
category for air (Proposed Rule, Section 2.2.1.2). The mobility
factor would measure the tendency of a hazardous substance to
migrate as a gas or as particulates. All hazardous substances
available to migrate to the air would be evaluated for gas
mobility. In addition, if the substance can migrate as a
particulate, the site would be evaluated for particulate
mobility. If a hazardous substance is present in a documented
release, the assigned mobility value for that substance would
be the maximum. The mobility of substances not found in
observed releases would be calculated in the same way gas and
particulate mobility are calculated under potential for release,
with the higher of the two scores being used if a substance
could migrate as either a gas or particulate.

The purpose of the mobility factor is to increase the
accuracy of the waste characteristics factor category by
taking into account the differing abilities of substances to
migrate and, therefore, the relative threats posed by their
release. The mobility and toxicity values for each substance
would be combined in a matrix to reflect the importance of
both in assessing risk; the substance with the highest toxicity/
mobility value would be used to assign the factor value (Proposed
Rule, Section 2.2.1.3). Combining toxicity and mobility to
select the substance would lead to selecting the substance
that poses the most significant threat, thus increasing the
accuracy of the HRS.

Targets

The current HRS evaluates three target factors: population
within a four-mile radius, distance to a sensitive environment,
and land use. The proposed HRS would revise these three
factors and add a factor to reflect the risk to the MEI.

Several of the proposed changes to this factor category
are discussed in Sections V C 6, 8, and 9 — the extension
of the sensitive areas definition, the distance weighting
factors, and the measurement of risk to the maximally exposed
individual. In addition, EPA is proposing changes specific
to the air pathway target factor category. While the Agency
proposes to retain the four-mile target distance limit for
humans, the target distance limit for sensitive environments
would be extended from one and two miles to four.

Public comments have suggested that the four-mile target
distance limit for the air pathway is too large. An EPA
study presented to the Science Advisory Board, however,
suggested that for sites with large emission rates of potential
carcinogens, individual risks may remain of concern even at
distances greater than four miles. In this study, EPA used a
range of plausible contaminant emission rates from Superfund
sites and a range of cancer potency values as input to a
Gaussian air dispersion model. The results provided information
on the range of risks due to air emissions found at varying
-------
-85-

distances from Superfund sites. The study — "Analysis of
the Air Target Distance Limit in the Hazard Ranking System"
(1987) — is available in the Superfund docket.

The Science Advisory Board reviewed the analysis
and recommended a dilution weighting scheme that would capture
the differences in concentrations at different distances from
the site. EPA is proposing to retain the current four-mile
target distance limit and would add weighting factors as
discussed in Section V c 8.

EPA is also proposing to revise the land use factor.
The current HRS air pathway considers five categories of land
use and assigns values to them depending on their distance
from the site. The highest assigned value for any of the
relevant land uses becomes the value used for the land use
factor. The proposed rule (Section 2.3.3) would change this
method of determining the land use value in three ways. First,
residential land use, now a single category, would be divided
into single-family residences and multi-family residences,
with the latter being assigned a higher value. Second, the
assigned value for land use would be multiplied by the distance
weighting factor. Third, the final land use factor would
be the sum of all the land uses within the target distance.
The inclusion of all land uses would provide better dis-
crimination; the greater range of assigned values and the
distance weighting would provide a more accurate assessment
of the potential risk.

The sensitive environment factor would be distance
weighted and all sensitive environments within four miles would
be evaluated and summed (Proposed Rule, Section 2.3.4). EPA
would like comments on whether the evaluation of sensitive
environments in the air pathway should be limited to terrestrial
sensitive environments.

The final score for the targets category would be the
sum of the four factors (population, sensitive environments,
land use, and MEI) (Proposed Rule, Section 2.3.5).

Air Migration Pathway Score

The air migration pathway score would be the product of
the likelihood .of release value, the waste characteristics value,
and the target value, normalized (Proposed Rule, Section 2.4).

4. Onsite Exposure Pathway

Figure 4 shows the structure of the proposed onsite
exposure pathway. There is no current onsite exposure
pathway. To parallel the proposed rule, this section discusses
the three factor categories for the resident population, then
for the nearby population.

CERCLA (Section 105(8)(A)) required EPA to take into
account the potential for direct human contact in setting
-------
ONSITE EXPUbUKt HAI nwAY
REVISED MRS*
LIKELIHOOD OF X
EXPOSURE
3 OBSERVED
CONTAMINATION
RESIDENT POPULATION THREAT

WASTE CHARACTERISTICS X TARGETS
3 TOXICITY
3 HIGH RISK POPULATION
3 TOTAL RESIDENT
POPULATION
3 TERRESTRIAL SENSITIVE
ENVIRONMENTS
LIKELIHOOD OF X
EXPOSURE
3 ACCESSIBILITY/ 3
"FREQUENCY OF USE
3 HAZARDOUS WASTE QUANTITY
NEARBY POPULATION THREAT

WASTE CHARACTERISTICS X

3 TOXICITY
TARGETS

3 POPULATION WITHIN
1 MILE
•The current Hf?S Includes » d/rtcf contact pathway, but that pathway is not used in calculating the
overall HRS migration score.
-------
-87-

priorities for the NPL. When the HRS was promulgated in
1982, EPA explained that hazards from direct contact with
hazardous substances would be addressed and controlled by a
CERCLA emergency response action prior to remedial action and
therefore the direct contact pathway did not need to be included
in the HRS migration score. The direct contact portion of
the current HRS calculates the potential for direct exposure
to hazardous substances in a way that essentially parallels
the migration pathways, but is not included in the score used
to determine if a site should be on the NPL.

In developing the proposed revisions, EPA analyzed the
Records of Decisions (RODs) produced during the first four
years of the Superfund program. (A ROD is the documentation
of the decision process associated with the selection of a
remedy for a site.) This analysis showed that in over 50
percent of the NPL sites, direct contact was listed as one of
the considerations in selecting the remedial action.
This analysis indicated that some significant risks from
direct contact may not have been completely addressed by
removal actions and should be of concern in determining
priorities for remedial action. The analysis is available in
the Superfund docket.

Based on its review and its experience, and in order to
better respond to the mandate in CERCLA Section 105(8)(A)
(now CERCLA 105 (a)(8)(A)), EPA is proposing to add a separate
onsite exposure pathway, similar to the direct -contact pathway,
that would be included in the calculation of the total HRS
site score (Proposed Rule, Section 5.0). EPA considered
incorporating direct contact exposures in the other migration
pathways. However, soil ingestion at sites probably constitutes
the most significant direct contact threat. The likelihood
of soil ingestion represents a distinctly different mode of
exposure than found in the other pathways. Therefore, EPA
decided that a separate onsite exposure pathway would more
directly and more accurately reflect the potential threat.

The proposed onsite exposure pathway score would consist
of two population groups evaluated for the three factor categories
(likelihood of exposure, waste characteristics, and targets), and
is constructed in a way similar to the surface water threats.
The first group is the resident population, including people
living on a property where contamination is (or can be inferred
to be) significantly above background levels, people attending
schools or day care on such property, or sensitive environments
that have become contaminated (Proposed Rule, Section 5.1).
The second group consists of the nearby population, composed
of people who have access to a contaminated area (Proposed
Rule, Section 5.2). EPA asks for public comments on this
breakdown.

Resident Population Threat

Likelihood of Exposure. The revised HRS would evaluate
-------
-88-

the resident population likelihood of exposure (Proposed
Rule, Section 5.1.1) based on the presence of contamination
and not on release potential, as in the other pathways, because
no migration of contaminants off-site is necessary for exposure
to occur; people live on or attend school or day care on the
site, or the contamination is in a terrestrial sensitive
environment.

The proposed HRS would require documented, analytic
evidence of contamination above background levels in order to
assign a score for the pathway. The criteria for contamination
would be the same as for the other pathways (see Section V C 4);
samples would be taken within a specified depth below the
surface.

As set forth in the revised HRS, it would be possible to
infer that properties are contaminated, even though no soil
samples demonstrate contamination on these properties, if
surrounding properties show contamination. This approach
would also require that the likely mechanisms of transport
(overland flow, air, etc.,) be considered along with topography
and other factors to determine whether such interpolation of
sampling results is reasonable. The Agency considered the
alternative approach of only counting targets living on
properties where sampling had demonstrated contamination;
however, that approach would most likely result in either
much higher costs for site inspections due to the increase in
soil sampling or in less accuracy, if insufficient samples
were taken. While the Agency has proposed the approach of
interpolating from soil sampling and other information to
demonstrate contamination, comments are solicited both on the
approach selected and on guidance for implementing it. The
guidance will need to address such issues as methods for
establishing background levels of hazardous substances
and interpretation of negative sample results within the
boundaries of the contaminated area. In addition, the Agency
recognizes that the approach of using property boundaries to
define contaminated residential land may be problematic in
certain situations, such as Federal installations or Indian
lands, where residential areas could be within the contaminated
"property" and be at considerable distance from the hazardous
substances. The Agency is seeking comment on how such
situations could be addressed in the revised HRS.

Waste Characteristics. Toxicity would be the only factor
in this factor category and would be calculated as discussed
in section v c 2.

Targets. The three target subgroups considered in the
resident population factor would be children under seven, the
total resident population, and sensitive environments (Proposed
Rule, Section 5.1.3). These values for these three factors
would be added together to obtain the targets category score.
Children under seven are considered a high risk subgroup
because they have much higher soil ingestion rates than other
-------
-89-

people. The children counted in the high risk .group would
include those attending school or day care on contaminated
property plus those who live on the contaminated property.
Individual children could be counted only once in this factor
category. These high risk individuals would be assigned
scores five times that of individuals in the rest of the
resident population (Proposed Rule, Section 5.1.3.1).

The total resident population would include everyone who
lives or goes to school on the property except those individuals
already counted under the high risk population (Proposed
Rule, Section 5.1.3.2).

A high-risk population of 10 or a total resident
population of 50 would be required to assign the maximum
target score if there are no affected sensitive environment
targets. EPA decided that onsite exposures to a very small
number of people warrant assigning a high priority to the
site because individual risks can be very high. Also, onsite
exposures can lead to an extremely high level of public concern.

The Agency requests comments on the division of the
population and on the relative weights. It also requests sugges-
tions on how the high risk population can best be determined.

Any contaminated terrestrial sensitive environments
would also be assigned a value (Proposed Rule, Section 5.1.3.3).
For this pathway, sensitive environments include only terrestrial
environments; aquatic ecosystems would be addressed in the
surface water pathway.

The resident population threat score would be calculated
by multiplying the likelihood of exposure, waste characteristics,
and targets categories (Proposed Rule, Section 5.1.4).

Nearby Population Threat

Likelihood of Exposure. The likelihood of exposure
factor category in the revised MRS would evaluate the relative
risks a site.poses to the nearby population by considering
the quantity of hazardous waste on the site and the site's
accessibility/frequency of use. The hazardous waste quantity
would be evaluated based on the total area! extent of the
contamination (Proposed Rule, Section 5.2.1.1). Contaminated
area would be used in the onsite exposure pathway for the
hazardous waste quantity factor because, for the onsite pathway,
this factor evaluates the probability that wastes will be
encountered, not the severity of exposures.

The accessibility/frequency of use factor would also
evaluate the likelihood that wastes will be encountered.
Accessibility refers to natural barriers or measures taken to
limit access. Frequency of use is assigned a value based on
estimates of use. Documented contamination of school property,
parks, etc., would be assigned the maximum value because, by
-------
their very nature, the public is attracted to them. Schools
onsite are included in this factor because other members of
the community besides students use them. The value assigned
to other contaminated properties would decline as the numbers
of barriers increase, with the lowest value assigned to areas
protected by a combination of natural and artificial barriers
that completely surround the site and by guards who control
entry at all times (Proposed Rule, Section 5.2.1.2). No site
would receive a score of zero for this factor because EPA
considers that no system can provide a completely effective
barrier.

Accessibility/frequency of use and waste quantity would
be combined in a matrix to assign a value for the likelihood
of exposure factor category (Proposed Rule, Section 5.2.1.3).

EPA specifically requests comments on the appropriateness
of basing the estimate of likelihood of exposure for the nearby
population on site area and accessibility/frequency of use,
the criteria used to assign accessibility/frequency of use,
and the scales assigned to site areas.

Waste Characteristics. Toxicity would be the only factor
in this category and would be assigned the same value for the
nearby population as it would be for the resident population.

Targets. Individuals would be counted in the nearby
population if they live or go to school or day care within a
one mile travel distance of the contaminated site. As described
in Section V c 8, the nearby population would be distance
weighted. The potential for exposure of nearby populations to
contaminated soils and wastes is expected to be significantly
less than resident population exposures. The Agency proposes
to weight nearby populations at least 20 times lower than
resident populations to reflect a reasonable estimate of the
relative exposure levels. Individuals farther from the site
would be weighted even lower to reflect the assumption that
frequency and probability of access decrease with increasing
distance from the site (Proposed Rule, Section 5.2.3).

The nearby population targets section does not include
a factor for sensitive environments. EPA concluded that
relative to the harm measured by the terrestrial sensitive
environments factors in the resident populations section and
in other pathways, the threats to sensitive environments
located near areas of contamination are much less, and therefore
inclusion of a sensitive environment factor is not warranted.
In addition, sensitive environments would be evaluated in the
other pathways.

The nearby population threat score is calculated by
multiplying the values for likelihood of exposure, waste
characteristics, and targets (Proposed Rule, Section 5.2.4)
-------
-91-

.e Exposure Pathway Score

The final pathway score would be calculated by adding
the resident population score and the nearby population
score, subject to a maximum (Proposed Rule, Section 5.3). As
discussed above, the maximum for the pathway is also the
maximum for either the resident population score or the nearby
population score.

5. Fire and Explosion

Although, the current HRS evaluates the risk of fire and
explosion at sites to determine if removal actions may be
required, the score for the fire and explosion pathway is not
included in the final HRS migration score. EPA's experience
indicates that the fire and explosion pathway would not provide
a useful basis for scoring a site for remedial action. • The
potential for fire and explosion is evaluated in another part
of the Superfund program, the removal program, to determine
if a removal action is necessary. Therefore, the proposed
HRS would not include a fire and explosion pathway.

E. CERCLA Section 125

Section 125, added by SARA, requires EPA, in revising the
HRS, to address facilities that contain substantial volumes
of waste specified in Section 3001(b)(3)(A)(i) of RCRA. These
wastes include fly ash wastes, bottom ash wastes, slag wastes,
and flue gas emission wastes generated from combustion of
coal and other fossil fuels. Section 125 requires EPA to
revise the HRS in a manner which assures the appropriate
consideration of the quantity, toxicity, and concentrations
of the hazardous constituents present in such wastes in
comparison with other wastes; the extent of, and potential
for release of such hazardous constituents; and the degree of
risk these hazardous constituents pose to human health and
the environment.

The Agency believes that the proposed revisions to the
HRS address the requirements of Section 125 in a number of
different areas. First, the toxicity factor has been revised
to include chronic and carcinogenic risks. The revised toxicity
factor will provide for a better indication of the comparative
toxicity of substances and will provide greater discrimination
among sites. Thus, the toxicity of fly ash wastes will be more
accurately reflected in HRS scores.

Second, to more fully consider the quantity and concen-
tration of hazardous constituents at fly ash waste sites, the
revised HRS will incorporate a tiered approach for calculating
the hazardous waste quantity factor. Such a tiered approach
would use the best data available at a site to calculate
waste quantity, including constituent concentration data, if
adequate. The revised HRS would consider the concentration
of a hazardous substance in three ways: (1) by assigning a
-------
-y <:-

higher score to populations drinking water with contamination
that exceeds a health-based benchmark; (2) by outlining specific
criteria for determining the significance of an observed
release, thus improving the way the HRS evaluates risk; and
(3) by distance/dilution weighting targets subject to potential
contamination.

Third, the revised HRS will consider the extent of, and
potential for, release of hazardous constituents from fly ash
waste sites into the environment by the observed release
criteria, the revised method for calculating hazardous waste
quantity, and the addition of factors that would improve the
way the HRS evaluates the potential for hazardous substances
to be released. In the ground water pathway, such factors
include the revised depth to aquifer/hydraulic conductivity
factor, the sorptive capacity factor, and the mobility factor.
To improve the potential to release evaluation .in the surface
water pathway, the revised HRS would replace the current
potential to release factors with two new groups of factors,
overland flow and potential to release by flood. In addition,
the revisions to the persistence factor in the surface water
pathway to include mechanisms for attenuation other than
biodegradation would provide a more accurate assessment of
the potential for hazardous substances to migrate. In the
air pathway, the potential for a hazardous substance to be
released would be considered by the addition of a potential
to release mechanism, which would take into account source
type, source size, and the mobility of hazardous substances
at the site.

Fourth, the degree of risk to human health and the
environment posed by such constituents would be appropriately
considered in the revised HRS by:

o Revising the toxicity factor to include
chronic toxicity;

o Improving the calculation of hazardous waste
quantity by enabling the HRS to use more complete
data, if available;

o Adding a mobility factor to the ground water and
air pathway that would better assess the potential
for contaminants to migrate;

o Revising the evaluation of potential to release in
the ground water and surface water pathways;

o Adding a potential to release category in the air
pathway;

o Specifying criteria for determining when an observed
release is significantly above background;

o Using health-based and ecological benchmarks for
weighting the targets actually exposed to contamination;
-------
-93-

o Adding distance and dilution weighting for targets
potentially exposed to contamination.

flipina Wastes

Although SARA did not require EPA to revise the HRS with
specific reference to mining wastes, the Agency received a
number of comments on mining waste issues in response to the
previous NFL rulemakings and to the ANPRM. The primary concern
of the commenters was that the HRS may be biased against high
volume, low concentration wastes because it does not adequately
consider quantity, toxicity, and concentration of hazardous
constituents.

In considering these issues, EPA evaluated studies conducted
by commenters and conducted additional studies to determine
whether HRS scores for mining sites versus non-mining sites
were too high relative to the potential hazard they posed.

One study was a comprehensive analysis of the HRS
scoring patterns of 406 sites on the NPL (mining and non-mining)
plus 297 sites considered but not on the NPL at that time.
The basic finding was that mining and non-mining NPL sites do
not differ significantly in their scoring patterns for observed
releases, population/distance scoring, or toxicity/persistence
scoring. • Mining sites generally do score higher on the hazardous
waste quantity factor. However, hazardous waste quantity is
a relatively less important determinant of HRS scores than
several other factors. In addition, the maximum hazardous
waste quantity score in the current HRS (2,500 tons and higher)
covers a wide range of quantities reported at sites (e.g.,
5,000,000+ tons). This large upper category of the scoring
range diminishes the relative impact of very large quantities
of waste.

A second study provided relevant information on waste
and constituent quantities at six high-volume waste sites.
For three of the sites, the quantity of hazardous substances
present was estimated using site-specific information on
constituent concentrations and amounts. For the other three
sites, constituent quantities were estimated based on the
quantity of hazardous waste reported on HRS scoring sheets
for the site and generic constituent concentration ranges for
the appropriate mining industry segment/district. The estimated
quantity of hazardous constituents present at each site exceeded
2,500 tons, which is the cutoff value for the maximum hazardous
waste quantity score. Therefore, these six sites would have
received the maximum hazardous waste quantity score even if
only the quantity of hazardous constituents present had been
evaluated rather than the quantity of waste. In fact, at the
six sites, the estimated amount of hazardous constituents
exceeded the total amount of hazardous wastes at more than 60
percent of other NPL sites.
-------
A third study compared HRS scores with the results of an
analysis of potential dangers for six actual mining waste
sites. The sites were chosen primarily on the basis of data
availability. Site information relating to potential risks to
human health and the environment was compiled for all four
migration pathways. Although the six sites were not randomly
selected and cannot be construed as representative of all
mining waste sites, some conclusions can be drawn. All six
sites were associated with a high potential risk rating in at
least one exposure route. In addition, they demonstrate that
any or all HRS pathways may be associated with significant
potential risk at mining waste sites.

Within the six sites, higher HRS scores generally were
associated with higher potential danger ratings and also with
sites having a large number of potentially dangerous exposure
routes. This result gives limited evidence that HRS scores
may correlate with potential risk at mining waste sites.
While all three studies covered a limited number of sites,
they do suggest that the HRS score does not unfairly
treat mining waste sites.

EPA requested the assistance of the Science Advisory
Board regarding the applicability of the HRS in scoring mining
waste sites. The studies discussed above are summarized in a
report prepared for the Science Advisory Board deliberations
entitled "The Superfund Hazard Ranking System (HRS): Applicability
to Mining Wastes Sites" (ICF, Inc., July 1987). The report
and the studies are available in the Superfund docket. The
Board examined the scientific issues pertinent to waste and
site characteristics and past HRS experience scoring mining
waste sites. The Board concluded that, based on past experience,
there is no evidence to demonstrate that the HRS is biased
against these sites. However, the Board cautioned that the
current HRS has the potential for bias when calculating a
score based on potential to release. The Board suggested
ways to improve the HRS in regard to large volume waste sites,
including modifying the toxicity factor to reflect characteristics
of metals, incorporating concentration and mobility factors,
and adding transformation parameters. The new mobility factors
in air and ground water, the revised persistence factor in
surface water, and the new sorptive capacity factor in ground
water will improve the accuracy of the revised HRS in evaluating
the potential risk posed by mining waste sites.

The Agency also requested the Science Advisory Board's
assistance on a related subject — the feasibility of using
concentration data in determining the hazardous waste quantity
factor. The report presented to the Science Advisory Board —
"The Superfund Hazard Ranking System (HRS): Feasibility of
Using Concentration Data in a Revised HRS" (ICF, Inc., July
22, 1987) — is available in the Superfund docket.

In response to the issue of using concentration data in
calculating the hazardous waste quantity factor, the Board
-------
-95-

analyzed two options besides the current method (see Section
V C 3). The proposed tiered approach is based on the Board's
recommendation. From comments it has received, EPA expects
that sufficient data may be available at certain types of
sites, which could allow these sites to be scored using the
highest tier in calculating the waste quantity factor.

The Board's conclusions and recommendations are available
in the Superfund docket.

VI. REQUIRED ANALYSES

A. Executive Order No. 12291

Under Executive Order No. 12291, the Agency must judge
whether a regulation is "major11 and thus subject to the
requirement of a Regulatory Impact Analysis. The notice
published today is not major because the rule will not result
in an effect on the economy of $100 million or more, will not
result in increased costs or prices, will not have
significant adverse effects on competition, employment,
investment, productivity, and innovation, and will not
significantly disrupt domestic or export markets.

An initial economic analysis entitled "Economic Impact
Analysis in Support of the Proposed Revisions to the Hazard
Hanking System" (U.S. EPA, January, 1988) was prepared to
estimate the incremental costs associated with alternatives
to the current MRS. This analysis compared the estimated
cost of the revised HRS with the current HRS and with two
alternative ranking systems — the Department of Energy's
Remedial Action Priority System, the model that, with the
revised HRS, did well in the site ranking panel review, and
the revised HRS with a direction of ground water flow factor
included. The analysis indicates that the revised HRS will
cost more than the current HRS, but would be less costly than
either of the other alternatives. The results of evaluating
sites using the current HRS or any other alternative model
are those costs incurred to collect the data and score a
site. The best estimate of the average cost of the current
HRS is $58,200 per site. The best estimates for the average
cost per site for the alternatives are $147,600 for the proposed
revised HRS, $217,000 for the revised HRS plus ground water
flow direction, and $261,700 for the Remedial Action Priority
System. The economic impact analysis is available for inspection
in the Superfund docket.

Based on the results of the economic analysis, EPA has
concluded that the proposed HRS is not a major rule under
Executive Order No. 12291. The proposed HRS is expected to
impose total costs of $56.0 million and expected to impose
costs on society of $9.0 million over current HRS expenditures,
well below the $100 million annual effect on the economy that
defines a major rule. At this point, it is impossible to
predict whether the revised HRS would result in more or fewer
sites being included on the NPL.
-------
-96-
* a

This proposed rule has been submitted to the Office of
Management and Budget (OMB) for review as required by Executive - ,
Order No. 12291.

B. Regulatory Flexibility Act

In accordance with the Regulatory Flexibility Act of
1980, Federal agencies must evaluate the effects of a rule
on small entities and examine alternatives that may reduce
these effects. EPA certifies that the proposed HRS will
not have a significant impact on a substantial number of
small entities.

Small businesses generally do not pay for HRS activities
and therefore, most firms will not be affected by the proposed
changes. In some cases, a responsible party may be required
to pay HRS costs. EPA prepared an analysis of the potential
impact the revised HRS would have on firms required to pay
for HRS activities. The results of the financial analysis
demonstrate that four out of five sample small firms had the
assets or income to enable them to finance HRS action. (See
Appendix A of the economic report.)

C. Paperwork Reduction Act

The information collection requirements in this proposed
rule have been submitted for approval to the Office of Management
and Budget under the Paperwork Reduction Act. 44 U.S.C.
3501 et seq. An Information Collection Request document has
been prepared by EPA (ICR No. 1488) and a copy may be
obtained from Cafl M. Koch, Information Policy Branch, EPA,
401 M St., S.W. (PM-223), Washington, D.C. 20460 or by calling
(202) 382-2739.

The public reporting burden for this collection of
information is estimated to vary from 1280 to 1500 hours,
with an average of 1390 hours per response, including time
for reviewing instructions, searching existing data sources,
gathering and maintaining the data needed, and completing and
reviewing the collection of information.

Send comments regarding the burden estimate or any other
aspect of this collection of information, including suggestions
for reducing this burden, to Chief, Information Policy Branch,
PM-223, U.S. Environmental Protection Agency, 401 M St.,
S.W., Washington, D.C. 20460; and to the Office of Information
and Regulatory Affairs, Office of Management and Budget,
Washington, D.C. 20503, marked "Attention: Desk Officer for
EPA." The final rule will respond to any OMB or public comments
on the information collection requirements contained in this
proposal.

List of Subjects in 40 CFR Part 300

Air pollution controls, Chemicals, Hazardous materials,
-------
-97-
Intergovernmental relations, Natural resources, Oil
pollution, Reporting and recordkeeping requirements,
Superfund, Waste treatment and disposal, Water pollution
control, Water supply.
Lee M. Thomas Date
Administrator
For the reasons set out in the Preamble, Title 40 of the
Code of Federal Regulations is proposed to be amended as follows:
Part 300 ~ National oil and Hazardous Substance Pollution
Contingency Plan

1. The authority citation for Part 300 is revised to read as
follows: Authority: 42 U.S.C. 9605, 9618, 9625(a); 33
U.S.C. 1321(c)(2); E.O. No. 11735, 38 FR 21243; E.O. No.
12580, 52 FR 2923.

2. Appendix A is revised to read as follows:
-------