.
United States Office of May 30, 1986
Environmental Protection Policy Analysis
Agency Office of Policy,
Planning and Evaluation
c/EPA Santa Clara Valley
Integrated Environmental
Management Project
,
Revised Stage One Report
-------
SANTA CLARA VALLEY
INTEGRATED ENVIRONMENTAL MANAGEMENT PROJECT
REVISED STAGE I REPORT
May 30, 1986
Keith Hinman
Don Schwartz
Eileen Soffer
Regulatory Integration Division
Office of Policy Analysis
Office of Policy, Planning, and Evaluation
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
ACKNOWLEDGEMENTS
The authors give special credit and thanks to Forest Reinhardt and David
Morell for their crucial role in designing the Santa Clara Valley IEMP study
and their significant contribution to the October 1985 draft of this report.
We would also like to thank John Wise, Sam Napolitano, Bob Currie, and Dan
Beardsley for their managerial support. For their substantial technical assistance,
the authors thank Greg Browder, Rod Lorang, Andy Manale, Palma Risler, Sue
Perlin, Gary Silverman, Tim Smith, and David Sullivan.
While we have benefitted greatly from the generous assistance of people
too numerous to name, we would especially like to thank the following individuals
and organizations for their valuable help throughout the first Stage of the
Santa Clara Valley Integrated Environmental Management Project:
Members of the Intergovernmental Coordinating Committee:
Nancy lanni, City of San Jose (Committee Chair)
Thomas Ferrito, Town of Los Gatos (Vice-Chair)
Sharon Albert, City of Gilroy
Lynn Briody, City of Sunnyvale
Rod Diridon, Bay Area Air Ouality Management District
Patrick Ferraro, Santa Clara Valley Water District
Homer Hyde, Regional Water Quality Control Board (former)
Dianne McKenna, Association of Bay Area Governments
Peter Snyder, Regional Water Quality Control Board
Susanne Wilson, Santa Clara County
Members of the Public Advisory Committee:
Kenneth Manaster, Santa Clara University (Committee Chair)
Eugene Leong, Association of Bay Area Governments (Vice-Chair)
Delia Alvarez, Santa Clara County Health Department
Cliff Bast, Hewlett-Packard, Industry Environmental Coordinating Council
George Adrian, Santa Clara County Water Retailers (former)
Mike Belliveau, Citizens for a Better Environment
E. H. Braatelein, Jr., Water Pollution Control Plants
Patrick Kwok, Water Pollution Control Plants
Peter Cervantes-Gautschi, Central Labor Council
Ann Coombs, League of Women Voters
Greg Cummings, San Jose Chamber of Commerce/Cummings Environmental
Will Danker, Western Oil and Gas Association/Chevron
Jim Dufour, Semiconductor Industry Association (former)
Don Fast, Industry Environmental Coordinating Committee (former)
Milton Feldstein, Bay Area Air Quality Management District
Roxanne Fynboh, Cal-OSHA
Bernice Giansiracusa, County Health Department (former)
Roger James, Regional Water Quality Control Board
Peter Jones, Fire Marshalls' Association
Ed Miller, Bay Area Air Quality Management District
-------
Dan Kriege, Santa Clara Valley Water District
June Oberdorfer, San Jose State University
Elisabeth Pate-Cornell, Stanford University
Steve Pederson, Semiconductor Industry Association
David Roe, Environmental Defense Fund
William Sanborn, Santa Clara Chamber of Commerce
Ted Smith, Silicon Valley Toxics Coalition
Greg van Wassenhove, County Agriculture Commission
Jerone Wesolowski, California Department of Health Services
Kirk Willard, Industry Environmental Coordinating Committee
Scott Yoo, San Jose Water Company
Bay Area Air Quality Management District:
Dario LeVaggi
Lew Robinson
Steve Hill
Tom Perardi
Tom Umeda
Toch Mangat
California Department of Health Services:
Cliff Bowen
Dick McMillan
Jim Stratton
Regional Water Quality Control Board:
Don Eisenberg (former)
Adam Olivieri (former)
Peter Johnson
Santa Clara Valley Water District:
David Chesterman
John Sutcliffe
Rich Pardini
Tom Iwamura
County Health Department:
Steve Brooks (former)
Glenn Hildebrand
Lee Esquibel
While this report would not exist without the support of these people
and organizations, EPA and the authors are responsible for the content.
-------
TABLE OF CONTENTS
CHAPTER ONE INTRODUCTION
CHAPTER TWO GENERAL METHODOLOGY
CHAPTER THREE ANALYSIS OF RISKS FROM EXPOSURE TO AIR TOXICS
CHAPTER FOUR ANALYSIS OF RISKS FROM TOXICS IN DRINKING WATER
CHAPTER FIVE ANALYSIS OF RISKS FROM TOXICS IN SOUTH SAN FRANCISCO BAY
CHAPTER SIX CONCLUSIONS
APPENDIX A PUTTING RISKS IN CONTEXT
APPENDIX B ADDITIONAL INFORMATION ON NON-CANCER HEALTH EFFECTS
-------
EXECUTIVE SUMMARY
-------
EXECUTIVE SUMMARY
SANTA CLARA VALLEY
INTEGRATED ENVIRONMENTAL MANAGEMENT PROJECT:
STAGE ONE REPORT
This report presents the results of the first phase of the Santa
Clara Valley Integrated Environmental Management Project (IEMP), an
innovative project designed to address the environmental and public
health problems posed by toxic chemicals in California's Santa Clara Valley.
The IEMP, sponsored by the U.S. Environmental Protection Agency, is an
effort to improve public health protection and environmental management
by applying the best scientific knowledge and management skills available
to the problems found in the Santa Clara Valley. The project's goals
are:
0 to evaluate and compare the health risks - of cancer and other
chronic, toxic effects - from toxic pollutants in the environment;
0 to use this evaluation to set priorities for further analysis
and possible control;
0 to work closely with government agencies and the community to
manage environmental public health problems.
Traditionally, EPA has developed regulations aimed at controlling
the health and environmental effects of a single industry, or a single
pollutant in a single environmental medium (such as air or water). While
substantial environmental improvement has been achieved with this approach,
some drawbacks have also become apparent: often pollution controls merely
shift the problem from one medium to another; little attention is paid to
whether Agency programs, taken as a whole, reduce health risks in the
most efficient or cost-effective way; and rarely do national standards
account for the site-specific nature of a problem.
In contrast, the integrated environmental management approach takes
account of the transfer of toxics across media - in land, air, surface
water, and groundwater. In addition, the integrated approach is founded
on the concepts of risk assessment and risk management in which estimates
of risk to public health are used as a common currency for comparing a
variety of pollution problems. Finally, by focusing on one ccmmunity, in
this case the Santa Clara Valley, the approach can assist communities in
developing environmental management strategies tailored to their unique
problems and characteristics.
Integrated' environmental management is intended to be a practical
tool for controlling pollution that threatens public health. EPA, in
partnership with state and local leaders, can use estimates of the public
health impacts of a wide range of environmental problems to compare
those problems and set priorities for risk management. Setting priorities
provides a way of working through an environmental agenda by targeting
the worst problems first in order to get the most risk reduction (and
thus public health benefit) for any given level of resources.
-------
-2-
Ihe Santa Clara Valley IEMP
The Santa Clara Valley IEMP is one of EPA's early efforts to gain
field experience with this alternative approach. Similar integrated
environmental management projects have analyzed the health risks from
environmental toxic chemicals in Philadelphia and Baltimore.
EPA chose to conduct an IEMP in the Santa Clara Valley in part
because the industrial base and environmental concerns were substantially
different from those of other areas under study. In the last three to
four decades, the area's population has grown rapidly to its current
level of about 1.4 million. In addition, the local economy has shifted
increasingly from agriculture to industry. In the 1970's, the area exper-
ienced a rapid growth in electronics and other high-tech, computer-
related industry, becoming a world leader in this field.
Currently, the northern Santa Clara Valley is well populated and has
an industrial economic base. The southern part of the Valley, by contrast,
remains more sparsely populated, with an economy based on agriculture.
The southern Santa Clara Valley, however, is expected to experience
significant population and industrial growth in the coming decades. The
IEMP study area, which roughly corresponds to the Santa Clara Valley, is
shown in Figure 1.
Some of the Santa Clara Valley's environmental concerns are at
least partly a result of its unique industrial base. In recent years,
the discovery of groundwater contamination caused by leaks and spills
from underground tanks and other waste storage areas has generated widespread
public concern; many of these leaks occurred on the grounds of electronics
firms. Other sources of toxics in the local environment are common to
most urban areas, and include automobiles, dry cleaners, sewered industrial
liquid wastes, and disinfection of drinking water supplies. The southern
Santa Clara Valley has high nitrate levels in its groundwater from past
agricultural activity.
The decision to conduct an IEMP in the Santa Clara Valley followed
extensive discussions by EPA with state and local officials, industry
representatives, and public interest groups. EPA was especially impressed
by the local response to groundwater problems associated with underground
storage tanks. In 1982-83, a coalition of local elected and regulatory
officials, industry representatives, and environmental leaders responded
effectively to the discovery of groundwater contamination in the Santa
Clara Valley. Working together, these local leaders drafted a new model
Hazardous Materials Management Ordinance (HMMO) to regulate the storage
and handling of industrial chemicals; most of the cities and the county
then enacted these ordinances within their respective jurisdictions. The
IEMP could thus build upon an unusually active coalition of local interests
cormitted to effective management of environmental risks.
-------
-3-
Figure 1 Santa Clara Valley IEKP Study Area (Shown by -I \-
^.'-.STANISLAUS .
#£,;; COUNTY:>'
Cupertino
C«mpb«H ,
Saritogs
SANTA CLARA
Monte* -
Sereno G.to. COUNTY
i : -.. >tei^^'=:^>,-^'«vi^>:--.^ i. i ( ' ^ /PTX--
^VV^-^^gll^ Jgtf$-.
: -, -,;: .Y.::-'-^ft^
'£*!&.?'#-''^
':"^!;.^:''^;v^'".::''v^'f ';/'::^?:vv^-^^^
v'..-,-' ..;:'.' '. '>/-i ; '^''-''v/v/vV''.'/''::^:'-'-:':'^^''^';COUNTY .-'.'..-'';;'-:.'.:;'
'' :'->- ^
-------
-4-
Public Participation
In conducting the Santa Clara Valley IEMP, the EPA has put a great
deal of emphasis on public participation and cooperation with other
agencies. At the project's outset, EPA established two advisory committees:
an Intergovernmental Coordinating Cormittee (ICC), consisting of local
elected officials and board members of regulatory agencies; and a Public
Advisory Committee (PAC), including staff of regulatory agencies, industry
and public interest groups, and others.
The committees have provided a public forum in which to discuss
complex environmental problems and sensitive issues of public health,
and a process by which to build understanding and consensus within the
community. In addition, the committees have provided a vehicle for
fostering cooperation among the regulatory agencies and local leaders
who need to work together to manage environmental risk effectively.
local participation through the advisory committees has substantially
improved both the quality of the IEMP analysis and the opportunities for
the project to make useful contributions.
METHODOLOGY
The IEMP applies the techniques of risk assessment and risk manage-
ment to environmental problems. Risk assessment is a means of evaluating
the potential health impact of exposure to chemicals in the environment.
Risk assessment allows decision-makers to compare the potential effects
of different pollutants (such as trichloroethane and benzene), exposure
pathways (such as air and drinking water) and sources (such as underground
tanks and automobiles), using a common denominator of human health risk.
By providing estimates of the comparative impacts of different toxic
chemicals and sources, risk assessement can be used to identify the most
serious problems.
Risk management is the process of controlling the health risks
identified through risk assessment. Risk management considers not only
the level of risk posed by a particular pollutant or source but also the
feasibility and cost of control, public preferences, and institutional
capabilities. Setting priorities for research and risk reduction is a key
aspect of risk management.
This report presents the results of the first phase of the IEMP,
which emphasized risk assessment of a broad set of toxic pollutants and
sources. The project's second stage, now beginning, will emphasize risk
management of a selected set of priority problems.
-------
- 5 -
Risk Assessment
This study uses risk assessment to evaluate and compare the potential
health risks from toxic pollutants in the air, land and water. Several
measures of risk are used in this report, including estimated risk to
particular individuals and projected risk for an entire population.
Risk to an individual is defined as the increased probability that
an individual exposed to one or more chemicals will experience a particular
adverse health effect during the course of his or her lifetime. It is
important to realize that the risk estimated for a particular type of
exposure is the incremental risk beyond that which a person faces from
exposure to other environmental or hereditary causes of disease, sometimes
referred to as the background rate of disease. In this report, we present
two types of estimated individual cancer risks: (1) average individual
risk, for the typical individual, and (2) risk to the most-exposed individual
(MEI), who may be particularly close to a source or is highly exposed
for some other reason. As explained below, for non-cancer effects, this
study estimates whether or not exposures appear to be high enough to
place a person at increased risk of an adverse health effect.
Risk to the population is the expected increased incidence (number
of cases), above the background rate, of an adverse health effect in an
exposed population. In this report, we present potential increased
numbers of cases of cancer resulting from estimated exposure to particular
chemicals and pollution sources. For health effects other than cancer,
estimates of the number of people exposed at levels high enough to pose
some increased risk are presented.
The two key elements in estimating risk are a chemical's estimated
potency, or toxicity, and human exposure to that chemical. EPA estimates
a chemical's toxicological potency on the basis of available scientific
evidence. Scientific data typically consist of laboratory studies of
animals exposed to a chemical under controlled conditions, or epidemiologic
studies of human exposure to a chemical, usually in an occupational
setting. Exposure to a chemical is estimated by measuring or estimating
the concentration at which a chemical is present in the air or water,
and then making assumptions about how much air a person breathes or
water he or she drinks. Finally, potency and exposure estimates are
combined to estimate individual and population risks. This methodology
is illustrated in Figure 2.
This project examined cancer and a number of other toxic effects,
such as birth defects, neurobehavioral effects, and effects of the immune
system, blood, liver, and kidney, that might result from long-term exposure
to environmental pollutants. Since methods of estimating cancer risks
are fairly well developed and accepted, it was possible to develop
estimates of the potential individual risks and aggregate incidence
(number of cases) for exposure to carcinogens.
EPA's Carcinogen Assessment Group (GAG) has established a method for
evaluating the potential cancer risk from a substance. First, CAG evaluates
the weight of evidence that a substance poses a cancer risk to humans.
-------
-6-
FIGURE 2 IEMP RISK ASSESSMENT METHODOLOGY
EXPOSURE
ASSESSMENT
n=i
Modeling of
Fate and
Transport
Monitoring
Ambient
Pollution
Concentrations
PATHWAY
TO EXPOSURE
LAB;:::
EXPERIMENTS
nnnrmnnn
UUUUUUUU
:EPIDEMIOLOGICAL
:'-. : STUDIES :'.'".
POTENCY
ASSESSMENT
Hazard
Identification
Quantitative
Potency
Estimates
ESTIMATED HEALTH RISK
-------
- 7 -
For those substances that may pose some risk, CAG provides a quantitative
estimate of their potency, or toxicity. The IEMP used the CAG evaluations
and, in a few cases, developed additional evaluations for substances not
yet studied by CAG.
No equivalent and accepted techniques exist for estimating the
probability or incidence of non-cancer effects (an experimental technique
for making such estimates is now being evaluated by scientists within and
outside EPA). Therefore, in evaluating the potential health risks for
effects other than cancer, this study relied on "no-effect" thresholds,
also referred to as EPA Reference Doses (RfDs) or Acceptable Daily Intake
levels (ADIs). Thresholds, or RfDs, represent an estimated dose below
which adverse health effects are assumed not to occur in most people.
In evaluating non-cancer effects, the IEMP estimated the number of people
who might be exposed at levels above an estimated no-effect threshold,
and therefore might be at risk of a toxic effect. However, the IEMP
could not estimate the possible number of cases that might occur as a
result of such exposures.
The IEMP performed an initial screening exercise to identify, frcm a
master list of about 1800 pollutants, those chemicals most likely to pose
an environmental health risk in the Santa Clara Valley. Using a combination
of exposure and toxicity criteria, the project initially identified
about 50 chemicals that might pose such risks. For this report, exposure
and health risks were estimated for 41 pollutants - all those for which
sufficient exposure evidence and toxicological data could be found.
These chemicals, an indication of their suspected toxic effects, and
their likely routes of exposure are shown in Table One.
In estimating the possible toxic health effects of such a diverse set
of pollutants, sources and exposure routes, the IEMP encountered a number
of very significant uncertainties and data gaps. In general, the IEMP
approach to this problem was to use conservative, or pessimistic,
assumptions likely to overstate possible health impacts. In addition,
the study made extensive use of sensitivity analysis of key issues, in
which health risks are estimated under several different assumptions.
Sensitivity analysis is useful in showing whether important results are
"robust," i.e., whether they hold up under a variety of different
assumptions. In pinpointing the importance of alternative assumptions in
affecting estimates of health risk, such analysis also can help to identify
those areas where research to provide better information is most important.
It is important to remember that the estimates of individual health
risk and aggregate incidence from exposure to toxics presented in this
report should not be interpreted as precise or absolute estimates of
future health effects. The simplifying assumptions and uncertainties in
both the toxicology and the exposure components of this study are simply
too great to justify a high level of confidence in the predictive value
of the results. (The important limitations and uncertainties are summarized
below.) The value of these estimates lies in their usefulness for con-
paring problems to one another, developing a rough notion of the magnitude
of possible effects, and setting priorities for risk management.
-------
-8-
TABLE ONE
TOXICS IN SANTA CLARA VALLEY
SUMMARY OF HEALTH EFFECT AND EXPOSURE DATA
TOXIC SUBSTANCE
METALS AND MINERALS
ARSENIC
BARIUM
BERYLLIUM
CADMIUM
CHROMIUM
FLUORIDE
LEAD
MERCURY
NICKEL
NITRATES
SELENIUM
SILVER
ZINC
ORGANIC CHEMICALS
BENZO(A)PYRENE (BAP)
BENZENE
BROMOFORM
CARBON TETRACHLORIDE
CHLOROFLUORCCARBON
(CFC-113)
CHLOROFORM
CHLORODIBRCMOMETHANE
CHLORAMINES
POTENTIAL
ADVERSE HEALTH EFFECTS
Type of Effect Considered in IEMP 2
CANCER
X4
X
X5
X 5
X
X
X
X
X
X
X
NON-CANCER
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
EXPOSURE PATHWAYS l
Sources of Information on Presence
of Toxics in Santa Clara Valley 3
OUTDOOR AIR
MONITORED
MONITORED
MONITORED
MONITORED
MONITORED
-
MONITORED
MONITORED
-
-
MONITORED
ESTIMATED
MODELED
-
SHORT-TERM
MONITORING
MODELED 6
MODELED
-
_
DRINKING WATER
MONITORED
MONITORED
-
MONITORED
MONITORED
MONITORED
MONITORED
MONITORED
-
MONITORED
MONITORED
MONITORED
MONITORED
MODELED
MONITORED
MONITORED
MONITORED
MONITORED/
MODELED 7
MONITORED
MONITORED
-------
-9-
TABLE ONE (cont.)
TOXICS IN SANTA CLARA VALLEY
SUMMARY OF HEALTH RISK AND EXPOSURE DATA
TOXIC SUBSTANCE
DICHLOROBENZENE
DICHLOROBRCMO-
METHANE
1,2 DICHLOROETHANE
1,1 DICHLORO-
ETHYLENE (DCE)
1,2 DCE
DBCFCP
ETHYLENE DIBROMIDE
ETHYLENE OXIDE
GASOLINE VAPORS
GLYCOL ETHERS
ISOPROPYL ALCOHOL
METHYLENE CHLORIDE
METHYL ETHYL
KETONE (MEK)
PERCH LOROETHYLENE
(PCE)
PESTICIDES*
PHENOL
TOLUENE
1,1,1-TRICHLORO-
ETHANE (TCA)
TRICHLOROETHYLENE
VINYL CHLORIDE
XYLENE
POTENTIAL
ADVERSE HEALTH EFFECTS
Type of Effect Considered in IEMP 2
CANCER
X
X
X
X
X
X
X
X
X
(8)
X
X
NON-CANCER
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
EXPOSURE PATHWAYS l
Sources of Information on Presence
of Toxics in Santa Clara Valley -*
OUTDOOR AIR
MODELED 6
_
MODELED 6
MODELED 6
-
-
ESTIMATED
MODELED
MODELED
MODELED 6
MODELED 6
MODELED
_
MODELED
-
MODELED 6
MODELED
MODELED
MODELED
-
MODELED
DRINKING WATER
-
MONITORED
-
MONITORED/
MODELED 7
MONITORED
MONITORED
MODELED
-
-
-
-
MODELED
MODELED
MONITORED/
MODELED 7
MONITORED*
-
MODELED
MONITORED/
MODELED 7
MODELED
MODELED
MODELED
-------
-10-
FOOTNOTES TO TABLE ONE:
The IEMP also performed limited analysis of the possible risks
to a hypothetical individual regularly consuming contaminated fish
caught fron the South Bay; see text and Table Six. The South Bay was
not thought to be an exposure pathway by which toxics affected many
people, and thus is not included on this table.
2 "X" indicates evidence of adverse, chronic health effect in animals or
humans. This table summarizes the type of potential adverse health
effect (cancer or non-cancer) considered for purposes of the IEMP
report. For a more complete discussion on pollutant selection and
toxicological evaluation of adverse health effects from pollutants
see chapter 2 - General Methodology.
3 The IEMP used different types of information to estimate the potential
exposure of Santa Clara Valley residents to some level of a toxic
substance. Monitored data are obtained by collecting and analyzing
samples fron the air or water in Santa Clara Valley. Modeling is a
way of estimating the ambient environmental concentration of a pollutant
by calculating the estimated dispersion pattern from sources known to
emit the substance. Estimated exposure is done in different ways
depending on available data as described more completely in the full
report.
4 There is seme dispute as to the carcinogenicity of low levels of
arsenic in drinking water. See text.
5 Cadmium and (hexavalent) chromium are assumed to be carcinogenic
through inhalation only, not ingestion.
6 These pollutants were modeled only to estimate exposures to most-
exposed individuals (MEIs) near certain sources.
7 Current exposure to this chemical was derived from monitoring data.
Possible future exposure was modeled.
8 In accordance with current EPA policy, the IEMP does not consider TCA
to be a carcinogen in its base case. For sensitivity analyses, however,
the IEMP does examine the impact of TCA as if it were a carcinogen.
* Drinking water sources have been monitored for a number of pesticides.
However, little evidence of pesticide contamination was found. See
chapter 4.
-------
- 11 -
AIR ANALYSIS
In California, the regulation of sources of air pollution is the
responsibility of the state Air Resources Board (ARE) and local Air
Quality Management Districts. The ARB establishes emissions requirements
for motor vehicles and has oversite responsibilities for control of
other sources of air pollution in the state. The local Bay Area Air
Quality Management District is directly responsible for regulating non-
vehicular sources of air pollution in the Santa Clara Valley.
While the historical emphasis of air pollution control efforts has
been on criteria pollutants, such as those causing smog, this study examines
toxic pollutants, which may pose health risks at comparatively low
environmental levels. The IEMP Stage I analysis of risk from exposure
to toxics present in outdoor air involved the study of three classes of
toxic air contaminants: organic gases, heavy metals, and organic particulates.
Organic gases. The IEMP analysis of exposure to organic gases
focused on eleven specific organic compounds plus gasoline vapors, a
mixture of compounds. This broad class of chemicals comes from many
sources. Some gases, such as benzene and ethylene dibromide, are emitted
to the atmosphere from fuel combustion or evaporation. Motor vehicles
are a major source. Emissions of other organic gases, juch as 1,1,1-
trichloroethane and methylene chloride, result (generally through evaporation)
from the use of solvents by electronics firms, other industrial and
commercial establishments, and households.
To analyze exposure to and risk from organic gases in outdoor air,
the IEMP and the Bay Area Air Quality Management District (AQMD) developed
estimates of toxic gas emissions from various sources, and then modeled
pollutant dispersion to arrive at estimated pollutant concentrations in
the ambient air. The analysis examined 25 major point sources, such as
semiconductor facilities and other industrial plants; and a variety of
small, dispersed area sources, including motor vehicles, industrial
solvent applications, fuel combustion for home heating and dry cleaners.
The organic gases emissions inventory and the dispersion model were used
to estimate human exposure and risk from different sources and pollutants.
The IEMP also estimated toxic organic releases and risks from three
sources not included in the AQMD inventories: sewage treatment plants,
municipal landfills, and groundwater aeration facilities.
Metals. Analysis included eight toxic metals, such as arsenic,
chromium and cadmium. These metals are released to the air primarily as
a result of various forms of combustion (metals are present in trace
quantities in most fuels). Airborne metals may also be present as a result
of windblown dust, which may contain the settled particles from past emissions.
EPA relied on long-term monitoring data from downtown San Jose to
estimate the concentrations of metals throughout the Valley and to project
risks from exposures to toxic metals. No emissions inventory for metals
has yet been developed. While the monitoring data are fairly reliable,
countywide projection of these concentrations is problematic. Since it
-------
- 12 -
is likely that roost metals sources are dispersed area sources (likely to
result in a more even distribution pattern than pollution dominated by
point sources), and that downtown San Jose metals concentrations are
probably somewhat higher than average for the Valley (as is the pattern
with most other pollutants), EPA judged that an appropriately conservative
assumption for this screening analysis was to estimate risks as if the
single site's monitoring data were representative of the Valley.
Organic particulates. Organic particulates, such as benzo(a)pyrene,
are toxic organic chemicals present in the air primarily in particulate,
rather than gaseous, form. Sometimes called products of incomplete
combustion (PICs), they are the result of fuel combustion from motor
vehicles, home heating sources (such as fireplaces and wood stoves) and
other sources.
No local monitoring or emissions data exist for organic particulates.
EPA made rough estimates of local levels of these chemicals by scaling
national and other data to local levels based on known sources of organic
particulates (such as residential heating and gasoline combustion) and
on local monitoring data for total suspended particulates, the best
available proxy for organic particulates. These estimates are adequate
for identifying the general magnitude of the problem, but better local
data would be needed to support regulatory actions.
DRINKING WATER ANALYSIS
Drinking water in the Santa Clara Valley comes from three sources:
local groundwater, surface water imported through the South Bay and
Hetch Hetchy Aqueducts, and local surface water. About half the drinking
water in the Valley is groundwater, and about half is surface water.
Large volumes of imported and local surface water are used to recharge
the groundwater basin artificially, both to prevent the depletion of the
aquifer and to store water supplies at relatively low cost.
Nineteen water retailers - some municipal and some private - deliver
water to the Valley's consumers, under regulation by the California
Department of Health Services. The Santa Clara Valley Water District
(SCVWD) imports and treats surface water (some imported water is purchased
directly from the City of San Francisco via the Hetch Hetchy Aqueduct)
and is responsible for overall management of the Valley's groundwater
resources. The San Francisco Bay Regional Water Quality Control Board
(RWQCB) has primary responsibility for protecting groundwater quality
(although the SCVWD and municipal authorities are also involved in pro-
tecting groundwater quality). The state Department of Health Services
(DOHS) has primary responsibility for ensuring the quality of drinking
water delivered by major public systems.
-------
- 13 -
Toxic contamination problems, as well as available data, differ
for surface water and groundwater. The issues studied in the Stage 1
analysis include:
o By-products of water treatment (disinfection). Disinfection
of water by chlorination and related processes results in the
creation of chloroform and other "trihalomethanes." Disinfection
of drinking water is necessary to protect people from diseases
that might otherwise result from microbial contamination.
Since most groundwater is not disinfected (microbial contamination
is not usually a problem with groundwater in the Santa Clara
Valley), this is primarily a concern with surface water.
o Metals and minerals. A number of inorganic substances,
including metals that may cause toxic effects, are found in
drinking water in the Santa Clara Valley. Most of these substances
are probably from natural background sources (e.g., substances
naturally present in the soil), although some may be fron past
or present man-made contamination. Metals and minerals are
present in both imported surface water and in groundwater.
o Pesticides. Runoff from agricultural areas through which
imported surface water travels may contaminate the water with
pesticides. Groundwater may be contaminated through local
pesticide use.
o Industrial chemicals from tanks, pipes and spills. The
contamination of local groundwater through leaks of underground
storage tanks, piping or simply through sloppy handling has
became a significant local concern since the discovery of a
leak at Fairchild Camera & Instrument in 1981. About 100
sites involving industrial contamination of soil or groundwater
have since been discovered in the Valley; EPA has added six
of these sites to its Superfund National Priority List, and has
proposed adding an additional twelve. Many sites involve
contamination by industrial solvents, such as trichloroethylene
and perchloroethylene.
o Fuels from tanks, pipes and spills. Santa Clara Valley has
about ten times as many fuel storage tanks as industrial chemical
storage tanks. The San Francisco Bay Regional Water Quality
Control Board has documented over 400 leaks and spills from fuel
tanks. Toxic contaminants in fuels include ethylene dibromide
and benzene.
o Organic chemicals from other sources. Organic chemicals may
contaminate groundwater by leaking from sewer lines that contain
industrial wastewater. While toxic wastes are formally barred
from Class III sanitary landfills (the only landfills sited in
Santa Clara Valley), such wastes may be present in household and
conmercial waste. According to the RW3CB, there is evidence of
historical disposal of organic chemicals in municipal landfills.
-------
- 14 -
Other potential sources of organic chemical contamination of
groundwater include above-ground chemical tanks and storage areas;
residential use of chemicals such as pesticides and cleaning
agents; leakage from septic tanks; illegal dumping; and dry
wells.
o Nitrates. Nitrates, which can cause methemoglobinemia, or "blue
baby syndrome," may be present in groundwater as a result of
fertilizer use, animal waste, and leakage from septic tanks and
sewered wastes. Parts of southern Santa Clara Valley have high
nitrate levels, and the City of Morgan Hill is under order from
the state Department of Health Services to come into compliance
with state and federal standards for nitrates in drinking water
by 1988.
Different methods were used to assess exposure and risks from these
different sources and pollutants. Historical monitoring data were used
to estimate exposure to metals and minerals. To estimate exposure to
trihalomethanes, the IEMP obtained monitoring data that reflect recent
changes in treatment practices and water quality. Direct monitoring was
also used to assess exposure to pesticides, but these data are less
complete. Risks from nitrates in the groundwater of the south County
were estimated by calculating the number of infants who might be exposed
to nitrate levels high enough to be of concern, based on monitoring
data. Current risks from contamination of drinking water by industrial
chemicals were estimated based on recently collected monitoring data for
public systems. (No comparable data were available for private wells used
as sources of drinking water.)
Estimating the extent of possible future exposure to groundwater
contamination from leaking fuel and industrial tanks and other sources
was the most conplex part of the drinking water analysis. Current groundwater
contamination, as well as future leaks and spills, may affect drinking
water wells that are currently unaffected, or worsen contamination at
already affected wells. At the same time, recently instituted programs
to improve tank construction, monitor groundwater near potential sources,
clean up contamination sites and monitor drinking water wells will signifi-
cantly reduce risks from what they might have been in an unregulated
world. To estimate possible risks over time, the IEMP modeled possible
future contaminant releases, movement and impact on drinking water wells.
This effort took into account variations in hydrogeology; voluntary
replacement of old tanks; and regulatory programs to prevent and clean
up contamination, and to monitor drinking water wells.
In general, the drinking water analysis based on monitoring data is
more reliable than that based on modeling. The analysis of future risks
from groundwater contains the greatest uncertainties. Because of this,
the modeling effort was consistently conservative, so as not to underestimate
potential risks. In addition, extensive sensitivity analysis was performed
to examine alternative assumptions for key factors such as the size of
tank leaks and the effectiveness of regulatory actions.
-------
- 15 -
SOUTH SAN FRANCISCO BAY SURFACE WATERS
Surface water contamination in the South San Francisco Bay and its
tributary streams occurs as a result of "non-point source" runoff from
both urban and rural areas, and through releases from identifiable "point"
sources, including sewage treatment plants and industrial facilities.
Fairly high levels of metals exist in the water and in the sediment of
the Bay, and both metals and organic chemicals (mostly pesticides) have
been found in fish and shellfish tissue.
While toxic (and conventional) pollutants are of concern because
of their potential impact on the health of the aquatic ecosystem, their
likely impact on human health appears to be small by comparison to the
other exposure routes examined. The IEMP was unable to estimate the number
of people exposed to toxics through South Bay surface waters, but it
seems unlikely that such exposure is widespread. South Bay water is not
a drinking water source. In addition, relatively little swimming occurs,
because of limited access. The main exposure route of concern appears
to be possible consumption of contaminated fish or shellfish. Since it
is possible that some persons consume such fish regularly, the IEMP
calculated possible individual risks for a hypothetical individual
consuming significant quantities of contaminated local fish.
LIMITATIONS AND UNCERTAINTIES
An understanding of the uncertainties and limitations that underlie
the IEMP analysis is critical to a proper interpretation of its results.
Limitations in the scope of what was studied, and uncertainties in both
the exposure and toxicological data, argue against taking the estimates too
literally. Nevertheless, decision-makers must often act now to protect
against health threats from toxic chemicals and cannot afford to wait for
scientific certainty. The IEMP analysis uses the best information
available today to estimate health risks from toxics so that decisions
that cannot wait will be as informed as possible.
Limitations in Scope
The reader should recall that this analysis does not directly
examine disease incidence in the local population and attempt to link it
with environmental exposure. Because the analysis is not an epidemiologic
study, it is not intended to and does not answer questions such as what
may have caused a statistically higher rate of birth defects in the Los
Paseos area. Instead, the IEMP attempts to evaluate what health effects
might result from current and future environmental exposures.
This analysis does not attempt to estimate the health risks from
all chemicals that individuals may be exposed to in their daily lives.
The IEMP did not estimate risks from indoor air contaminants, nor those
from occupational exposures. (The IEMP has commissioned a scoping study
on occupational exposures, which is now in progress.) Similarly, risks
from contaminants in food are not estimated. Qnitting analysis of these
-------
- 16 -
routes of exposure does not imply that they are unimportant; indeed, it
is quite possible that risks from any of these exposure pathways could
exceed risks from the exposures that we did examine. The IEMP decided
not to assess these exposure routes because of resource limitations and
because they are outside EPA's traditional purview and area of expertise.
The imp chose not to analyze exposure to and risks from conventional
pollutants in air and water (such as ozone and oxides of nitrogen and
sulfur in air, and oxygen-depleting substances and oil and grease in
water) because EPA believed it could make a more significant contribution
by concentrating on less well understood and less regulated toxic chemicals
(largely organic chemicals and heavy metals thought to be potentially
hazardous at low levels).
Finally, the IEMP did not estimate risks from possible infrequent,
accidental releases of toxic chemicals, such as a major release of a
toxic gas. (The study did estimate risks from more frequent and predic-
table accidental releases, such as tank leaks and chemical spills.) The
probability and magnitude of such an accident is very difficult to estimate,
and the likely risk from such an event is therefore difficult to quantify.
The omission of such events from this analysis does not imply that possible
accidents are not an important environmental and public health concern.
In sum, it should be clearly understood that this report is not an
analysis of health risks from all possible exposures to potentially
dangerous chemicals in the Santa Clara Valley.
Limitations in Exposure Data
Beyond these intentional limitations in scope, the study's exposure
and toxicological estimates are uncertain in a number of potentially
important ways. On the exposure assessment side, one limitation of the
analysis is that it did not exhaustively examine all sources and pollutants.
While the IEMP has tried to identify and assess risks from the most
significant sources and pollutants, it was unable to estimate exposure
to some chemicals, such as arsine and phosphine, because of a lack of
data. In reality, of course, chemicals not included in the study may
also pose some health risk.
Even where exposure data were available, those data varied significantly
in their quality. Thus, the resulting exposure estimates vary in their
reliability. Those based on extensive monitoring, such as for trihalome-
thanes and inorganic substances in water, are probably fairly good.
Those based on less extensive monitoring, such as for metals in air
(based on a single long-term monitoring station) are somewhat less
reliable.
Exposure estimates derived from modeling also vary in their reliability.
Estimates of exposure to toxic organic chemicals in air, calculated using
a dispersion model, are dependent primarily on the quality of the emissions
estimates and other factors such as meteorological data. The range
-------
- 17 -
of possible error for most pollutants is probably well under an order
of magnitude. The analysis of the future risks from groundwater con-
tamination, which relies heavily on engineering assumptions and modeling
of future events, is more uncertain. Where there are significant uncer-
tainties, such as in the groundwater exposure analysis, we have attempted
to make assumptions that are likely to err on the side of overestimating
possible health impacts. In addition, we have performed sensitivity
analysis of particularly important variables, such as possible chemical
reactions and the effectiveness of regulatory actions.
Limitations in Toxicological data
Estimates of the potential health effects of particular chemicals are
designed to be conservative (i.e., more likely to overestimate toxic
health effects than to underestimate them) in several ways. Health
effects observed in laboratory animals are assumed to be a reasonable
indicator of potential effects in humans. In converting the animal data
to predicted human responses, and in extrapolating from high doses to
low doses, EPA uses models that yield a plausible upper-bound estimate
of potency rather than a "best guess" estimate.
On the other hand, many substances of potential concern have never
been evaluated scientifically, or have not been evaluated in sufficient
detail to allow estimation of effects on humans. For example, lead
(present in air, water, and dust) is thought to pose a health risk to
children at ambient levels; currently, however, EPA has no established
way of estimating individual risks or numbers of possible cases. EPA is
likely to be aware of the dangers from many of the most potent chemicals,
since the evidence for their toxicity will typically be the most obvious;
however, it is possible that some chemicals about which we currently
know little may someday be demonstrated to be toxic.
Because of the many uncertainties and potential emissions, it is
impossible to say whether the total risk estimates presented here are
over- or underestimates of total toxic health risks from pollutants in
air and drinking water. For those chemicals for which the IEMP was able
to make guantitative estimates of exposures and risks, it is more likely
that risks are overestimated than underestimated. To the extent that
toxic chemicals about which we currently know little have been left out,
risks may be underestimated. The value of the IEMP methodology is that
it allows an evaluation and comparison of the health risk from chemicals
and pollution sources about which we know something. Management of
these risks, based on the best current information, can proceed, while
research continues on the effects of chemicals about which little is
currently known.
-------
- 18 -
RESULTS OF STAGE I RISK ASSESSMENT
A brief summary of estimated environmental exposures is presented
below, followed by a presentation of estimated health risks from
environmental toxics.
Exposure
Detailed estimates of exposure to toxic chemicals are too lengthy to
present in this summary; the interested reader is referred to the full
report. In air, dispersion modeling of toxic organic substances indicates
that pollutant concentrations are generally highest in the northern part
of the study area, which is more industrialized and more heavily populated.
Monitoring data for toxic metals indicates that they are present in the
Santa Clara Valley's outdoor air at low levels - in some cases, the lower
end of the range of concentrations is below the detection limit for the
analytic equipment used.
Overall, concentrations of air toxics modeled or monitored in the
Santa Clara Valley appear to be similar to or lower than pollutant con-
centrations typical of urban areas. Estimated average concentration
levels for most chemicals examined were below 5 micrograms per cubic
meter (ug/m-^).
Estimated exposures to most-exposed individuals (MEIs) near
sources of air toxics (such as semiconductor facilities, dry cleaners and
traffic intersections) were typically five to one hundred times higher
than the average concentration levels. The difference between average
and MEI exposures was greater for chemicals such as ethylene oxide and
chloroform whose emissions were dominated by a few point sources, and
less for chemicals such as xylene and toluene, which are emitted by
many dispersed sources.
About half the population in the Santa Clara Valley is exposed to
trihalomethanes in treated drinking water, at levels that are fairly
typical for disinfected water (about 20-80 micrograms per liter (ug/1)).
Highly exposed individuals are estimated to be exposed to THM levels at
the high end of this range but below the 100 ug/1 standard.
Thirty-six public wells and about 56 known private wells have been
affected by industrial chemicals. In the majority of cases where a
source has been identified, the pollution has resulted from leaking
underground tanks or associated chemical spills. Some operating public
wells are serving water containing 1,1,1-trichloroethane, perchloroethylene,
l,l,2-trichloro-l,2,2-trifluoroethane, carbon tetrachloride, and a few
other chemicals in the low parts per billion, well below current state
drinking water standards. (The highest concentration level recorded at
an operating public well is seven parts per billion, or ug/1). About
129,000 people are currently drinking water from public wells with low
levels of contamination.
-------
- 19 -
Recent testing by the County Health Department of 171 private wells
found that about 8% of the wells were affected by detectable contamination
by synthetic organic chemicals, and that almost 40% were affected by
bacteriological contamination (i.e., they were unsanitary). Wells were
not selected by random sampling, so these figures are not necessarily
representative of other private wells in the Valley.
Modeling of possible future drinking water contamination under
conservative (pessimistic) assumptions yielded estimated exposures signifi-
cantly higher than current levels, and included some pollutants (such as
gasoline constituents) not yet seen in drinking water wells. Concentration
levels were estimated to be significantly higher at private wells than
at public wells, because public wells benefit from greater regulatory
and natural hydrogeologic protection. The IEMP estimated that future
exposure to contaminated groundwater sources of drinking water could
affect about 15% of the population, in addition to those already affected.
In the southern parf of the county, several public wells and an
unknown number of private wells contain levels of nitrates that are above
state and federal standards. Large systems exceeding standards are
under order to comply by 1988. Little evidence was found of pesticide
contamination of drinking water, either in local groundwater or in imported
surface water.
Health Risks
1. Overall Cancer Risk; EPA's findings suggest that the estimated
cancer risks from the toxic chemicals and sources studied are apparently
a small proportion (well below one percent) of total cancer cases in the
Valley. Since any level of exposure to a carcinogen is assumed to pose
some risk, all 1.4 million residents of the Santa Clara Valley are projected
to face some level of increased cancer risk as a result of environmental
exposure. EPA estimated that exposure to the pollutants and sources
examined may be responsible for about four cases of cancer per year; an
estimated 3,600 cases of cancer occur annually in Santa Clara County.1
This finding, although tentative, provides an important perspective on
health risks from toxic chemicals in the outdoor air and drinking water
in the Santa Clara Valley as compared to other possible means of exposure
to toxic substances, such as smoking, diet, occupation, and indoor air.
However, it is important to keep in mind that this study examined a
relatively small number of known toxic chemicals; exposure to many thousands
of chemicals in the air and drinking water about which we know little
may also be a source of significant, although currently unknown, health
risk.
1 Ratio of estimated cancer cases to estimated cancer deaths, from 1983
national data from the American Cancer Society: 1.93. Cancer deaths in
Santa Clara County. 1984: 1,879. 1,879 cancer deaths X 1.93 cases/death
= 3,626 estimated cancer cases in Santa Clara County.
-------
-20-
Average individual cancer risk estimates for typical individuals
exposed to toxics in both air and drinking water indicate a potential
increase in cancer probability of about 200 in a million over a lifetime.
This estimate of increased risk is the projected cumulative risk for
exposure to all sources and pollutants examined. Projected individual
cancer risk is a small proportion of the total lifetime cancer risk for
an average person of about one in four (250,000 in a million). Of course,
individuals who are particularly highly exposed to chemicals, by virtue
of their proximity to a source or for some other reason, may face
significantly higher-than-average cancer risk. (Risks to such highly-
exposed individuals are discussed below.)
2. Non-cancer Risks; EPA estimated that about 10% of the population
in the Santa Clara Valley may be exposed to chemicals at levels high enough
to pose a risk of effects other than cancer. Populations estimated to be
at risk of non-cancer health effects due to exposures above no-effect
thresholds are shown in Table Two.l
The IEMP estimated that exposure to benzene in the air could pose an
increased risk of lowered blood cell counts to about 100,000 people in
the Santa Clara Valley. This exposure is the most widespread exposure, at
a level above an estimated no-effect threshold, of any chemical examined.
Benzene is released primarily by vehicles.
1 TCA has not demonstrated any teratogenic potential in published
studies conducted using rodent species. Therefore, the IEMP base-case
analysis assumes that exposure to TCA poses no risk of fetal effects. An
unpublished study, which has not undergone scientific peer review, reports
fetotoxic effects (cardiac malformations) in rat pups exposed in utero to
TCA (Dapson et al., 1984). In order to assess the importance to Santa
Clara Valley residents of further research on this issue, the IEMP uses
the Dapson study to examine the possible impact of TCA under the alternative
assumption that exposures above an estimated threshold based on that
study's results could pose the risk of fetal effects. THE SENSITIVITY
RESULTS SHOULD NOT BE INTERPRETED AS INDICATING WHETHER OR NOT A RISK IN
FACT EXISTS; EPA RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK MANAGEMENT
DECISION-MAKING OR REGULATORY ACTION. Under this alternative assumption,
the IEMP projects that about 3,000 people, mostly those using private
wells, could be exposed to levels of TCA in their drinking water that
exceed the estimated threshold. In addition, most-exposed individuals
downwind of an industrial facility are projected to be exposed at levels
above the estimated threshold in the air. These findings suggest that
more research is appropriate, both on actual levels of exposure and on
TCA's potential adverse effects. The National Toxicology Program has
commissioned a project to repeat the limited Dapson study; results are
expected in Fall of 1986.
-------
-21-
TABLE TWO
POPULATIONS ESTIMATED TO BE AT RISK OF NON-CANCER HEALTH EFFECTS
IN SANTA CLARA VALLEY, BY POLLUTANT
POLLUTANT
EXPOSURE POTENTIAL
PATHWAY HEALTH EFFECTS
POPULATION
EXPOSED ABOVE
NO-EFFECT
THRESHOLD
PRIMARY
SOURCE(S)
VOLATILE ORGANIC
CHEMICALS
Benzene Air
1,1 Dichloro- Water
ethylene
Methylene Water
Chloride
1,1,1 Trichloro- Water
ethane 2
Trichloro- Water
ethylene
Vinyl Water
Chloride
Blood 100,000 Motor Vehicles
Liver, kidney 20 - 340 Underground Tanks
Liver, fetal
Liver, neuro-
behavioral
Liver, neuro-
behavioral
Liver, kidney,
cardiovascular
0-50 Underground Tanks
10 - 100 Underground Tanks
0-10 Underground Tanks
0-10 Underground Tanks
METALS AND
INORGANIC SUBSTANCES
Nitrates
Water
Blue baby
syndrome
50 - 100 Fertilizer, Septic
Tanks
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF THE POPULATIONS POTENTIALLY AT RISK OF DISEASE ARE ONLY ROUGH
APPROXIMATIONS. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND CHEMICAL
TOXICITY. See Text. UNLIKE CANCER RISK ESTIMATES IN THIS REPORT, THESE ARE
ESTIMATES OF POPULATIONS EXPOSED AT LEVELS THAT MAY POSE HEALTH RISK; THEY ARE
NOT ESTIMATES OF POSSIBLE NUMBERS OF CASES OR PROBABILITY OF DISEASE.
1 In many cases, risks for most-exposed individuals (MEIs) were estimated
without estimating populations involved. Such risk estimates are presented
in table six.
2 The IEMP conducted sensitivity analysis on TCA for possible fetal effects. See
footnote to text.
-------
-22-
In the southern part of the Santa Clara Valley, nitrate contamination
of groundwater supplies of drinking water is above threshold levels
estimated to pose an increased risk to infants of methemoglobinemia, or
blue baby syndrome. The IEMP estimates that up to 50 or 100 babies may,
at any one time, be exposed to nitrate levels high enough to pose risk.
In addition, under sane assumptions about the way groundwater
contamination may affect drinking water supplies, the IEMP projects
that several hundred people who drink from private wells could be at
increased risk of a variety of effects - including birth defects and
neurobehavioral, cardiovascular, liver, blood and kidney effects - from
industrial contaminants from tank leaks and spills. Concentration levels
in public well water are projected to remain below no-effect thresholds,
even under conservative assumptions.
Substantial evidence exists that lead may cause toxic effects,
including blood effects and decreased IQ, particularly in children, who
are most sensitive to it. Lead is present in air, dust and water, as a
result of combustion of leaded gasoline, use of lead solder in pipes,
and other sources. The IEMP was unable to calculate risks from lead
exposure in this analysis because of a lack of an accepted EPA method
for doing so. The IEMP hopes to estimate health risks from lead in the
Santa Clara Valley as a part of follow-on work in Stage II.
It is important to note that exposure to toxic chemicals in the air
or drinking water may pose some health risk at levels below estimated
thresholds if exposures from other sources - such as diet or occupation -
are significant. Even in this instance, comparisons with estimated
thresholds provide a useful indication of the significance of the portion
of exposure due to outdoor air or drinking water. Environmental exposures
at or near estimated thresholds are likely to pose a more significant
added risk than exposures well below a threshold.
3. Risks by Exposure Route; outdoor air and imported surface
supplies of drinking water appear to be the major exposure routes by
which toxic contaminants in the ambient environment are likely to affect
most people. The estimated breakdown of cancer risk by exposure route
is shown in Table Three.
Estimated toxic health risk through different exposure routes (e.g.,
air or drinking water) generally reflects the extent of exposure to toxic
chemicals through those routes. Exposure to air toxics is the most
widespread; everyone breathes the air and all 1.4 million Santa Clara
Valley residents are estimated to be at some increased health risk from
toxic air pollutants. Not surprisingly, toxic chemicals in the air are
estimated to pose the most significant health risks among the exposure
routes studied, over two estimated additional cancer cases per year.
-------
-23-
TABLE THREE
ESTIMATED INCREASE IN CANCER INCIDENCE
IN SANTA CLARA VALLEY, BY EXPOSURE PATHWAY
EXPOSURE PATHWAY
POINT ESTIMATE
OF ANNUAL
INCREASE IN CANCER
INCIDENCE (Range)
WEIGH1 OF EVIDENCE
FOR CARCINOGENICITY
Air
2.2 (0.8 - 7.8)
A-B2
Drinking Water
Surface Water
Groundwater
1.3 (1.3 - 8.3)
0.06 (0.04 - 0.3)
A-B2
A-C
TOTAL
3.6 (2.1 - 16.4)
A-C
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS
OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND
POTENCY, AND ARE THEREFORE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE
THEM. See text.
1. The weight of evidence of carcinogenicity for the compounds included in
the analysis varies greatly, from very limited to very substantial. According
to EPA's categorization of levels of evidence of carcinogenicity, A = proven
human carcinogen; Bl = probable human carcinogen (limited hunan evidence);
B2 = probable hunan carcinogen (insufficient human evidence but sufficient
animal evidence); C = possible human carcinogen D = not classifiable; E= no
evidence.
-------
-24-
Cancer risks from inported drinking water supplies are estimated to
be somewhat lower than those from air sources - slightly over one
additional case per year. This estimated risk results primarily from
exposure to disinfection by-products, to which half the Valley's population
is exposed. (See Conclusion 5 for more details).
One of the more striking findings of this study is that overall
risks from consumption of contaminated groundwater are estimated to be
low (about 1-2% of the cancer risk among the sources examined in this
study, or about one additional cancer case every 15 to 30 years).
Estimated cancer risks from current levels of exposure at public wells
are lower: one estimated additional cancer case every 800 years. The
primary reasons for this finding of relatively low risk from groundwater
are that natural hydrogeologic protection and a number of regulatory
programs and voluntary actions in effect or soon to go into effect are
expected to limit most people's exposure. The IEMP estimates that no
more than about 25% of the people in the Santa Clara Valley are likely to
be exposed to groundwater contamination, compared to about 100% to air
contaminants and about 50% to trihalomethanes in surface water. It is
important to note that while our analysis of future groundwater contamination
involves substantial uncertainties, this conclusion of relatively low health
risks holds up under a wide range of alternative assumptions and appears
fairly solid.
The key hydrogeologic factor is the presence of an aquitard, or
clay layer, over much of the Valley protecting public drinking water
sources. While this clay layer has, for the most part, prevented
contamination near the surface from reaching deep drinking water supplies,
there is concern that such contamination could occur either through
abandoned wells that may function as conduits, or through faults in the
confining layer itself. The recent discovery of deep groundwater contamination
in Mountain View (which has not yet affected public drinking water wells)
provides the first strong evidence of contaminant transfer through conduit
wells in the Santa Clara Valley. This finding is consistent with the
IEMP analysis, which suggests that conduit well transport is likely to be
more significant than contamination through the major clay confining
layer itself, and that a number of public wells may eventually be affected
in this way.
One important set of regulatory programs estimated to reduce
groundwater contamination and human exposure are the local Hazardous
Materials Management Ordinances, which have become models for hazardous
materials control in other areas. These ordinances reduce contamination
at the source by requiring groundwater monitoring near underground tanks,
improved tank construction standards, and better chemical handling
processes. Other important regulatory and response actions include
clean-up actions at existing contamination sites, public drinking water
well monitoring for a broad range of organic chemicals (to be required
annually), and a policy of closing any public well contaminated above
state action levels. Voluntary actions taken by firms, such as underground
tank replacement and improved handling procedures, are also likely to
reduce future groundwater contamination.
-------
-25-
Analysis of the effectiveness of all of these programs indicates
that, in combination, they may reduce health risks by roughly one hundred
times (e.g., risks with these programs in place may be only 1% of what
risks would have been without them). Other programs, including efforts
to seal abandoned wells that may act as contaminant conduits, and efforts
to monitor and protect private wells, are also likely to reduce health
risks from groundwater contamination.
Exposure through the outdoor air and drinking water is direct, as
people take in pollutants through breathing or drinking. Contamination
affecting the San Francisco Bay and local surface streams, by contrast,
was judged to be only indirectly related to human exposure, largely
through body contact or fish consumption. Exposure through these routes
appears to be relatively small by comparison with air and drinking water
exposure.
Most hazardous wastes are exported from the Santa Clara Valley for
recycling or disposal elsewhere, and thus pose little local risk. Those
local risks we could identify from hazardous waste storage and handling
appear to be primarily through groundwater contamination, and were
analyzed under that exposure pathway. Accidental releases, such as those
resulting from transportation accidents, also have the potential to affect
soil and groundwater.
Although the IEMP explicitly examined a number of potential issues
of pollution transfer from one medium to another, none appeared to be
very significant in terms of public health risk in the Santa Clara Valley.
For example, the study estimated the possible toxic organic chemical air
emissions front sewage treatment plants, groundwater aeration/clean-up
sites, and sanitary landfills. Air emissions from these sources were
estimated to be fairly small in comparison to other sources of toxic
organic gases.
4. The toxic environmental contaminants posing the most significant
health risks in the Santa Clara Valley are, for the most part, the same as
those found in the other urban environments. A relatively small number of
toxic chemicals, including the trihalomethanes (primarily in drinking
water), and benzene, gasoline vapors, carbon tetrachloride, benzo(a)pyrene,
chromium and arsenic (primarily in air), account for about 92% of aggregate
cancer risk estimated in this study. National studies and data from
other areas show that estimated exposure levels in the Santa Clara Valley are
similar to, and in some cases lower than, ambient concentrations found in
other urban areas. It should be noted that the less-developed southern
Santa Clara Valley has a contamination problem typical of many agricultural
areas: high nitrate levels in the groundwater. A summary of estimated
cancer risks by pollutant is presented in Table Four.
As a class, volatile organic compounds account for the majority
(about 58%) of the cancer risk estimated in this study. Heavier organic
chemicals, such as benzo(a)pyrene, comprise an estimated 20% of aggregate
cancer risk. Metals and inorganic substances account for about 22% of
total estimated cancer risk.
-------
-26-
TABLE
ESTIMATED INCREASE
FOUR
IN CANCER INCIDENCE
IN SANTA CLARA VALLEY, BY POLLUTANT
POLLUTANT
(WEIGHT OF EVIDENCE)1
VOLATILE ORGANIC
CHEMICALS
Trihaloiethanes (B2)
Benzene (A)
Carbon Tetrachloride (B2)
Gasoline Vapor (B2)
1,1, Dichloroethylene
Perchloroethylene (B2)
Ethylene Oxide (Bl)
Other (A-C)
TOTAL, VOCs
ORGANIC PARTICULATES
BENZO(A)PYRENE
GROUP (B2)
METALS AND INORGANICS
Chroniun (A) **
Arsenic (A) ***
Cadmiun (Bl) **
Other (A-B2)
TOTAL, METALS
TOTAL, ALL CHEMICALS
STUDIED
All Exposure
Pathways
1.3
0.3 (0.3- 1.2)
0.2
0.1 (0 - 0.4)
0.04
0.04
0.03
0.03
2.1 (1.9- 3.5)
0.7 (0.01-1.3)
0.4 (0 - 4.0)
0.3 (0.2- 7.4)
0.07 (0.04-0.1)
0.03 (0 - 0.07)
0.8 (0.2-11.6)
3.6 (2.1-16.4)
POINT ESTIMATE OF ANNUAL INCREASE
IN CANCER INCIDENCE (Range)
Surface
Air Water Groundwater
<0.01 1.3 <0.01
0.3 (0.3- 1.2) <0.01 <0.01
0.2 <0.0001
0.1 (0 - 0.4)
0.04
0.03 <0.01
0.03
0.02 <0.01 0.01
0.7 (0.6- 1.9) 1.3 0.06
(0.03 - 0.3)
0.7 (0.01-1.3)
0.4 (0 - 4.0) 0 0
0.3 (0.2- 0.4) 0 (0-7) ***
0.07 (0.04-0.1) 0 0
0.03 (0 - 0.07) 0 0
0.8 (0.2- 4.6) 0 (0-7) 0
2.2 (0.8- 7.8) 1.3 0.06
(1.3 - 8.3) (0.03 - 0.3)
-------
-27-
FOOTNOTES TO TABLE FOUR
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS
OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND
POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM.
See text.
1 The weight of evidence of carcinogenicity for the compounds listed varies
greatly, fron very limited to very substantial. According to EPA's categori-
zation of levels of evidence of carcinogenicity, A = proven human carcinogen;
Bl = probable human carcinogen (limited human evidence); B2 = probable human
carcinogen (insufficient human evidence but sufficient animal evidence);
C = possible human carcinogen; D = not classifiable; E = no evidence.
* The weight of evidence identified for trihalomethanes is that for chloroform
only.
** Neither (hexavalent) chromium nor cadmium is thought to be carcinogenic
in water. See chapter 4.
*** There is some dispute over the carcinogenicity of arsenic in water. See
text. Arsenic exposure listed under surface water is for combined exposure
to surface water and groundwater.
-------
-28-
Arsenic risk in drinking water is a significant question mark in
this analysis. Arsenic accounts for as little as 0 to as much as 66% of
total estimated cancer risk, depending on assumptions about its toxicity.
Some evidence exists that arsenic in drinking water may cause a form of
skin cancer known as "Blackfoot Disease." Applying EPA's standard risk
estimation techniques, the IEMP would estimate up to seven additional
cancer cases a year from exposure to the levels of arsenic found in Santa
Clara Valley water (these levels are fairly low in comparison to those
found in many areas). Substantial disagreement exists as to the carcino-
genicity of lew levels of arsenic in drinking water, however, and EPA's
Office of Drinking Water believes that the levels of arsenic found in
Santa Clara Valley water are well within safe limits. This uncertainty
does not affect the estimate of lung cancer from airborne arsenic; the
evidence for this effect is much stronger.
The cancer risk from chromium in the air is another significant
uncertainty in this analysis. Monitoring data do not distinguish between
hexavalent chromium (thought to pose a risk of lung cancer) and other
forms (not considered carcinogenic). Depending on assumptions about the
proportion of chromium that is hexavalent, estimated cancer risk ranges
from none to four additional cases per year. Based on studies conducted
elsewhere and a tentative identification of local sources, this
study conservatively assumed that about 10% of airborne chromium was
hexavalent.
5. The pollution sources posing the most significant overall
health risks appear to be similar to high-risk sources identified in
other urban areas. However, the sources of some of the most important
environmental toxics are uncertain.Identifying sources is important for
risk management, since pollution control decisions aimed at reducing risk
must be directed at known sources of risk. Table Five presents a preliminary
breakdown of cancer risk by source type, making some assumptions about
the sources of chemicals whose origin is not well understood.
Most (about 77%) of the estimated risk from air exposure, particularly
for the toxic metals and organic particulates, is from sources that are
only tentatively identified. The AQMD does not maintain emissions
inventories for the metals and organic particulates, as it does for toxic
organic gases. Since consideration of control actions requires a knowledge
of the sources of risk, this report has identified the collection of data
on the sources and emissions of these substances as an important research
need. In Stage II, the AQMD, with assistance from EPA, plans to compile
a metals emissions inventory. The IEMP has done some preliminary analysis
of the possible sources of many of the substances of concern. However,
more definitive source identification would be required in most cases
before risk management control actions can be taken.
Preliminary analysis of likely sources suggests that the primary sources
of toxic chemicals in the air may be dispersed area sources, such as residen-
tial heating and motor vehicles. These sources appear to emit the bulk
of the benzene, gasoline vapors, benzo(a)pyrene, and metals that are
-------
-29-
TABLE FIVE
ESTIMATED INCREASE IN CANCER INCIDENCE
IN SANTA CLARA VALLEY, BY SOURCE TYPE
SOURCE TYPE
(EXPOSURE PATHWAY)
[WEIGHT OF EVIDENCE]
ESTIMATE OF ANNUAL
INCREASE IN CANCER
INCIDENCE
PERCENT OF
TOTAL ESTIMATED
CANCER INCIDENCE
Drinking Water
Disinfection
(Surface Water) [B2] *
1.3
36%
Fuel Combustion
for Residential
Heating
(Air) [A-B2]
0.63 - 1.1
18-31%
Motor Vehicles
(Air) [A-B2]
0.63 - 0.67
18-19%
**
Cement Plant
(Air) [A-B2]
0 - 0.5
0-14%
Unknown Sources/Back-
ground Contamination
(Air) [A-B2]
0.2
6%
Other Area Sources
(Air) [A-B2]
0.15
4%
Other Point Sources
(Air) [A-B2]
0.1
3%
Underground Industrial
Tanks
(Groundwater) [A-C]
0.05
1%
Underground Fuel Tanks
(Groundwater) [A-B]
TOTAL, ALL SOURCES
STUDIED
<0.01
3.6
100%
-------
-30-
FOOTNOTES TO TABLE FIVE
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND
ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. See text.
1 The weight of evidence of carcinogenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's categori-
zation of levels of evidence of carcinogenicity, A = proven human carcinogen;
Bl= probable human carcinogen (limited human evidence); B2 = probable human
carcinogen (insufficient human evidence but sufficient animal evidence);
C = possible human carcinogen; D = not classifiable; and E = no evidence.
Chlorofoon is considered a probable carcinogen. The upper end of this
range reflects the possibility that other THMs are also carcinogenic.
See chapter 4.
Source identification for residential heating and cement plant is
preliminary and uncertain. See chapter 3.
This point estimate derives from an estimated range of 0.2 to 7.4 annual
incidence. The point estimate does not include the potential risk from arsenic
in drinking water. There is substantial disagreement as to the carcinogenicity
of low levels of arsenic in drinking water. Conservative assumptions of
carcinogenicity, developed by EPA's Office of Research and Development, suggest
that the levels found in the drinking water in the Santa Clara Valley could
result in up to 7.2 additional cases per year. However, EPA's Office of Drinking
Water, which is responsible for setting standards, believes that low levels do
not pose risk.
-------
-31-
projected to cause most of the air toxics risk. Industrial point sources
do not appear to be significant contributors to aggregate risk, with the
possible exception of the coal-burning cement plant. This pattern of
many small and dispersed sources suggests that it may be difficult to
control major contributing sources so as to reduce risk.
Surface water risks are dominated by hazards posed by trihalonethanes
resulting from water disinfection. The presence of trihalomethanes in
treated drinking water involves a trade-off of one form of risk for another;
while chlorination introduces chloroform and other potential carcinogens
into drinking water supplies, it protects the population from the otherwise
much greater risk of infectious diseases such as cholera and typhoid.
Although discontinuing disinfection is not a viable option, there are
other disinfection methods that reduce the formation of trihalomethanes.
The Santa Clara Valley Water District (SCVWD) has recently implemented one
such treatment method, chloramination, at its two major local water treatment
plants. The IEMP projects that this change may reduce potential risks
substantially. In addition, the SCVWD has recently commissioned a study
of still other disinfection techniques, such as ozonation, that might
reduce risks further. Given the apparent importance of these chlorinated
organic chemicals relative to other sources of toxic health risk, such
analysis of alternatives may be appropriate, both in the Santa Clara Valley
and elsewhere.
"Background" contamination is contamination not linked to any known
current source. Such contamination may be from natural sources, such as
minerals in the soil, or from prior agricultural or industrial activities.
Background contamination, largely from persistent levels of carbon
tetrachloride in the air, is estimated to account foe about 5% of total
cancer risk from sources and pollutants studied. However, if pessimistic
assumptions about the carcinogenicity of arsenic in drinking water are
correct, the risk from background contaminants increases to well over
half of all estimated cancer risk in the Santa Clara Valley.
The major groundwater contamination sources examined, underground
fuel and solvent tanks, are estimated to account for about 1-2% of the
total cancer risk among sources examined.
6. Some individuals, who live near pollution sources or are highly
exposed for other reasons, face toxic health risks that appear to be
significantly higher than average. Estimates of potential risk to these
most-exposed individuals (MEIs) are shown in Table 6.
In contrast to the estimates of low overall risks from groundwater
contamination, people drinking from private wells appear to be vulnerable
to potentially significant levels of exposure and risk as a result of
leaks from underground tanks. Individuals who obtain drinking water from
private wells are more vulnerable to risk because these wells are shallow
(and thus not protected from surface contamination by an intervening clay
layer) and are not typically monitored. Risks to individuals who may be
-------
-32-
TABLE SIX
ESTIMATED HEALTH RISK TO MOST-EXPOSED INDIVIDUALS
IN SANTA CLARA VALLEY
EXPOSURE PATHWAY
& SOURCE TYPE
INCREASED
LIFETIME
CANCER RISK
iCHANCES IN
A MILLION)
POLLUTANT
(WEIGHT OF EVIDENCE)2
POTENTIAL
NON-CANCER
HEALTH EFFECTS3
AIR;
TRAFFIC 300
INTERSECTIONS
HOSPITALS 200
PHARMACEUTICAL 100
MANUFACTURER
COMPUTER EQUIPMENT 40
MANUFACTURER 4
INDUSTRIAL 30
FACILITY 4
Benzene (A)
Benzo(a)pyrene(B2)
Cadmium (Bl)
Ethylene Dibrcmide(B2)
Ethylene Oxide (Bl)
Ethylene Oxide (Bl)
Benzene (A)
Methylene Chloride (B2)
Methylene Chloride (B2)
Benzene (A)
Blood, fetal
Blood, fetal
FUEL PIPELINE
DRY CLEANERS
SEWAGE TREATMENT
PLANTS 4
GAS STATION PUMP
GROUNDWATER
AERATION 4
20
10
5
0.2
Benzene (A)
Perchloroethylene (B2)
Chloroform (B2)
Benzene (A)
Methylene Chloride (B2)
Perchloroethylene (B2)
Trichloroethylene (B2)
Benzene (A)
Gasoline Vapors (B2)
Methylene Chloride (B2)
Trichloroethylene (B2)
-------
-33-
TABLE SIX (cont.)
ESTIMATED HEALTH RISKS TO MOST-EXPOSED INDIVIDUALS
IN SANTA CLARA VALLEY
EXPOSURE PATHWAY
& SOURCE TYPE
INCREASED
LIFETIME
CANCER RISK
(CHANCES IN
A MILLION)1
POLLUTANT
(WEIGHT OF EVIDENCE)2
POTENTIAL
NON-CANCER
HEALTH EFFECTS3
GROUNDWATER;
UNDERGROUND TANKS 5
(AT PRIVATE WELLS)
20,000
1,1 Dichloroethylene (C)
Vinyl Chloride (A)
Perchloroethylene(B2)
Ethylene Dibronide (B2)
Methylene Chloride (B2)
Chloroform (B2)
Benzene (A)
Trichloroethylene (B2)
1,1,1-Trichloroethane
Liver, kidney,
Liver, kidney,
cardiovascular
Liver, fetal
Liver, neurobehavioral
Liver, neurobehavioral
*
FERTILIZER, SEPTIC
TANKS
Nitrates
Methemoglobinemia
(Blue baby syndrome)
SURFACE WATER;
DRINKING WATER
TREATMENT
BACKGROUND
100*
Trihalonethanes (B2)
0 - 7,000*** Arsenic (A)
SOUTH SAN FRANCISCO BAY;
SHRIMP CONSUMPTION 7 0 - 6,000*** Arsenic (A)
MUSSEL CONSUMPTION 7 20 - 200 PCB (B2)
Chlordane (B2)
DDT (B2)
STRIPED BASS
CONSUMPTION 7
80 - 16,000
PCB (B2)
Cadmium
Mercury
Liver, neurobehavioral,
kidney, reproductive
Kidney, reproductive,
liver, birth defects
8
-------
-34-
FXX)TNOTES TO TABLE SIX;
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESTIMATES OF INDIVIDUAL RISK AND DISEASE INCIDENCE
ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CON-
SERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND ARE MORE LIKELY TO
OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. See text.
1 Except in the case of underground tanks at private wells, estimated cancer
risk is for all pollutants combined frcm given source. For underground
tanks, estimate is for pollutant posing the greatest cancer risk. In each
case, pollutants for each source are listed in decreasing order of estimated
cancer risk.
2 The weight of evidence of carcinogenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's
categorization of levels of evidence of carcinogenicity, A = proven human
carcinogen; Bl = probable human carcinogen (limited human evidence);
B2 = probable human carcinogen (insufficient human evidence, but sufficient
animal evidence); C = possible human carcinogen; D = not classifiable;
E = no evidence.
3 Non-cancer health effects are reported only if exposures are above estimated
thresholds for such effects.
4 If TCA is assumed to be carcinogenic, total estimated cancer risk is at or
slightly above level presented.
5 Estimated impacts are for "high" release, base case; see chapter 4.
6 Estimated taste and odor threshold is very slightly below the estimated
threshold for blood effects.
7 Estimated risks for fish consumption are for a hypothetical individual
who regularly consumes contaminated fish or shellfish caught in the South
Bay. Assumed consumption is 5 to 52 pounds of fish per year. Note that
the IEMP has no actual data on the number of people eating fish from the
South Bay, although we believe that number is small.
8 Estimated exposure value for mercury in striped bass is just slightly
under the lowest estimated human threshold.
* The IEMP conducted sensitivity analysis on TCA for possible fetal effects.
See footnote to text.
** Risk for system with highest estimated average risk; some individuals
may be exposed to higher risks.
*** Considerable controversy exists as to the carcinogenicity of arsenic by
ingestion. See text.
-------
-35-
highly exposed to industrial chemicals in their private wells were
estimated to be potentially higher than risks fron any other source
examined. The estimated risk to the most-exposed individual drinking
from a private well is quite uncertain and should not be interpreted
literally. However, the potential vulnerability of this group is clear,
and this is the important conclusion for risk management. Current efforts
by the County Health Department to monitor some private wells appear to
be a useful first step in addressing this problem.
Persons living near a highly congested intersection were estimated
to face the highest individual cancer risk frcm exposures to toxic air
contaminants such as benzene. The risk facing individuals living near
hospitals and exposed to the stenlant ethylene oxide (ETO) was estimated
to be nearly as large. These potential risks from high exposures to air
toxics are substantially lower than the estimated risks for highly exposed
individuals at private wells. The comparatively high estimated risk
near intersections reinforces the importance of vehicles as a source of
air toxics risk - both to the general populace and to highly exposed
individuals. Ethyleno oxide from hospitals, on the other hand, is not
projected to be a major source of risk for most people but nevertheless
appears to pose comparatively high risks near the source. Because of
uncertainties about ETO emissions, and the possible reactivity of the
chemical once released, estimated emissions and exposure levels should
be confirmed before control actions are taken. Since use of ETO as a
disinfectant is not unique to the Santa Clara Valley area, this finding,
if confirmed, may have implications for other areas.
Although we lack actual consumption data, estimated risks to a
hypothetical individual regularly consuming significant quantities of
contaminated fish or shellfish caught fran the South San Francisco Bay
appear to be fairly high. Concentrations of PCBs, pesticides, mercury,
and other metals in shrimp, mussels, and striped bass may pose a significant
risk. Possible effects include cancer, neurobehavioral, reproductive,
kidney, and liver effects (estimated thresholds for non-cancer effects
are exceeded only under a "high" consumption estimate of one pound per
week of contaminated fish). We stress that these exposure estimates are
conservative, and that we have no data on the number of people consuming
contaminated fish from the South Bay. Nevertheless, these estimates
suggest that regular consumption of fish or shellfish from the South Bay
may pose significant health risks. This finding is consistent with a
health advisory issued by the state Department of Health Services, warning
pregnant women not to eat striped bass.
7. One of the more important implications of this analysis is
that groundwater contamination may be an economic and natural resource
issue as well as a risk issue. IEMP estimates of future risk depend on
many actions that we assume will be taken in the future. For example,
the study assumes that public drinking water wells will be closed when
contaminated above action levels and replacement supplies obtained; it
also assumes that the Hazardous Materials Management Ordinances will be
implemented, although this has not yet fully occurred. While IEMP projects
-------
-36-
that these actions will be largely successful in controlling risk, they
could be extremely expensive. The direct economic costs of contamination
prevention and response include the costs of tank replacement, clean-up,
monitoring and well closure. In addition, groundwater contamination
causes a potentially significant indirect natural resource cost: the
loss of clean, local groundwater.
The IEMP analysis illustrates the difference between drinking water
health risk and groundwater resource iitpacts. Under the rather pessimistic
assumptions used in this study, health risks to people drinking groundwater
from public wells are projected to be comparatively stall. Yet, about 78
public wells serving about 200,000 people are projected to be affected by
fuel or industrial contamination, with 20% to 30% of the wells contaminated
above state action levels.
Contamination above action levels requires well closure or treatment.
In some cases, contamination below action levels has also led to removing
a well from service. Clean-up of contaminant plumes can also have a
significant impact on the groundwater resource, as large quantities of
groundwater are pumped, cleaned and discharged to the Bay. This water
must be replaced with recharge water imported from the Sacramento Delta.
While the IEMP estimates of the number of wells likely to be affected
are intentionally pessimistic, they clearly indicate the importance of
examining the natural and economic resource 'impacts, as well as the
health effects, of groundwater contamination and programs to address it.
Thus, the low aggregate risk estimates presented in this draft report
do not imply that groundwater contamination is not an important environmental
management issue. Despite the comparatively low estimated aggregate
risks, it may be appropriate to assess groundwater control and treatment
options in terms of their potential risk, cost, and resource impacts.
8. This study identified many scientific uncertainties and data
gaps that may be appropriate research priorities for regulatory agencies
or others. A few of the most important include:
0 Hydrogeology: A better understanding of Santa Clara Valley
hydrogeology - in particular the effectiveness of the major clay
confining layer or aquitard - would improve the ability to protect
the groundwater resource effectively.
0 Pollutant transport and transformation: In particular, better
understanding of the speed with which fuel contaminants degrade
after release into the environment is critical to determining the
importance of leaking fuel tanks as a groundwater contamination
source.
0 Monitoring data: Two of the more critical uncertainties in this
Stage I analysis - levels of organic chemicals in the ambient air
-------
-37-
and in private wells - are being addressed by local agencies.
Local data on organic particulates in air would be valuable also;
the IEMP plans to sponsor the collection of such data as a part
of Stage II.
Source data: Better information on sources of metals and organic
particulates in air is needed to assist in the development of risk
management strategies. The Bay Area Air Quality Management District,
with EPA support, will be compiling a metals emissions inventory
in Stage II.
Non-Cancer effects: Development of a method of estimating possible
disease incidence for effects other than cancer would allow a
more complete analysis of toxic health risks. Some key chemicals
of concern in the Santa Clara Valley have been identified in this
report. These issues are being pursued within EPA and by scientific
peer review groups.
Next Steps
This Stage I Report for the Santa Clara Valley Integrated Environmental
Management Project presents the results of the lEMP's comparative analysis
of toxic environmental health risks. Potential health effects were
analyzed, and exposure pathways, pollutants and sources compared in
terms of health risk. A draft of this report has been reviewed widely
by EPA's two advisory committees, the Intergovernmental Coordinating
Committee and the Public Advisory Committee, and by other interested
agencies, scientists and individuals. It is now undergoing scientific
peer review by a group of scientists at Rutgers University.
The Stage I Report findings are intended to provide the basis for
Stage II of the IEMP, which will focus on managing risks: identifying
priority issues, analyzing control options for dealing with those problems,
and implementing solutions. Stage II will also expand and improve upon
the problem definition developed in Stage I.
EPA, in consultation with its IEMP advisory committees, has developed
a Stage II workplan to guide the project's future work. This workplan
identifies risk management priorities, taking into account public concerns
and ongoing programs. It outlines research priorities, analyses of
pollution control options, and a management strategy that EPA and its
local partners hope will lead to discussions and actions that protect
public health and the environment more effectively.
-------
CHAPTER ONE
INTRODUCTION
-------
CHAPTER ONE
INTRODUCTION
I Introduction 1-1
II The Inception of Integrated Environmental Management 1-2
III The Concepts of Risk Assessment and Risk Management 1-3
IV The Uses of Pisk Assessment and Risk Management 1-4
V The Santa Clara Valley IEMP 1-5
A Stage I: Evaluating Toxic Risk 1-6
B Stage II: Developing Risk Management Strategies 1-8
VI Scope of the Santa Clara Valley IEMP 1-8
A Geographic Boundaries ,«. 1-8
B Imports and Exports of Toxics from the Study Area 1-8
C Analytic Scope 1-10
-------
CHAPTER ONE
INTRODUCTION
In the 1980s, effective management of toxic risk has risen to the top of
the environmental agenda. An increasingly concerned public is demanding safe
drinking water, cleaner air, and control over the risks posed by toxic materials.
In response, federal, state, and local governments have been seeking new ways
to control toxic environmental risks by combining risk assessment with risk
management.
As part of this pursuit of new approaches to environmental policy, the
U.S. Environmental Protection Agency (EPA) in early 1984 began a special project
in Santa Clara County, California an area better known as "Silicon Valley."
This effort, termed the "Integrated Environmental Management Project," or IEMP,
is an innovative attempt to manage complex toxic environmental problems. The
IEMP applies the best scientific knowledge and rrian^jument skill? to a broad
set of environmental and public health issues.
EPA has designed the project to identify and define risks to public health
posed by toxic snvironmental contaminants; to compare these risks to one
another to set research and management priorities; and to develop approaches
to manage such risks more effectively. EPA is cooper i':c ing closely with stiate
and local governments, industry, envirormental groups, universities, and
members of the public. The Santa Clara Valley IEMP is part of a national
effort including earlier EPA studies in Philadelphia and Baltimore.
This report presents the results of Stage One of the Santa Clara Valley
IEMP: a broad analysis of public health risks from toxic chemicals in the
area's air and drinking water. We compare risks among major exposure pathways
(air, drinking water from groundwater supplies and from surface water supplies);
for 30 different pollutants; and from various sources of toxic contamination.
This Stage One risk assessment is essentially a screening exercise to set
research and management priorities among toxic environmental problems. Although
we estimate health risks from toxics, the Santa Clara Valley IEMP is not designed
to produce unequivocal statements about the absolute risks attributable to
these substances. Analytical and data uncertainties in this study are simply
too great to allow reliable projections of such absolute health impacts.
Nevertheless, this risk assessment provides estimates of the approximate magnitude
of several environmental problems, and compares these problems against one
another. The IEMP analysis thereby allows us to set research and management
priorities for addressing public health risks.
In Stage Two of the IEMP, EPA will further assess risks identified in
Stage I, where appropriate, and analyze options designed to manage those risks
identified as priorities. These analyses will encompass possible control
strategies, their effectiveness in reducing risk, and their costs. Stage Two
will be unable to cover all local toxic environmental issues; therefore, EPA
will work with state and local agencies to address the most important issues in
terms of the risks they pose, the possibilities for taking effective action,
the extent of public concern, and the cost-effectiveness of controls. In sum,
-------
1-2
Stage One has assessed a set of toxic environmental risks in the Santa Clara
Valley (risk assessment); Stage Two will focus on ways to better assess and
manage those risks (risk management).
THE INCEPTION OF INTEGRATED ENVIRONMENTAL MANAGEMENT
Traditionally, the U.S. has responded to its environmental problems by
passing new laws, each focused on one environmental medium: a Clean Air Act, a
Clean Water Act, a Safe Drinking Water Act, and so on. EPA has organized
itself accordingly: in major program offices responsible for air pollution
control, water pollution control, regulation of the disposal of hazardous
wastes, or regulation of pesticides and toxic substances. Most state and local
environmental agencies mirror this Federal format.
In the early 1970s, EPA concentrated on the most dramatic, obvious forms
of pollution: the smog and dust over our cities, and the oxygen depletion in
our rivers and lakes. The U.S. has made great progress in controlling these
kinds of pollution, though problems remain. EPA has begun focusing on the
more difficult issues involving toxic substances. Even in very small amounts,
these chemicals may endanger public health or biological systems. Of the
65,000 chemicals used commercially, EPA has focused on some 400 due to their
potency (toxicity), their persistence, or their pervasiveness in the environment.
The job of reducing risks from toxic chemicals is larger and more complex
than anyone expected some years ago. Problems we once regarded as solved
turn out not to be solved. Each new scientific revelation seems only to
lengthen the agenda of possible actions. Sophisticated monitoring devices
can detect minute amounts of toxic pollutants, outdistancing our knowledge of
the health effects attributable to exposure to such low concentrations.
Moreover, toxic substances move through the environment in complex
patterns. Our ability to monitor and control these movements from the initial
source of contamination to the eventual public exposure is constrained by our
limited comprehension of migration pathways, chemical transformations, and
abatement and control techniques. The chemicals may be found in our drinking
water supplies, and may be present in the air we breathe. If placed on the
ground they may migrate either to the air (volatilize) or to the water (leach).
Largely because of our historical single-issue approach, environmental
regulation often results in the transfer of pollutants from one medium or
location to another. Industrial wastewater treatment, for example, shifts
pollutants from water to land. It produces a relatively clean effluent acceptable
for discharge to water, plus a residual sludge containing many of the original
pollutants in concentrated form; the sludge is then disposed of on the land.
Similarly, air pollution control devices such as scrubbers remove pollutants
from smokestacks, but their sludges containing inost of the pollutants then
require land disposal.
Laws and regulations to control environmental contamination vary in their
objectives, stringency, and approach, resulting in inconsistencies across
programs. The Clean Air Act, for instance, sets national air quality goals
while the Clean Water Act requires dischargers in different industries to meet
technology standards regardless of water quality. The various statutes call
-------
1-3
for differing enphases on protecting human health, considering economic costs,
and so on.
In 1982, EPA developed the concept of integrated environmental management
as one response to the growing challenge of toxic pollution, and created a
new Integrated Environmental Management Division in its Office of Policy Analysis.
The IEMP focuses on toxics because of their potential health effects and because
of the more rudimentary level of existing control when compared to conventional
pollutants. Integrated environmental management refers to evaluating and
controlling public health risks from various pollutants, pathways, and sources,
from a multi-media perspective. This approach assures that we account for
inter-media transfer issues, and allows us to set consistent priorities so that
we can choose better risk reduction strategies.
The EPA effort has had two basic parts: industry studies, and geographic
analyses. The IEMP studies of the iron and steel, oil refining, and pulp and
paper industries have concentrated on understanding the range of pollutants
they generate. EPA identified alternative pollution control strategies to reduce
the risk to public health and the environment from these industries.
The lEMP's geographic projects are intended to provide ways for local,
state, and federal governments to manage risks from toxic pollution in an
areawide setting. EPA traditionally sets national standards for industrial
categories that contribute to pollution problems. While this has been a useful
and powerful approach, it results in overcontrol in some areas and undercontrol
in others. The IEMP departs from this method by examining toxic contamination
problems, and possible solutions, in a particular geographic area. The Santa
Clara Valley IEMP has built upon the initial geographic projects in Philadelphia
and Baltimore.
THE CONCEPTS OF RISK ASSESSMENT AND RISK MANAGEMENT
The concepts of risk assessment and risk management have captured the
attention of EPA's leaders during the past several years, and have played a large
role in developing the concept of integrated environmental management. Since
1983, EPA has increasingly used risk assessment to identify and measure toxic
problems, and to compare them to one another to establish priorities ("Risk
Assessment and Management: Framework for Decision Making," EPA, December,
1984).
In essence, risk assessment allows one to evaluate the health effects
from exposures to toxic contaminants by estimating the risk from that exposure.
Quantitative estimates allow us to compare risks from different chemicals,
pathways, and sources.
In the IEMP, we define individual risk as the increased probability that
an individual exposed to one or more chemicals will experience a particular
adverse health effect during the course of his or her lifetime. Risk to the
population is the expected increased incidence (numbers of cases) above the
background rate in an exposed population.
While risk assessment is rooted in scientific principles and knowledge
of toxicology, epidemiology, and environmental exposure, it is inherently
-------
1-4
imprecise necessary data are often limited, if available at all. Consequently,
the results of a risk assessment depend heavily on assumptions about physical
relationships and biochemical transformations, among other factors.
Despite its uncertainties, risk assessment provides a useful framework
for comparing different environmental risks. As long as one understands its
uncertainties, risk assessment can be useful in screening problems and establishing
priorities for risk management. In a 1983 speech to the National Academy of
Sciences, former EPA Administrator William Ruckelshaus distinguished between
risk assessment and risk management:
Scientists assess a risk to find out what the problems are. The process
of deciding what to do about the problems is risk management. The second
procedure involves a much broader array of disciplines, and is aimed
toward a decision about control. Risk management assumes we have assessed
the health risks of a suspect chemical. We must then factor in its
benefits, the costs of the various methods available for its control, and
the statutory framework for decision.
THE USES OF RISK ASSESSMENT AND RISK MANAGEMENT
Risk Assessment and Risk Management Help Set Priorities
As already noted, there are thousands of chemicals in commerce, and an
unknown number of contaminants and unintended by-products. These pollutants
are proper targets for Federal, state, or local regulation if they pose significant
risks to health or the environment. Too often, however, regulatory priorities
are set in response to public pressures to address particular environmental
issues of the day, rather than on the basis of systematic analysis of risks and
costs. This would not be a problem if the issues of greatest public concern
also posed the greatest risk. At times, however, perception and risk diverge
widely. Vfe cannot know which problems are the most serious in terms of public
health without some kind of systematic risk assessment. Risk assessments help
identify and set priorities among those pollutants posing significant risks.
Some priorities will always be set by Congress or state legislatures, or
through the press of emergencies. However, a consistent analytic approach is
necessary to set an agenda for effective risk management in Santa Clara County,
and more broadly. Analyzing alternative strategies for reducing risk provides
a useful basis for such an approach. Time and resources are best spent
regulating those toxic chemicals to which people are actually exposed and
for which practical controls are possible. It makes no sense to spend limited
resources- regulating chemicals, even highly toxic ones, to which no one will
ever be exposed, or for which there is no capability to impose additional
controls.
Risk Management Allows Balanced Decision-making
Scientific evidence suggests that toxic chemicals can cause cancer and
other diseases. We are exposed to a complex mixture of chemicals through
air, water, and food. Given the level of public anxiety, extreme points of
view often emerge when regulatory actions may involve substantial health or
-------
1-5
economic impacts in the face of scientific uncertainty. This situation may
polarize debate and bring public policy to an impasse.
Consistent use of risk assessment and risk management can assist by explaining
the scientific basis for the risk estimates, including the confidence we have
in such numbers; by placing the risk reduction expected from regulation in
context with other risks and opportunities for their reduction; and by explaining
the broader social values which affect risk management. Moreover, communication
of policies based on well-articulated scientific principles is perhaps the most
important element in creating a strong base of public understanding of EPA,
state, and local regulatory actions, thereby building government credibility.
Risk Assessment and Risk Management Produce More Efficient Policies
The existing patchwork of government authority to control pollution needs
to be woven together more coherently, beginning at the analytical level and
continuing through to regulatory decisions. Differences in the ways public
agencies manage risk remain, but risk management can use administrative flexibility
to make more efficient use of the resources available to reduce risk. Enhancing
efficiency through risk management means examining the regulatory options and
selecting those that reduce risk the most for any given level of resources.
THE SANTA CLARA VALLEY IEMP
In 1983, EPA decided to conduct its third geographic IEMP. in California's
Santa Clara Valley some 50 miles south of San Francisco. Better known as
"Silicon Valley," this area is one of the fastest-growing in the country. Since
1951, electric and electronic equipment manufacturing has emerged as the area's
largest industrial employer, increasing from about 3,000 workers in 1951 to
over 100,000 today. Rapid growth in the Santa Clara Valley is expected to
continue, particularly in San Jose and in the southern parts of the Valley.
Population growth and industrialization have brought certain environmental
problems; some of these are common to urban areas, while others are products
of local factors. Air pollution, primarily from automobiles, can be a problem.
Santa Clara County exceeds Federal standards for ozone (a measure of smog) and
carbon monoxide several days each year. Sources of toxic air contaminants,
such as organic chemicals and heavy metals, include automobiles, dry cleaners,
industrial sources, and other combustion sources. No Federal or state standards
exist for most air toxics, which are analyzed in the IEMP-
Drinking water is subject to contamination from several sources. About
half of the Valley's water supply comes from the Delta area or from the Sierras.
This water may contain various minerals and metals, and is subject to pesticide
contamination. Disinfection of this water creates trihaloroethanes, which may
be carcinogenic. Although trihalomethanes pose risks, disinfecting surface
water is necessary to reduce risks from bacteriological contamination.
In addition to imported surface water, local groundwater supplies about half
of the Santa Clara Valley's drinking water. In recent years, the discovery of
groundwater contamination caused by leaks and spills from underground industrial
and fuel tanks and other sources has generated widespread public concern.
Other potential sources of groundwater contamination include municipal landfills,
sewage pipes, and pesticide use. Nitrates, the result of agricultural and
-------
1-6
septic tank practices, are above standards in some parts of the South County
area.
Managing municipal solid wastes poses a significant challenge; so does
managing hazardous waste generated by local industries. Sewering industrial
liguid wastes requires continuing control. In addition, accidental releases
of hazardous materials, such as spills on highways or large releases at
manufacturing facilities, are also of concern.
The decision to conduct an IEMP in the Santa Clara Valley followed extensive
discussions by EPA with state and local officials, industry representatives,
and public interest groups. The impressive local response to the groundwater
problem associated with underground storage tanks was of special interest to
EPA. In 1982-83, a coalition of local and state government agencies, industry
representatives, and environmental group leaders responded to the discovery
of groundwater contamination by enacting a new Hazardous Materials Management
Ordinance (HMMO) to regulate the underground storage and ultimate disposition
of industrial chemicals. The IEMP thus could build upon an active coalition
of local interests.
In addition, local and regional environmental management and planning
agencies in the area - the San Francisco Regional Water Quality Control
Board, the Santa Clara Valley Water District, the Bay Area Air Quality Management
District, the Association of Bay Area Governments, and the Santa Clara County
Health Department, among others had strong professional reputations and
access to significant sources of information on toxics. EPA has been working
closely with these agencies, and has established a new relationship among
federal, state, and local governments, with cooperation supplementing the
traditional delegation and oversight.
EPA established an Intergovernmental Coordinating Committee (ICC), consisting
of county supervisors, mayors, city council members, and members of the boards
of regional regulatory agencies to advise EPA on the IEMP. This committee
meets regularly with EPA to discuss the progress of the project, to suggest
modifications, and to ensure that the project remains useful to state and local
government.
While improved ties among various agencies are important to the lEMP's
success, so is the active involvement of industry and the public. Therefore,
EPA has been working actively with industry, the environmental community,
universities, and technical staff of local agencies through a Public Advisory
Committee (PAC), which has provided extensive useful comments on the analytic
components of the Project.
Stage I; Evaluating Toxic Risk
To gain greater knowledge about toxic environmental pollution in the
Santa Clara Valley, and the attendant risks to public health, the IEMP has
compiled existing data, collected new information, and analyzed exposure and
risk from toxics from many sources in all environmental media. This effort-
has concentrated on about thirty chemical substances. This Stage I report
presents the results of this risk assessment.
-------
1-7
Risk assessment combines estimates of the toxicity, or potency, of substances
with estimates of human exposure to those substances. Risk assessment can be
divided into the four specific steps mentioned below and discussed in detail in
Chapter 2.
1. Hazard Identification
We first examine the evidence about a chemical to determine if enough
there may be a link between that chemical and health effect, and thereby to
determine if we should consider the chemical a potential environmental hazard.
2. Potency Assessment
If evidence suggests that a pollutant may be hazardous, we then guantitatively
assess the adverse health effects associated with different levels of exposure
to that chemical. This assessment is expressed as that particular chemical's
potency. These potency values vary widely for different chemicals, health
effects, and pathways of exposure.
3. Exposure Assessment
We then estimate the numbers of people in the Santa Clara Valley that
may be exposed to these chemicals at particular concentrations in air or
drinking water. Chapters 3 (air) and 4 (drinking water) present details on
the lEMP's exposure assessments.
4. Risk Estimation
With the first three steps completed, we then estimate risks to individuals
and to populations as a whole by combining our potency and exposure estimates.
The intent has been to answer the following kinds of questions: What
are the primary sources of these different pollutants? Through what environmental
pathways do these pollutants typically move? How many people are likely to
be exposed to these pollutants, and in what amounts? What is the probability
of different health effects from these exposures? We can then use this
assessment of risks to help set priorities for risk management.
Given the uncertainties about the mechanisms of chronic health effects,
and given a clear preference to ensure an extra margin of safety in the face
of such unknowns, the IEMP risk assessment incorporates a series of conservative
assumptions. These assumptions do not result in a worst-case analysis; we
preferred to construct a more plausible but still conservative analysis.
Since Stage I is primarily a screening process, we have been more willing
to accept significant limitations than EPA would if it were conducting an
assessment for a national regulation. We tolerate a higher degree of uncertainty
in our toxicology and exposure estimates, and in our subsequent risk estimates.
Resource and time constraints and the breadth of the effort prevent us from
analyzing individual issues in as much detail as might otherwise be possible.
We have attempted to strike a balance between the desire for exhaustive and
definitive analysis, and the need for timely results, useful to decision
makers, at a reasonable cost.
-------
1-8
Stage II: Developing Risk Management Strategies
This comparison of risk provides a basis for developing a risk management
strategy. In Stage II, the IEMP will further assess risks, where appropriate,
and analyze different risk management controls in terms of risk reduction and
cost. Local, state, and federal agencies, and the private sector, will then
make decisions on managing risk. These parties are already making decisions
on the management of risk in the Santa Clara Valley. The IEMP framework is
intended to help in making these decisions more effectively.
SCOPE OF THE SANTA CLARA VALLEY IEMP
Any analytic effort must define its scope: what it does and does not include.
Such decisions involve trading off competing objectives and are inherently
somewhat arbitrary. We describe below the geographic and analytic boundaries of
this analysis, and the reasons for selecting those boundaries.
Geographic Boundaries
Since environmental pollution does not respect the boundaries defining our
political jurisdictions, selecting the physical boundaries for a study such as
the IEMP requires a balancing of several competing aims. Our selection of the
lEMP's geographic boundaries resulted from three objectives: definition of self-
contained air and groundwater basins; maintenance of a manageable analytic
scope; and limitation to a reasonable number of political jurisdictions.
The principal study area for the Santa Clara Valley IEMP covers all of the
urbanized, industrialized portions of Santa Clara County (see map). The northern
boundary runs along the border between Santa Clara County and the counties to the
north (San Mateo and Alameda). The study area is bounded on the east and west by
the ridges on either side of the Santa Clara Valley: the Santa Cruz Mountains to
the west, and the Diablo Range to the east. The southern border is an imaginary
east-west line crossing the Santa Clara Valley just south of Gilroy. This study
area includes the municipalities of Campbell, Cupertino, Gilroy, Los Altos, Los
Altos Hills, Los Gates, Milpitas, Monte Sereno, Morgan Hill, Mountain View, Palo
Alto, San Jose, Santa Clara, Saratoga, and Sunnyvale, and some of the unincorporated
areas of Santa Clara County.
Imports and Exports of Toxics from the Study Area
We defined the study area's boundaries to minimize the importance of the
movement of toxics into and out of the Valley. Our analysis focuses on exposure
and risk to people within the Santa Clara Valley, and the pollution sources of
concern are also generally within the Valley. We have attempted to include the
effects of pollutants imported into the Valley when these pollutants pose
significant health risks. For example, in Chapter 4 we examine surface water
imports of pesticides, organics, and metals via the Hetch Hetchy and South Bay
Aqueducts.
IP addition, pollutant exports from the Santa Clara Valley affect other
areas. For example, the Santa Clara Valley exports most of its hazardous
waste to landfills outside of the County. We have not studied pollutant exports
in our risk analysis.
-------
1-9
Figure 1-1: Santa Clara Valley IEKP Study Area (Shown by -t
^>;;: COUNTY :,
SANTA CLARA
COUNTY
-V.'';, :^f'-'--\J%''-':}!'y^^
jj&tf:.:-; ' w^:^^-^'; '^" "''' .''-- i;/
-------
1-10
Analytic Scope
Although the IEMP is more comprehensive than most EPA analytic efforts, we
are not examining risks from all pollutants. We decided to focus on "toxic"
pollutants (i.e., those capable of causing adverse health effects at relatively
low levels of exposure) because such pollutants are relatively less studied and
less controlled than more "conventional" pollutants, such as ozone precursors in
air and oxygen-depleting substances in water. Conventional pollutants may also
affect health, although generally not at the low levels at which more toxic
pollutants may be harmful.
We did not study all toxic pollutants in the Santa Clara Valley in detail.
We decided which pollutants to study largely on the basis of the available
data on their toxicity and their presence in the environment (see the the
discussion on pollutant screening in chapter 2). We have tried to identify
and study those toxic pollutants which we believe are most likely to pose
health risks in Santa Clara Valley. Nevertheless, it is likely that some of
the pollutants we are not studying because of a lack of data on their presence
or toxicity may also be toxic.
We did not examine in detail every exposure pathway for toxics in the Santa
Clara Valley. Our focus is on exposure to contaminants in outdoor air and drinking
water (both surface water and groundwater). We are not considering exposures
fron voluntary intakes such as cigarette smoking and diet, nor exposures to
indoor residential air, occupational toxics, or toxic residues in food.
While exposure from smoking and diet are clearly significant, the authority
of any governmental entity to regulate them is, at best, unclear. We chose not
to examine exposure from toxic residues in food because of the national rather
than local scope of the food distribution network. Occupational exposure is
essentially outside of EPA's jurisdiction, and we lack the resources to address
it adeguately. (The IEMP has, however, commissioned a scoping study on occupational
exposures.) Although indoor residential air may arguably fall within EPA's
purview, and recent studies suggest that indoor air may contribute more to
exposure from certain substances than does outdoor air, such exposures vary
enormously among households in the same neighborhood, making an assessment
extremely difficult. Finally, given our limited resources and previous EPA
work oh criteria air pollutants, we judged that we could contribute more by
concentrating on less well understood and less regulated toxic chemicals.
In summary, we have omitted a detailed study of risks from each of these
types of exposures for one or more of the following reasons:
0 given our resource constraints, we could contribute only marginally to an
understanding of the problem;
0 it is not clear what regulatory authority EPA possesses to address exposures
in these areas; or
0 the problems are inappropriate for lEMP's methods of geographic analysis.
Expanding the lEMP's scope to address any of these issues would require
substantial effort, focusing on further problem definition rather than analysis
of control options. While such efforts may be important, performing further
-------
1-11
problem definition work in the IEMP would necessarily reduce our ability to
study control options in those areas where we have examined risks. This
would reduce the lEMP's effectiveness as a management tool.
Although the IEMP is necessarily limited in scope, it is important to
place the risks for ambient air and drinking water in context by conparing
them to other types of risks County residents may face, and to exposures in
other geographic areas. We attempt to make such comparisons, when appropriate,
in the Stage I analysis.
-------
CHAPTER TWO
GENERAL METHODOLOGY
-------
CHAPTER TWO
GENERAL METHODOLOGY
I Introduction 2-1
A What Risk Assessment is Not 2-1
II Overview of Methodology .2-2
III Toxicological Potency Assessment 2-7
A Pollutant Selection 2-8
B Toxicological Evaluation of Pollutants 2-9
1 Cancer Effects 2-9
2 Non-Cancer Effects 2-11
IV Evaluating Exposure to Toxic Pollutants 2-15
VI Interpreting Risk Assessment Results 2-16
-------
CHAPTER TWO
GENERAL METHODOLOGY
The purpose of Stage I of the Santa Clara Valley IEMP is to compare health
risks attributable to different sources (aut-^nobiles, underground tanks, etc.)/
pollutants (organics, metals, etc.), and exposure pathways (air and drinking
water) in order to set research and management priorities. To make those
comparisons, we must first estimate the risks quantitatively. This chapter
discusses our methods of estimating and comparing risks. These methods are
generally consistent with the approach to quantitative risk assessment for
toxic air and drinking water contaminants that EPA has used in other recent
studies.
The limitations and uncertainties in our risk assessment deserve considera-
tion. First, the risk estimates are based primarily on existing knowledge
about pollutant releases, ambient conditions, and chemical toxicity; these data
vary widely in quality and are almost always incomplete. Second, the exposure
estimates incorporate a series of simplifying assumptions: although these
assumptions are necessary, they remain open to question and may be controversial.
Third, the potency estimates are necessarily based on current knowledge of the
toxicological effects of various substances. Considerable controversy exists
about the degree of hazard posed by different pollutants, and about whether some
are hazardous at all. Finally, resource and time constraints and the breadth
of our focus prevent us from analyzing individual issues in as much depth as
might be possible. We have attempted to strike a balance between the desire
for exhaustive and definitive analysis and the need for timely results at a
reasonable cost.
To assure that our data, assumptions and analytic methods are as good as
they can be, we have sought and will continue to solicit review from a range of
knowledgeable sources: EPA regional and headquarters staff, state and local
government officials, industry representatives, university faculty, and members
of public interest groups. Our risk assessment procedures must be seen in light
of their basic objective: to allow us to compare one risk to another, rather
than to make definitive statements about the absolute risk posed by a particular
substance, pollution source, or exposure pathway.
What Risk Assessment Is Not
It is important to realize that our analysis does not directly examine
disease incidence in the local population and attempt to link it with
environmental exposure. For many of the exposures and health effects with
which we are concerned, such epidemiologic study is difficult because background
incidence is high, exposures to particular individuals are difficult to quantify,
and a variety of factorsgenetics, occupational exposures, food, exposures
from other places, and many othersmight be the cause of an observed effect.
Also, many effects may take years to manifest themselves, and would not be evident
in the current population.
Because it is not an epidemiologic study, the IEMP risk assessment is not
intended to and does not answer questions such as what caused a statistically
higher rate of birth defects in the Los Paseos area, near the site of a major
2-1
-------
2-2
solvent leak at Fairchild Camera and Instrument. Instead, it attempts to evaluate
what health effects might result from current and future environmental exposures.
While different in approach and interpretation, risk assessments and epidemiologic
work are complementary. Risk assessment can help to identify populations and
geographic areas that appear to be at risk and that therefore might be appropriate
subjects for an in-depth epidemiologic study. Epidemiologic analysis increases
scientific understanding of the relationship between exposure and health effects,
thereby strengthening the basis of risk assessment. Epidemiologic analysis may
also, in some cases, be useful for confirming specific hypotheses suggested by
risk assessment.
The uncertainty in the lEMP's risk assessment assumptions is great enough
that results should be considered rough indicators of the probable magnitude of
effects, not as precise, site-specific predictions of effects. Risk assessment
estimates probabilities of various health effects, and thus is most useful for
analyzing generic problems and for generalizing about large populations. For
example, it is far more useful for examining the overall health risks from
groundwater contamination than for speculating about the likely health effects
of a specific plume of groundwater contamination on a small number of affected
persons. The greater usefulness of risk assessment for drawing broad rather
than narrow conclusions springs from several factors: (1) the toxicological
potency estimates take into account the variation in susceptibility in a
population, and thus are more reasonably applied to a large rather than a small
group; and (2) statistically, even if the estimates of likely effects are
correct, the percentage range of error in predicted effects is less when
projections are applied to larger populations.
OVERVIEW OF METHODOLOGY
The following overview of the methodology for quantifying health risks
applies, in theory, to both cancer and non-cancer health effects. However, the
Stage I Report contains quantitative risk estimates for cancer effects only. A
more detailed discussion of our analysis for both types of health effects
begins below.
In the IEMP, we define risk to an individual as the increased probability
that an individual exposed to one or more chemicals will experience a particular
adverse health effect during the course of his or her lifetime. (An average
lifetime is assumed to be 70 years.) Risk to the population is the expected
increased incidence (number of cases), above the background rate, of an adverse
health effect in an exposed population. For cancer risks, we present both
types of quantitative estimates in the Santa Clara Valley IEMP.
The IEMP risk assessment combines estimates of toxicological potency,
derived from laboratory and occupational studies, with estimates or measurements
of local contamination (or exposure) levels to estimate risk to the local
population. This approach involves a relatively high degree of uncertainty,
and, in most cases, cannot easily be verified or disproven by observation. It
is useful, however, because it is analytically straightforward and is the only
method available for estimating the future health effects of current and future
exposures.
-------
2-3
The two key elements in estimating risk are a chemical's estimated
toxicological potency and an individual's exposure to that chemical.
The toxicological potency assessment involves (1) a qualitative 'hazard
identification' to determine if evidence exists that a chemical causes an
adverse health effect; and if so, (2) a quantitative estimate of the 'dose-response
relationship1 or potency of a chemical, linking specific quantities of a chemical
to particular effect levels. CXiantitative potency estimates typically relate
given dose (equivalent to exposure) levels to the percent incidence expected in
a population, or to a probability that an individual will be affected during
the course of his or her lifetime. (See Figure 2-1 for an example of a dose-
response curve.)
Exposure is estimated by measuring or estimating the ambient concentration
level at which a chemical is present in the environment and making assumptions
about the relationship between ambient conditions and actual exposures. These
standard assumptions, called exposure constants, account for the amount of air
or water a typical person takes in in a day, and the weight of an average
person.1 Vfe assume that an average person weighs 70 kilograms (or roughly 154
pounds), breathes 20 cubic meters of air each day (20 m^/day), and drinks 2
liters of water each day.
Finally, exposure and potency estimates are combined to estimate individual
and population risks. This methodology is illustrated in Figure 2-2.
If we make the simplifying assumption that the dose-response curve is
linear, potency (risk per unit of exposure) is constant at any level of exposure
in which we are interested (unless the chemical has a threshold, below which we
believe there is no effect; thresholds will be discussed below in the section
on non-cancer effects). Under these assumptions, lifetime risk to the exposed
individual is simply the product of exposure and potency:
R = E x P
(individual (exposure) (potency)
risk)
As discussed above, exposure is the product of ambient levels of the
pollutant in the medium of concern (air or drinking water) and exposure constants
(i.e., the standard assumptions about body weight, intake and absorption):
E = Y x Z
(exposure) (ambient concentrations (exposure constants)
in medium of concern)
1 A further calculation of the actual dose likely to be absorbed into the body
involves an estimate of the body's propensity to absorb a chemical to which it
is exposed. This conversion of exposure to dose is estimated for each chemical
and exposure route, and is contained within the potency estimate.
-------
FIGURE 2-1 DOSE-RESPONSE CURVE
INCIDENCE:
% OF EXPOSED
POPULATION
No Effect
Threshold
fief*
ro
i
->o-v
DOSAGE: MILLIGRAMS PER DAY
-------
2-5
FIGURE 2-2 IEMP RISK ASSESSMENT METHODOLOGY
EXPOSURE
ASSESSMENT
Modeling of
Fate and
Transport
Monitoring
Ambient
Pollution
Concentrations
"^S^% SOURC
PATHWAY
TO EXPOSURE
'LAB
EXPERIMENTS
nnnnnnnn
IE
UUUUUUUU
EPIDEMIOLOGICAL
STUDIES
POTENCY
ASSESSMENT
Hazard
Identification
Quantitative
Potency
Estimates
ESTIMATED HEALTH RISK
-------
2-6
Therefore, individual lifetime risk can be calculated by multiplying
ambient concentrations (Y) and exposure constants (Z) by potency (P):
(individual (ambient concentration (exposure (potency)
risk) in medium of concern) constants)
For example, we might wish to calculate the lifetime risk of cancer for
an individual exposed to ten micrograms per cubic meter of benzene in the air.
Filling in equation (3) presented above, the ambient concentration (Y) is ten
micrograms per cubic meter (which is the same as 0.01 milligrams per cubic meter
or 0.01 mg/m3). Our exposure constants (Z), as discussed above, assume a
typical 70 kilogram person breathing 20 cubic meters of air each day (20m3/day).
Our potency value (P) is the cancer potency score for benzene inhalation developed
by EPA's Carcinogen Assessment Group (CAG) of 0.029 (mg/kg/day)~l. This value
indicates that an individual taking in one milligram of benzene per kilogram of
body weight per day (this would be 70 milligrams per day for a typical 70
kilogram person) for a lifetime has an estimated chance of about three in a
hundred of contracting cancer. (If the units are confusing, think of the score
as indicating cancer risk for a given dose of benzene.)
Using this information to fill in the equation presented above, we can
estimate risk as follows [THIS IS ONLY AN EXAMPLE]:
0.01
m-
20 m3
day 70 kg
0.029
(mg/kg day)
(individual (ambient (exposure constants) (potency)
risk) concentration)
Working through the arithmetic, individual risk (R) in this case is 8 x 10~5,
or eight chances in 100,000 of contracting cancer over a lifetime.
To calculate the aggregate expected increased incidence of disease, we
multiply the average individual risk by the number of individuals exposed:
I R x P
(aggregate expected (individual risk) (exposed population)
increased incidence
of disease)
Following this example, suppose that we estimate that 100,000 people are
exposed to 10 micrograms per cubic meter (or 0.01 mg/m3) of benzene. We would
estimate incidence as follows [THIS IS ONLY AN EXAMPLE]:
I = 8 x 10~5 x 100,000 people
(aggregate incidence (individual risk) (exposed population)
of cancer)
-------
2-7
The product of this equation is 8 cases over a 70 year period (the assumed average
lifetime in the individual risk calculations). By convention, incidence (unlike
individual risk) is reported in annual cases, so we divide our product, 8 cases,
by 70 to arrive at our estimate of 0.1 cases per year, or about one case every
ten years.
We estimate individual and population risks from ingestion of contaminated
drinking water in much the same way. Vfe first estimate a chemical's potency
(potency is not necessarily the same in drinking water as it is in air).
Concentrations of suspected toxic chemicals in the drinking water are measured
or estimated, then converted to doses by use of exposure constants (as stated
above, we assume that a typical individual drinks 2 liters of water a day).
Individual risk and aggregate incidence in a population are then calculated as
above.
We examine the potential risk of different health effects separately, even
if sources, pollutants and populations affected are the same. For example,
benzene exposure from automobile emissions may increase the risk of cancer as
well as the risk of "blood effects." In such a case, we examine the risks of
cancer separately from the risks of the non-cancer blood effects. In other
words, the risks from different health effects (cancer, kidney, liver, blood,
etc.) are not lumped together in the analysis.
By contrast, for a single, given health effect (e.g., cancer, kidney,
liver, etc.), we analyze risks separately and, when possible, we also aggregate
risk estimates across populations, sources, pathways, or pollutants. For
example, in our groundwater analysis we estimate the total (additive) risk of
cancer (the single, given health effect in this case) from exposures to percnloro-
ethylene; we also estimate the total (additive) risk of cancer from leaking
tanks: and we estimate the total (additive) risk of cancer from all groundwater
carcinogens. We are thus able to examine the risk of a given health effect
from different perspectives: as a function of individual contaminants (e.g.,
perchloroethylene), individual sources (e.g., leaking tanks), and pathways
(e.g., groundwater). This ability to use a given health effect, such as estimated
cancer risk, as a 'common currency1 for evaluating a variety of different
problems is one of the more useful aspects of risk assessment, allowing us to
compare problems and set priorities.
The remainder of this chapter is in three parts. The first part briefly
discusses our toxicology/potency work. The second section outlines some generic
aspects of our exposure analyses (we discuss the exposure assessments in detail
in chapter 3 on air, and in chapter 4 on drinking water). The chapter concludes
with a discussion of how IEMP risk assessment results should be interpreted.
TQXICOLOGICAL POTENCY ASSESSMENT
The lEMP's toxicological work has involved two major steps: (1) a pollutant
screening exercise to identify pollutants of significant toxicity likely to be
present in the Santa Clara Valley; and (2) a more detailed qualitative and
quantitative evaluation of pollutants identified in step (1) as being of potential
concern. As discussed in greater detail below, this second step is different
for cancer and non-cancer effects. Our toxicological evaluation yields the
following results:
-------
2-8
0 qualitative information on adverse health effects associated with
exposure;
0 estimated potency values without thresholds for carcinogenic effects;
0 estimated thresholds for most non-cancer effects; and
0 qualitative information on mutagenic effects.
Pollutant Selection
The IEMP performed an initial screening exercise to identify, from a
master list of about 1800 pollutants, those pollutants most likely to pose an
environmental health risk. Using a combination of exposure and toxicity criteria,
we initially identified about 50 chemicals that might pose such risks. For
this report, we estimated exposures and health risks for about 30 pollutants
all those for which we could find evidence of exposure and for which we had
toxicological data.
Toxicity criteria included the following:
o designation by the International Agency for Research on Cancer
(IARC) as a known or probable human carcinogen; or
o listing or active consideration of listing by EPA as a hazardous air
pollutant under Section 112 of the Federal Clean Air Act; or
o regulation by EPA under the Safe Drinking Water Act with a Maximum
Contaminant Level, or proposal by EPA of a Recommended Maximum
Contaminant Level under that Act; or
o scoring by EPA's Carcinogen Assessment Group (CAG) as a possible,
probable or known human carcinogen; or
o issuance of a Hazard Alert by the Hazard Evaluation System and
Information Service of the California Department of Health Services.
Exposure criteria included the following:
o detection in soil or groundwater in the Regional Water Quality Control
Board's leak detection program; or
o detection in drinking water by water retailers, either through routine
monitoring reported to the State Department of Health Services, or
through special monitoring undertaken by the Santa Clara Valley Water
District in 1982 (subsequently supplemented by monitoring in 1984/85
required by AB 1803); or
o listing by the Bay Area Air Quality Management District in its speciated
emissions inventory for point sources in Santa Clara County; or
o application as a pesticide in Santa Clara Valley in 1982 (as indicated
by pesticide records); or
o presence as an ubiquitous constituent of gasoline.
Details on the pollutant selection process are available upon request.
-------
2-9
Toxicological Evaluation of Pollutants
Cancer Effects
Assessnent of a pollutant's carcinogenic potential involves a qualitative
hazard identification followed by a quantitative estimate of potency (if the
qualitative assessment indicates that the substance is a suspected carcinogen).
The lEMP's carcinogen assessment methodology is consistent with EPA's Proposed
Guidelines for Carcinogen Risk Assessment as published in the Federal Register,
November 23, 1984.
The qualitative evaluation involves a review of the relevant literature to
determine whether sufficient evidence exists to establish a substance as a
suspected carcinogen. This review includes both human epidemiology studies and
animal experiments.
The qualitative evaluation results in an estimate of the strength of evidence
that a substance is carcinogenic. To summarize the strength of evidence, we
present the classification schemes developed by the International Agency for
Research on Cancer (IARC) and/or the scheme developed by EPA (whicn is an
adaption of the IARC system). In the IARC scheme, the evidence that an agent
produces cancer in humans is divided into three categories: known human carcinogen,
probable human carcinogen, or insufficient evidence to classify. The EPA
scheme stratifies the weight of evidence in a similar manner with five groupings:
carcinogenic to humans; probably carcinogenic to humans; possibly carcinogenic
to humans; not classifiable as to human carcinogenicity (not enough information);
and no evidence of carcinogenicity for humans (information suggests substances
are not carcinogenic). (For further discussion of the methodology for qualitative
assessment of carcinogenicity, see the EPA's Proposed Guidelines for Carcinogen
Risk Assessment in the Federal Register, November 23, 1984.)
If sufficient qualitative evidence exists to establish a substance as a
suspected carcinogen, the next step is a quantitative analysis to estimate the
degree of risk associated with exposure to that chemical. This second step in
toxicity evaluation results in a specific potency score, which is a quantitative
estimate of the dose-response curve for a given pollutant.
In estimating the potency of a known or suspected carcinogen, the IEMP uses
chemical-specific, peer-reviewed potency scores developed by EPA's Carcinogen
Assessment Group (CAG). toe also examine TCA, which some people consider a
potential carcinogen, but for which there no current CAG score exists. For
this substance, we perform sensitivity analyses, analyzing the impact under
alternative assumptions of carcinogenicity and non-carcinogenicity. These
analyses are discussed in greater detail in subsequent chapters.
Potency Evaluations: Some Simplifying Assumptions
In evaluating the potency of a substance, EPA's CAG relies several simplifying
assumptions that are used widely in this type of analysis. These assumptions
allow risk assessors, albeit with some uncertainty, to extrapolate from observed
laboratory or epidemiologic data to situations involving different exposure
levels and different populations. Important types of assumptions include: (1)
interpreting lab or animal data; (2) extrapolating data to other species or
populations; (3) extrapolating high-dose effects data to low-dose exposure
-------
2-10
situations; and (4) estimating the effects of exposure t;o combinations
of toxic chemicals.
In estimating the potency of a suspected carcinogen, EPA's CAG assumes
that cancer is a non-threshold health effect. This assumption means that any
level of exposure to a carcinogen is deemed to involve some risk. Because it
is impossible to measure directly the human risks associated with low levels of
exposure to carcinogens, scientists must rely on mathematical models to extrapolate
from effects at high animal or human occupational exposures to risks at low
environmental exposures. The extrapolation techniques are different for data
on humans and data based on animal studies. In estimating potency from animal
studies, CAG uses the most sensitive species observed and calculates the upper
95% confidence limit on the estimated slope of the low-dose range of the dose-response
curve; both assumptions tend to increase the estimated potency of the substance.
The mechanisms of carcinogenesis are not understood well enough'to develop
precise predictive models. Although several models may fit the observed data,
different plausible models may lead to large differences in projected risk at
low doses.
An important aspect of the procedure for extrapolating from high dose to
low dose effects is the assumption that the dose-response relationship is linear
in the low dose range covering most of the actual environmental exposures.
Thus, a unit increase in exposure produces the same increase in cancer probability
(individual risk) or incidence (aggregate risk) regardless of the background
exposure level. The linearity assumption simplifies both the estimation of
a dose-response curve and the use of the estimate (as in risk assessment).
Since scientists rarely have data on effects at this exposure level, this
simple assumption is probably as good as any other.
We emphasize that the linearized, multistage model used by CAG leads to a
plausible upper limit to the potency estimate that is consistent with widely
accepted mechanisms of carcinogenesis. Therfore, such an estimate does not
necessarily give a realistic prediction of the risk. The EPA's Science Advisory
Board (SAB) (in its preliminary review of EPA's proposed guidelines) noted that
the true value of the risk is uncertain, and for many substances the lower
bound estimate is zero. The SAB also noted that an established procedure that
is applicable to a variety of substances does not yet exist for making most
likely or best estimates of risk within the range of uncertainty defined by
upper and lower limit estimates. However, the SAB suggested that the Agency
strive to provide most likely or best estimates on a case-by-case basis where
the data and procedures are available.
The traditional approach to carcinogen risk assessment at EPA has been to
take the most conservative approach in developing potency estimates. In so
doing, EPA develops estimates that are unlikely to underestimate the true
potency of a chemical. We f'ol that such an approach is appropriate, particularly
in a screening analysis, tc ot underestimate potential human health impacts.
Therefore, we follow the traditional approach and use the plausible upper bound
potency estimates that CAG develops based on the linearized multistage model.
Another important simplification is the assumption that we can add the
cancer risks from various exposures to arrive at a total cancer risk to the
exposed individual. In some cases, the risks of cancer that we are adding are
for different types of cancer. For example, an individual may be at risk of
-------
2-11
lung cancer from air exposure to one substance, and also be at risk of liver
cancer from exposure to another substance. We add these two risks of cancer to
arrive an estimated risk of cancer for that individual. (Of course, it is also
possible to disaggregate risk estimates, so that the risks of lung cancer and
liver cancer, for example, are presented separately.)
The assumption that cancer risks are additive is potentially important
because individuals are typically exposed to many toxic chemicals. Additivity
implies that these pollutants are neither synergistic nor antagonistic in their
combined effect (i.e., that the presence of one chemical neither enhances nor
inhibits the effect of another). Synergism and antagonism probably do occur,
at least some of the time. However, most data analysis and experimentation on
toxicological potency attempt to isolate the effect of a single chemical.
Analyzing the effects of a chemical in combination with other chemicals is a
far more complex task, and such analysis has not been done for most pollutants.
In the absence of information on synergism and antagonism, the simple additivity
assumption seems most appropriate.
Finally, it is important to remember that various sub-groups within any
population exposed to a toxic chemical vary significantly in their sensitivity
to the substance. For example, pregnant women, young children, persons with
genetic impairments, and other often unidentified groups may be especially
susceptible to a particular chemical exposure. The potency values we use to
estimate risk reflect this variation; while some people may face higher risks
(and others lower risks), the incidence estimate is intended to reflect the
risk in a population with a typical distribution of sensitive persons.
A more detailed discussion of the important aspects of the potency evaluation
methodology, which we have summarized here, is contained in the EPA's Proposed
Guidelines for Carcinogen Risk Assessment in the Federal Register, November 23, 1984.
Non-Cancer Effects
The non-cancer health effects which we examine in this analysis include
systemic effects such as liver and renal toxicity, and respiratory, neurobehaviora1,
cardiovascular, blood and other effects; teratogenicity and reproductive effects;
and mutagenicity. The toxicological evaluation is somewhat different for
non-cancer effects than for cancer effects.
The crucial distinction between quantifying cancer and non-cancer effects
is the assumption that while cancer is a non-threshold effect, most other
environmentally caused health effects involve thresholds. This threshold
assumption means that non-cancer health effects are thought not to be caused IT/
exposure below some safe level (the no-effect threshold). Therefore, the first
step in the quantitative evaluation of a non-cancer effect is calculating a
no-effect threshold.
No-Effect Thresholds
In determining the threshold dose below which no observable adverse effect:?
are assumed to occur, we rely on EPA Reference Doses (RfDs; also referred to as
Acceptable Daily Intakes or ADIs) and on thresholds computed by IEMD toxicologists
and consultants. The IEMD thresholds were derived using the same procedures used
in estimating RfDs; however, these thresholds have not been subjected to full EPA
-------
2-12
review.
To calculate an RfD, EPA scientists collect the available animal and human
data and note the various dose levels (in milligrams per day) at which different
health effects are seen. The scientists then array the dose vs. effect infor-
mation, as illustrated in Figure 2-3, and identify the NOEL (the No Observed
Effect Level). The NOEL represents the highest dose tested that did not produce
observable results. The scientists also try to define the LOEL (the Lowest
Observed Effect Level), which is the lowest dose tested at which some type of
effect was seen. The LOEL is generally a mild effect or an effect which involves
a very small portion of the exposed population. They also try to define the
PEL (the Frank Effect Level), which is the dose that involves more serious
health problems.
Bear in mind that there will be many NOELs and LOELS for each chemical.
These levels depend on the doses selected and the health effects tested for by
the researchers. Research is very expensive and each experiment cannot be
exhaustive. Since EPA is dependent on available research, there is some inherent
uncertainty in the definition of NOELs and LOELs.
To calculate the RfD, EPA scientists select the lowest reliable NOEL and
divide it by safety factors. The selection of appropriate safety factors is
based on the nature of the study from which the NOEL was derived. Most safety
factors are multiples of ten with each one representing an extra degree of
uncertainty to account for extrapolation from animal data to the average human,
from the average human to the most sensitive subgroup, from subacute effects to
chronic effects, and from the LOEL to the NOEL. In general, the safety factor
is larger than the expected differences in the exposure levels producing these
effects, and is therefore a conservative estimate of the actual threshold.
Figure 2-3 illustrates that different effects may be associated with
different thresholds for the same substance. The RfD is calculated so as to
protect against all adverse effects and, as such, tends to be below all the
other health-specific thresholds. In some cases, however, more recent or
different data collected by IEMD toxicologists and consultants suggest a health
effect threshold that is below an older RfD. (Such a threshold is derived in
the same manner as an RfD, i.e., by taking the NOEL for that effect and dividing
by safety factors.)
The IEMP analysis uses the RfD-derived thresholds as well as thresholds
developed by IEMD toxicologists and consultants (using the same procedure as is
used to estimate RfD thresholds) to examine possible non-cancer health effects.
This examination involves comparing monitored or modeled concentration levels
with estimated human thresholds. Because a single substance may have many
different thresholds for the different health effects associated with it, we
examine each health effect for each substance separately.
If exposures to a substance are below the threshold for a given effect,
then we expect no increased risk of that health effect from that substance. If
exposures exceed the estimated threshold for a particular health effect, we
then indicate the possibility of increased risk of that effect and try to
estimate the population exposed to such concentrations.
-------
Incidence
(% of
population
affected)
2-13
RfD
NOEL LOEL
Dose (milligrams per day)
Other
Other PEL
Figure 2-3: Examples of Thresholds for a Single Substance
RfD (Reference Dose): EPA's no-effect threshold = NOEL divided by safety factors.
NOEL (No Observed Effect Level): highest dose at which no effect is seen.
LOEL (Lowest Observed Effect Level): lowest dose at which effect is seen.
Other other adverse health effect seen at this dose (e.g., liver or reproductive).
PEL (Frank. Effert Level)? more serious health effect seen at this dose.
-------
2-14
T/fe have restricted our evaluation of mutagens to a qualitative assessment.
Mutagenic effects cannot be expressed meaningfully in quantitative terms. Therefore,
we note positive evidence of mutagenicity by listing it as an effect on the
relevant tables.
Using thresholds in the way discussed above allows us to identify potential
risks of non-cancer effects without estimating the probability of disease or the
possible number of cases. As such, the thresholds alone provide a means of
screening risks.
A list of substances analyzed, their possible non-cancer health effects,
estimated quantitative thresholds (when available), information sources (i.e, is
the threshold derived fron an RfD or by IEMD), and the scientific evidence of the
health effect are presented in chapters 3 and 4.
Making Quantitative Estimates of Risks of Non-Cancer Effects
Originally, we had hoped to provide quantitative estimates of non-cancer
health effects in the Stage I analysis, as we do for cancer effects, to assess
the magnitude of risks associated with exposures above thresholds. However, for
a number of reasons we decided not to estimate these non-cancer risks quantitatively
in the Stage I Report.
The reasons for not making these quantitative estimates revolve around the
fact that the calculation and use of dose-response curves (the quantitative
estimate of potency) for non-cancer effects represent an innovative approach in
risk assessment. Since EPA has never developed a method for estimating the
probability of disease for non-cancer effects, the approach developed by IEMD
currently does not enjoy broad consensus within the Agency. In contrast, use of
CAG potency estimates and threshold values to estimate risks or levels of concern
is relatively straightforward; such indicators of risk have been used throughout
the Agency for many years. Because of the innovative nature of the IEMD non-cancer
health effects approach, various scientists are reviewing the approach IEMD
developed, including EPA's Science Advisory Board.
It is important to note that the decision not to estimate the potential
number of cases for non-cancer health effects does not reduce the capacity of
the Stage I analysis to set research and management priorities for Stage II.
By using the quantitative estimates of thresholds and comparing them to estimated
ambient concentrations of pollutants, we are able to identify substances (and
pathways and sources) that pose potential risks of non-cancer health effects,
and include them in the Stage II analysis. Therefore, quantitative estimates of
the magnitude of the potential risks are not necessary to perform the Stage I
screening analysis for non-cancer health effects.
Although we do not use the estimates in the Stage I Report due to the lack
of scientific consensus, a summary of important dose-response estimates for
non-cancer effects developed by IEMD is included in Appendix B. The Appendix
also contains a discussion of the scientific basis for not using the dose-response
information for projecting non-cancer health effects in the Stage I analysis.
-------
EVALUATING EXPOSURE TO TOXIC POLLUTANTS
Once we have identified hazards and calculated thresholds (for non-cancer
effects) and potency scores (for cancer effects), we combine this toxicological
information with exposure estimates to examine risk. In this section, we discuss
our basic exposure methodology and its application in assessing risk. We discuss
the exposure methodology in detail in Chapter 3 on air and in Chapter 4 on
drinking water.
Most of our exposure assessment efforts are directed at identifying ambient
concentrations of substances in the media of concern: air and drinking water.
We then make certain assumptions about how ambient concentrations relate to
actual human exposure. These assumptionsthe exposure constants regarding how
much air a person breathes or water he or she drinksare described above,
We estimate anbient concentrations in two ways; direct morn toring or
simulation modeling. There are a number of advantages to modeling emissions in
order to estimate ambient concentrations: modeling may provide tne only way to
estimate ambient conditions in situations where sucn conditions cannot be
observed (such as the future); it can take into account the geographic variability
of a large area, whereas monitoring data are often from a few specific points
and must be extrapolated over the area of concern; it is often less costly than
extensive monitoring; it allows us to estijnate exposures to the "maximum exposed
individuals" as well as to average individuals; and it links concentration
estimates, and hence exposures, to sources. Such source information is important
as a risk management tool because it allows us to estimate the impact of pollution
control options.
On the other hand, constructing a model of pollutant releases and resultant
ambient concentrations involves making assumptions about the important processes
between pollutant source and human recipient. Building such a model thus requires
an understanding of those processes, which is not necessary if one can simply
monitor ambient concentrations directly. Thus, the primary advantage of reliable,
long-term monitoring data is that they provide a more direct and often simpler
means of estimating ambient conditions. Even with direct monitoring, however,
interpretation of results can be difficult if data are limited, if lab contamina-
tion is suspected, or if any one of a number of other problems arises.
Ideally, we would have both monitoring and modeling data for each of the
substances studied. Used together, the monitoring data can serve as a check on
the modeling results. For this analysis, however, we relied on available data;
thus, we use different types of data in different situations.
We have relied on monitoring data to estimate current exposure and risks
from drinking water and from metals in the ambient air. For toxic organic
chemicals in the-air, we had little monitoring data and felt that the limited
data we did have were somewhat unreliable. As a result, we used the limited
monitoring data mainly as a rough check on our dispersion modeling of organic
air contaminants.
In some cases, we felt that current monitored contamination levels were
not good indicators of likely future levelssuch as when new regulations are
expected to decrease exposures or when contaminants move slowly from pollutant
source to human recipient. In these cases, we adjusted current contamination
-------
2-16
levels to account for expected future changes or made estimates based on our
understanding of the sources and processes involved (i.e., modeled the situation).
Vfe used both of these approaches to estimate future exposure and risk from
drinking water contamination, as described more fully in Chapter 4.
The key steps in modeling pollution are (1) estimating the number of sources
and the frequency and magnitude of releases; and (2) characterizing the processes
by which a pollutant is transported in air or water, including the speed of
transport, the extent of dilution or dispersion, and any chemical transformation
the pollutant might undergo (such as degradation to a non-toxic form). With
such assumptions, one can estimate resulting ambient concentrations at various
distances from the source. Estimating long-term average concentration levels,
which are of primary concern for evaluating chronic health impacts, is simpler
than short-term modeling, which must take greater account of variations in
meteorological conditions.
To determine individual lifetime risks and population incidence, we must
relate estimated ambient concentrations to exposed individuals. For most sources
analyzed in this report, we attempt to estimate the highest level of exposure
likely to be experienced by anyone (risk to the "maximum exposed individual" or
MEI), as well as the average individual lifetiine risk and population incidence.
To estimate the risk to a highly exposed individual, we typically need to know
how far an individual is from a source. To estimate population incidence, we
must identify the number of people exposed to a given pollutant concentration.
In this analysis, we used breakdowns of residential population by 5 kilometer
square grid, prepared by the Association of Bay Area Governments (ABAC) to
estimate air impacts. Populations exposed to drinking water contaminants were
estimated from flow and population data provided by water purveyors.
Our exposure methodology implicitly assumes that people are exposed to
estimated outdoor ambient air concentrations 24 hours a day. In fact, people
spend most of their time indoors, either at home or at work, while this simplification
overstates actual exposure to outdoor air, the bias introduced by this procedure
may not be too great, since recent EPA studies show that many outdoor air
contaminants are also found indoors at equal or greater concentrations. (In
addition, there are indoor sources of air contaminants to which people are
exposed which we did not include in this analysis.) Assuming exposure based on
residential population demography oversimplifies actual exposure patterns,
since many people breathe air and drink water from a number of different locations
in a day.
INTERPRETING RISK ASSESSMENT RESULTS
The estimates of individual health risk and aggregate incidence from
exposure to toxics presented in this Report should not be interpreted as precise
or absolute estimates of future health effects. The simplifying assumptions
and uncertainties in both the toxicology and exposure components are simply too
great to justify a high level of confidence in the predictive value of the IEMP
results.
The potency and threshold estimates used in this study are consistently
conservative in the direction of protecting health; they may overstate the
likely effects of chemical exposure but are unlikely to underestimate them. By
-------
2-17
contrast, our exposure estimates are not as clearly conservative; seme assump-
tions are conservative while others are our best guess of an actual value
(e.g., the speed of groundwater flow). Overall, we have tried to be somewhat
conservative in our exposure assessments, as is appropriate in a screening
exercise. However, we also made extensive use of sensitivity analysis to
examine alternative reasonable values for key but uncertain factors. We have
also made limited use of sensitivity analysis in our potency assessments.
Sensitivity analyses are useful in determining the importance of various
assumptions used in the base case analysis, and to determine the validity of
the conclusions we draw. A sensitivity case is similar to a "what if" case.
For example, we may assume in a base case that a substance degrades rapidly in
groundwater, but a sensitivity case can help us answer the question of "what
if this substance does not degrade?" If there is a significant difference
between the two results, this helps us to identify areas of particular uncertainty
and importance for further research. If the results are similar, we can be
more confident of the conclusions we draw because we know that some of the
different assumptions do not affect the results significantly.
We typically present risk estimates as ranges rather than point estimates,
reflecting the uncertainty of the underlying assumptions. It is important to
read the detailed methodology sections carefully to understand what confidence
to place in a particular estimate. The ranges presented often reflect alternative
assumptions about key variables affecting health risk, but do not necessarily
encompass the total range of possible effects. For example, we often estimate risk
using only a single, conservative potency estimate. If all other assumptions
are correct, it is possible that the health impacts will be overstated.
On the other hand, we may understate risks to the extent that we do not
estimate exposure and risks from sources and pollutants that in fact have toxic
effects. (We have tried to identify the sources and pollutants of greatest
concern, but we cannot be certain that we have not overlooked some sources and
pollutants.) Overall, our risk estimates probably overstate health impacts
more often than they understate them.
Because of the uncertainties involved in the IEMP methodology, the results
should not be interpreted too literally. For example, one should not conclude
that a source projected to cause three cases of cancer per year is clearly
worse than a source projected to cause two cases; given the overall uncertainty
of the analysis, these two results are virtually indistinguishable. On the
other hand, it is reasonable to conclude that a source projected to result in
ten cases a year is probably worse than one projected to cause only one case.
Despite the uncertainties, our risk estimates are useful for roughly assessing
the magnitude of the overall health threat from environmental toxics in Santa
Clara Valley; assessing the comparative risks from particular pollutants, sources
and pathways; comparing problems with one another; and setting priorities on
toxics problems.
-------
2-18
REFERENCES
1. Environmental Protection Agency: Proposed Guidelines for Risk Assessment;
Request for Comments, Federal Register, November 23, 1984.
.2. Environmental Protection Agency: Risk Assessment and Risk Management;
Framework for Decision-Making, December, 1984.
3. Environmental Protection Agency, Regulatory Integration Division: Health
Score Evaluation Documents for all substances in the IEMP analysis.
-------
CHAPTER THREE
ANALYSIS OF RISKS FROM EXPOSURE TO AIR TOXICS
-------
CHAPTER THREE
ANALYSIS OF RISKS FROM EXPOSURE TO AIR TOXICS
I Introduction 3-1
II Objectives 3-3
III Pollutants and Sources Considered. 3-3
IV Overview of the Methodology 3-5
V Methodology and Findings: Metals 3-23
A Monitored Levels of Metals in Santa Clara Valley's
Outdoor Air 3-16
B Possible Sources of Metals in Santa Clara Valley 3-20
C Estimated Exposures to Metals 3-23
D Estimated Health Risks from Exposures to Metals 3-31
1 Results by Pollutant 3-33
2 Results by Source Category 3-36
VI Methodology and Findings: Organic Gases 3-41
A Emissions of Organic Gases Fran Different
Source Categories 3-41
1 Emissions Estimates from the AOMD Inventory 3-43
2 Emissions Estimates from Three Additional Sources 3-49
B Modeled Exposures to Organic Gases 3-56
C Estimated Health Risks from Exposure to Organic Gases 3-66
1 Results by Pollutant 3-67
2 Results by Source Category 3-79
D Results of the Short-Term Monitoring of Organic Gases 3-84
VII Methodology and Findings: Organic Particulates 3-92
A Sources of Organic Particulates 3-92
B Methodology for Estimating Exposures 3-92
C Analysis of Health Risks 3-94
VIII Summary and Conclusions 3-96
IX Key Assumptions and Analytic Decisions in the Air Analysis 3-116
X References 3-121
-------
CHAPTER THREE
ANALYSIS OF RISKS FROM EXPOSURE TO AIR TOXICS
This chapter describes the analysis of human exposures to toxic pollutants
in the outdoor air that EPA and the Bay Area Air Quality Management District
(AOMD) conducted jointly as part of Stage I of the Santa Clara Valley Integrated
Environmental Management Project. The overall purpose of Stage I is to compare
human health risks attributable to various pollutants, sources, and exposure
pathways, and, by doing so, to shed light on appropriate priorities for improved
environmental decisions by federal, state, and local agencies, firms, and even
individuals. The analysis presented here is meant to allow broad comparisons
of the risks attributable to exposures to airborne toxics against the risks
posed by contaminants in drinking water or other exposure routes.
This analysis of risks from exposure to outdoor air toxics is based
primarily on existing data; this basis limits both the number of pollutants we
are able to consider and the confidence we can attach to the resulting estimates
of risk. The IEMP analysis has concentrated on a selected set of metals,
organic gases, and organic particulates, primarily because sufficient data to
assess exposures or toxicity, or both, were unavailable for other air pollutants.
For metals, historical monitoring data provided the basis for our exposure
estimates. For the organic gases, we relied primarily on dispersion modeling,
based on emissions inventories provided by the AOMD, in order to estimate
ambient concentrations. For organic particulates, we relied on surrogate
values for ambient concentrations, based largely on similar work conducted in
other cities. Each of these, estimation techniques, and the underlying data,
involve considerable uncertainty, and that uncertainty varies depending on the
pollutant. We have tried in every case to resolve the uncertainty in a
conservative manner so that for the pollutants examined, we are more likely to
overestimate than underestimate risks.
The risk analysis for air contaminants does not include risks due to
exposure to contaminants in indoor air, either in workplaces or in residences;
our analysis is confined to pollution in the outdoor ambient air.* Other research
indicates that indoor air may contribute more to total exposure for some compounds
than does outdoor air. (For example, see Office of Policy Analysis, "The Air
Toxics Problem in the United States: An Analysis of Cancer Risks for Selected
Pollutants," USEPA, May 1985.) Nevertheless, risks due to indoor air exposures
were not included in the analysis for a number of practical and jurisdictional
reasons.
One reason that we did not include risks due to indoor air pollutants is
that, to our knowledge, there are no local exposure or monitoring data on which
to base such ah analysis. Developing local data on indoor air is an expensive,
resource-intensive effort beyond the scope of the project. Studies on indoor
Throughout this chapter, we use "ambient air" to mean the outdoor air to
which people are actually exposed. The term is used, in air pollution
jargon, to distinguish between releases of pollutants to the air
("emissions") and the resultant impacts on the air ("ambient conditions").
-------
3-2
air from other cities vary widely in terms of their findings, making extrapolation
to a different, local situation difficult and beyond the scope of this project,
given its limited resources. Moreover, regulating indoor air is outside EPA's
traditional statutory mandates and thus potentially beyond our jurisdiction.
Similarly, the IEMP analysis does not include "criteria" pollutants
(including oxides of sulfur and nitrogen, ozone, carbon monoxide, lead, and
total suspended particulates).* EPA set national ambient standards for those
pollutants over a decade ago under the mandates of the Clean Air Act at levels
deemed adequate to protect human health. (For each pollutant, standards were
set based on "air quality criteria" that indicate the effects of the pollutant
on human health or welfare; hence the name "criteria" pollutant.) Hie air over
Santa Clara County meets those standards, except for some violations of the
ozone and carbon monoxide targets. Further, enormous amounts of scientific
expertise and governmental dollars have been invested in the past fifteen
years to gain an understanding of criteria air pollutants. The toxic air
pollutants we analyzed, in contrast, have been relatively little studied in
the past, and therefore, we judged that our ability to contribute to our
understanding of their effects was greater. However, in order to set the
risks from toxics in context and better inform risk management decisions on
air pollutants generally, the IEMP plans to examine the potential effects of
criteria pollutants in Santa Clara County as part of Stage II.
In addition, we did not include asbestos in our analysis. Outdoor air
exposures to asbestos have been raised as a concern within the County, specifi-
cally in Alviso, following the finding of asbestos wastes in the soil. Limited
soil and air sampling conducted by the California Department of Health Services
(DOHS) indicates that further work is necessary to define the extent of the
problem. EPA and DOHS have conducted some emergency remedial work (planting
vegetation and paving a dirt road and a school lot to keep dust levels down)
in the area of concern. Over the longer term, EPA and the Army Corps of
Engineers are overseeing work designed to further characterize sources; develop
concentration data in air, surface water, soil, and sediments; explore possible
remedial technologies; and develop environmental and ecological exposure and
risk estimates. Given the work underway, we decided not to use our limited
resources duplicating these efforts.
Finally, we did not attempt to estimate risks due to episodic releases of
toxic contaminants to air. Although EPA has, in recent months, concentrated
more of its attention on this issue at the national level, it is extremely
difficult analytically to estimate the probability of large-scale, episodic
releases or the magnitude of such releases; without such information, however,
we cannot estimate the exposures or risks which might result.
This chapter begins with a discussion of the objectives of the IEMP air
analysis. It then lists the air contaminants and sources of air pollution
which we considered. An overview of the methodology for estimating risk
follows. The main sections of the chapter describe, in detail, the analyses
by which we estimated risks from metals, organic gases, and organic particulates.
The last section of the chapter summarizes our conclusions about the nature
and approximate magnitude of the risks in Santa Clara Valley posed by those
outdoor air toxics we studied.
* Lead is discussed further in the section on metals.
-------
3-3
Objectives
The objectives for Stage I of the IEMP were shaped to some extent by the
limited data available on toxic air pollutants in Santa Clara Valley. One goal
was to improve our understanding of the human health risks attributable to
ambient air exposures from those substances for which data were available, and
to compare those risks to risks attributable to exposure to contaminants in
drinking water and other exposure routes. In addition, we wanted, to the
extent permitted by the data, to compare different air pollutants, and different
kinds of air pollution sources, in terms of their probable impact on human
health. At the same time, we wanted to put those estimates in perspective by
comparing air toxics risks in Santa Clara Valley to similar risks in other
urban areas and to other sources of risk to human health. A final goal was to
identify gaps in the existing data that could be filled during the lEMP's
second stage, or in subsequent work by EPA or other agencies.
Where possible, we have estimated the individual health risks to both
average and highly exposed individuals that are attributable to exposure to
various toxic air pollutants; we also attempted to estimate the aggregate
increased incidence of cancer attributable to those exposures.
Because of the breadth of the analytic objectives and the limitations in
available data, the Stage I work was designed as a "screening" exercise. That
is, we attempt to compare risks against one another in order to set research
and management priorities, rather than to arrive at unequivocal statements
about the absolute magnitude of air toxic risks in the Valley. Setting
priorities is facilitated by the use of a common basis for comparison, and
throughout the analysis we have presented two numeric measures of risk that
provide that comparison: expected increase in aggregate incidence of disease,
and lifetime individual risks. These quantitative statements are not, however,
an end in themselves, and because of the underlying uncertainties in this
study, they cannot be considered absolute predictors of risk or disease.
Their usefulness lies in the fact that they allow us to assess, very roughly,
the approxijnate magnitude of a number of environmental problems and compare
these problems to one another: they allow us to "screen" problems and set
priorities without projecting absolute estimates of risk.
Given this objective, we were more willing to use very rough data and
make assumptions than EPA would be if it were preparing a chemical-specific
risk assessment in support of a national regulation. Our goal, instead, was
to make the best use of existing data in order to set priorities among potential
problems in ambient air, in drinking water, and in other areas.
Pollutants and Sources Considered
The pollutants we considered in this analysis fall into three major
categories: metals, organic gases, and organic particulates. We estimated
risk attributable to exposures to eight metals, eleven organic gases, gasoline
vapors (which is a mix of compounds), and one class of organic particulates.
A handful of additional organic gases appear in the analysis, but were modeled
only to estimate risks to the most exposed individuals from particular sources.
The substances we considered are presented on table 3-1.
-------
3-4
Table 3-1
Pollutants Considered in the
Santa Clara Valley IEMP Air Analysis
Metals
Arsenic
Barium
Beryllium
Cadmium
Chromium
Lead
Nickel
Zinc
Organic Gases
Benzene
Carbon Tetrachloride
* CFC-113
(1,1,2-trichlorotrifluoroethane)
Chloroform
* Dichlorobenzene
* Dichloroethane
* Dichloroethylene
Ethylene Dibromide (EDB)
Ethylene Oxide (EtO)
Gasoline Vapors
* Glycol Ethers
2-ethoxyethanol (Cellosolve)
2-methoxyethanol (Methyl Cellosolve)
2-butoxyethanol (Butyl Cellosolve)
* Isopropyl Alcohol
Methylene Chloride (dichloronethane)
Perchloroethylene (tetrachloroethene)
* Phenol
Toluene
1,1,1-Trichloroethane (methyl chloroform)
Trichloroethylene (trichloroethene)
Xylene
Organic
Particulate
Benzo(a)pyrene
Group
These sobs Lances were modeled only to estimate risks to the most exposed
individuals from particular sources.
-------
3-5
The choice of contaminants for which we estimated health risks was based
primarily on the availability of existing data and the evidence of toxicity.
(See chapter 2 for a brief discussion of our pollutant screening method.) This
basis limits the number of pollutants we were able to consider. For instance,
we were unable to quantify exposures to all of the compounds used in semiconductor
and computer manufacture industries because of limitations in the data. Our
data on emissions from these industries were confined to a few well known contami-
nants, primarily solvents. A large number of other organic compounds and inorganic
gases (including arsine, phosphine, and boron compounds) are also used in these
industries, but we were unable to develop emissions estimates for them.
We have tried to focus our efforts on the more important substances, based
on our current understanding of toxic pollutants, but urban air is characterized
by the presence of hundreds of substances for which little information is available.
The short-term monitoring for organic gases conducted as part of Stage I (discussed
in greater detail elsewhere in this chapter) provides some evidence that a wide
variety of organic gases besides those examined in this study are present in the
ambient air in Santa Clara Valley. As such, it is possible that analyses such
as the IEMP air analysis, which focus on a select set of compounds, may account
for only a small percentage of health risk from airborne exposures.
Exposure data varied for the different classes of pollutants. The EPA has
analyzed samples of ambient air from an AQMD-run monitoring station in San Jose
for the metals over the past several years. The AQMD was also able to estimate
emissions for the organic gases we studied, from both large point sources like
factories and also from dispersed, small sources (called "area sources"),
including gasoline stations and motor vehicles. Data on organic particulates
were extremely sparse, but we included them in the analysis because a nationwide
analysis of air toxics conducted by EPA had shown that they could be significant
contributors to airborne risk. (Office of Policy Analysis, "The Air Toxics Problem
in the United States: An Analysis of Cancer Risks for Selected Pollutants,"
USEPA, May, 1985.)
The IEMP analysis attempted to consider all possible sources of these
pollutants, including area sources and certain types of sources which were
only recently discovered to be potentially significant emitters of air toxics
(for example, water pollution control plants, groundwater aeration facilities,
and landfills). The source categories examined are listed in table 3-2. The
AQMD inventory allows us, for the organic gases, to estimate the contributions
of different categories of sources to total ambient exposures in the Valley.
Our ability to do this for the metals and organic particulates is extremely
limited, however, since data on the sources of these pollutants are scarce and
largely speculative.
Overview of the Methodology
For all three classes of air pollutants, we estimated ambient levels as a
first step toward estimating risk: the methods we used to estimate those levels,
however, varied according to the available data. Figure 3-1 illustrates the
methodologies used for metals, gases and particulates.
For metals, we have ambient monitoring data from the monitoring station run
by the AQMD, and use those data directly to characterize ambient conditions.
-------
3-6
Table 3-2
Source Categories Considered in the
Santa Clara Valley IEMP Air Analysis
Point Sources
Large industrial sources
Water pollution control plants
Groundwater aeration facilities
Sanitary landfills
Area Sources
Small industrial sources
Gasoline stations
Dry cleaners
Mobile sources (cars and trucks)
-------
FIGURE 3-1 METHODOLOGY FOR ESTIMATING EXPOSURE TO
TOXIC AIR POLLUTANTS
ORGANIC GASES
o
ORGANIC
PARTICULATES
EXTRAPOLATIONS
EXPOSURE
ASSESSMENT
SHORT
TERM
MONITORING
AMBIENT
CONCENTRATIONS
-------
3-£
Since data were available from only one monitoring station in the County, we had
to rely on assumptions about how representative these data were of metals levels
elsewhere in order to estimate Valley-wide exposures.
No monitoring data for the organic gases of interest were available when
IEMP began in early 1984. While the project did collect a one week data set
from both stationary monitors and a mobile sampling unit, these data were
collected over such a short period that we could not, in general, use them as
quantitative indicators of annual average ambient conditions. Instead, we used
the AQMD's estimates of emissions of those gases; we then used a computer
model to generate estimates of ambient conditions based on those emissions
estimates. The model used to estimate organic gas concentrations begins with
estimates of emissions and tracks pollutant dispersion assuming no chemical
reactions. Because of this simplifying assumption, resulting estimates do not
account for compounds that may be formed or rapidly destroyed in the atmosphere.
We used the short-term monitoring results as a rough check on the model results.
For organic particulates, neither site-specific ambient monitoring data
nor site-specific estimates of emissions were available. We therefore had to
estimate ambient conditions based on levels found in comparable metropolitan
areas. These estimates of ambient levels are obviously subject to considerable
uncertainty and have larger error bounds than the estimates for organic gases
or metals.
Using the estimates of ambient levels of air toxics in all three categories,
ws then calculated ambient exposures to Valley residents and resultant risks. In
estimating individual risks attributable to air toxics, we used toxicological
information as discussed in chapter 2. To estimate aggregate increased incidence
of disease, we combined the information on individual risks with information on
population and population density provided by the Association of Bay Area
Governments (ABAC).
As noted in chapter 2, we assume that individual risk is the product of
exposure and potency. Exposure is estimated by making standard assumptions
about the relationship between ambient concentrations and actual exposure.
For instance, for air exposures the assumption is that an average person weighs
70 kilograms (or roughly 154 pounds) and breathes 20 cubic meters of air each
day (20 mVday). Potency is estimated by the slope of a dose-response curve,
which is assumed to be linear at the low levels of exposure typical for the
ambient environment. For each presumed carcinogen, therefore, we simply multiply
the estimates of individual exposures by the slope of the dose-response curve
in order to obtain a conservative estimate of individual risk.
For most other chronic toxic effects, we calculate human thresholds
(exposure levels below which no increase in risk is expected) and compare the
estimated exposure levels to those thresholds.
In estimating both thresholds and dose-response slopes, we use procedures
consistent with the proposed guidelines for risk assessment published by EPA
("Proposed Guidelines for Carcinogen Risk Assessment; Request for Comments,"
Federal Register Part VII vol. 49, No. 227, pp.46294-46331 [November 23, 1984].)
The procedures incorporate numerous assumptions that are discussed in chapter 2
of this report, in the risk assessment guidelines, and elsewhere in the risk
assessment literature. The assumptions are intended to be conservative; that
-------
3-9
is, we make the assumptions in such a way as to minimize the likelihood that we
are underestimating the risk.
We assume that we can multiply individual risks by exposed populations in
order to derive estimates of aggregate incidence of disease. We also assume that
risks of a particular health effect from exposure to several different pollutants
can be added in order to estimate risk to an individual exposed simultaneously
to many pollutants.
Table 3-3 presents data on the careinogenicity of the substances analyzed in
this chapter. Two types of information are presented: (1) a qualitative estimate
of the strength of evidence that a substance is carcinogenic; and (2) where a
substance is a suspected carcinogen, a quantitative estimate of the substance's
potency. To summarize the strength of evidence, we present the classification
scheme developed by EPA based on a similar classification scheme used by the
International Agency for Research On Cancer (LARC). The EPA classification
scheme is discussed in the proposed guidelines for carcinogenic risk assessment
referred to above. We briefly discuss these schemes in chapter 2. Quantitative
potency estimates are from EPA's Carcinogen Assessment Group (CAG).
Sane of the substances of concern as air toxics are chlorinated hydrocarbons.
These substances include carbon tetrachloride, chloroform, dichloroethane,
dichloroethylene, methylene chloride, perchloroethylene, trichloroethane, and
trichloroethylene. As noted on table 3-3, the evidence on carcinogenicity for
many of the chlorinated hydrocarbons is not as strong as it is for benzene or
for some of the metals. Considerable debate exists about whether some of
these chlorinated hydrocarbons are carcinogenic in humans. Much of the positive
evidence for these substances' carcinogenicity is the occurrence of liver
tumors in exposed laboratory mice. However, the mice used in these experiments
have a high background rate of liver tumors, and the scientific community
disagrees on the significance of these findings. While definitive scientific
consensus on the carcinogenicity of most of these substances has not been
reached, we follow EPA policy in this report in designating a substance as
"carcinogenic in humans," "probably carcinogenic in humans," or "possibly
carcinogenic in humans."
EPA's current policy is that 1,1,1-trichloroethane (TCA) should not be
considered a possible human carcinogen, and we follow that policy as our base
case. EPA formerly considered TCA a possible carcinogen, but has suspended
that classification because a key implicating study is currently under review.
Because of the uncertainty on this issue, and pending the outcome of the review,
we have performed sensitivity analyses of the possible impact of TCA if it
were a carcinogen. The value of this analysis is that it can indicate the
importance, in terms of local risk assessment, of resolving the uncertainty
on this issue. The sensitivity results, which appear in footnotes to the text
and tables, should be regarded as extreme upper-bound values. They are not
part of our base-case analysis.
Note that our analysis of cancer risk includes both benzene and gasoline
vapors. Although benzene is a constituent of gasoline, the type and site of
tumors associated with exposure to benzene appear to be different from those
associated with exposures to gasoline vapors. Because this indicates that the
cancer risk from benzene alone nay not account for the potential cancer risk
from exposures to gasoline vapors as a whole, we have included both in the
-------
3-10
Table 3-3
Carcinogenic Potency Values and Strength of Evidence
For Substances in the Santa Clara Valley
IEMP Air Analysis
Substance
Human
Level of Evidence*
Animal Classification
Potency Value**
per (ug/m3)-l
Metals
Arsenic
Barium
Beryllium
Cadmium
Chromium +6 (2)
Lead
Nickel
Zinc
S
-
I
L
S
-
S
-
I
-
S
S
S
-
S
-
A
-
B2
Bl
A
-
A
-
4.3 x 10-3
(1)
2.4 x 10-3
1.8 x 10-3
1.2 x 10-2
(3)
3.0 x 10-4 (4)
(1)
Organic Gases
Benzene S
Carbon tetrachloride I
CFC-113
Chloroform I
Dichlorobenzene -
1,2 Dichloroethane I
1,1 Dichloroethylene I
Ethylene dibromide I
Ethylene oxide L
Gasoline vapors I
Glycol ethers -
Isopropyl alcohol -
Methylene chloride I
Perchloroethylene I
Phenol
1,1,1-Trichloroethane -
Trichloroethylene I
Toluene
Xylene
Organic Particulates
Benzo(a)pyrene I
S
S
S
L
S
S
S
S
S
A
B2
B2
B2
C
B2
Bl
B2
B2
B2
B2
B2
8.0 x
1.5 x
(1)
2.3 x
(1)
6 x
0 x
x
0 x
0 x
(7)
(1)
4.1 x
4.8 x
(8)
(9)
1.3 x
(1)
(1)
2
5
2.2
1
7
10-6
10-5
10-5
10-5
10-5
ID'4 (5)
10-4
10-7 (6)
10-6
10~7
10-6
3.3 x 10-3
* EPA Classification: A = carcinogenic in humans? Bl = probably carcinogenic in
humans (limited human evidence); B2 = probably carcinogenic in humans (insufficient
human evidence but sufficient animal evidence); C = possibly carcinogenic in humans.
S = sufficient evidence; L = limited evidence; I = inadequate evidence.
**SOURCE: EPA's Carcinogenic Assessment Group (CAG).
upper-bound estimates of human effects.
Potency values are plausible
FOOTNOTES TO TABLE 3-3 APPEAR ON THE NEXT PAGE
-------
3-11
FOOTNOTES TO TABLE 3-3
At this time, there appears to be insufficient data for an adequate
evaluation of the potential carcinogenicity of this compound; therefore,
it is not analyzed as a carcinogen in this report.
Non-hexavalent chromium is not considered to be a carcinogen, hence there
is a potency value only for hexavalent chromium.
While there is limited evidence that some lead compounds induce tumors in
experimental animals, GAG has not yet made a determination as to its
potential carcinogenicity; therefore, lead is not analyzed as a
carcinogen in this analysis.
This potency value is for nickel from refinery dust; the GAG score for
nickel subsulfide is 6.0 x 10"^. This analysis uses only the score given
in the table; see text.
The potency value for EDB is based on studies of ingestion, not inhalation.
A more appropriate inhalation value is being generated and will, most
likely, be lower.
This potency value was converted from a value of 3.1 x 10~3 ppn assuming
that the average molecular weight of gasoline vapors is 110. See text for
a full discussion of the uncertainties involved in assessing the carcinogenic
potential of gasoline vapors and the use of the CAG value in this analysis.
No data regarding the potential carcinogenicity of these compounds were
located in the available literature. For that reason, an evaluation of
their carcinogenic potential is not possible.
No inhalation data regarding the potential carcinogenicity of this compound
were located in the available literature. For that reason, an evaluation
of the carcinogenic potential via inhalation is not possible.
In accordance with current EPA policy, we do not consider TCA to be a
carcinogen in our base case; for sensitivity analysis, however, we do
examine the impact of TCA if it were a carcinogen. For such alternatives,
which appear in footnotes to the text and tables, the potency value used
is 4.6 x 10~7, an extreme upper-bound value. See text.
-------
3-12
analysis. However, there is substantial uncertainty over the carcinogenicity
of gasoline vapors. EPA's Carcinogen Assessment Group (CAG) has calculated an
upper-bound cancer potency value for gasoline vapors. Much of the evidence of
carcinogenicity comes from a two-year animal study sponsored by the American
Petroleum Institute (API). Although the study was well conducted, its relevance
to human risk assessment is uncertain. One issue is that the animal models
used in the study may be highly susceptible to the observed effect, thereby
diminishing the relevance of the findings to humans. For purposes of this
analysis, we follow EPA policy and rely on the strength-of-evidence criteria
presented in table 3-3.
Another uncertainty centers on the question of vapor content. In the API
study, the gasoline was wholly vaporized for animal exposure; that is, the inhaled
mixture was identical to that in the liquid phase. However, this mixture is not
representative of the evaporative mix found in ambient situations. Some of the
larger hydrocarbon molecules comprising the liquid gasoline mixture are less
volatile and will therefore be present in lower proportions in ambient vapors.
The importance of this distinction is that certain subsets of the higher molecular
weight compounds are the ones that appear most likely to be responsible for toxic
effects observed in the male rats in the API study. As such, the API study may
overstate the toxic potential of gasoline vapors in the ambient environment.
Studies suggest that the wholly vaporized mixture may overestimate concentrations
of these higher molecular weight compounds by a factor of three to five relative
to the measured concentrations in the ambient environment. (For example, see Health
Effects Institute, "Gasoline Vapor Exposure and Human Cancer: Evaluation of Existing
Scientific Information and Recommendations for Future Research," September 1985.)
As a result, based on such findings and the professional judgement of EPA's
Office of Air Quality Planning and Standards, for our base-case analysis we
have divided our risk estimates by a factor of four. However, when appropriate,
we also indicate what the full range of possible risk might bewhich ranges from
zero to the undivided estimate based on the upper-bound CAG value.
A final consideration in using the CAG value for gasoline vapors is that the
potency estimate is given in terms of parts per million (ppm) whereas the exposure
information is in micrograms per cubic meter (ug/m3). in order to convert the
CAG value to the appropriate units, we need to know the molecular weight of the
vapors. However, gasoline vapors are composed of a mix of compounds of varying
molecular weight. In keeping with the professional judgement of EPA's Office of
Air Quality Planning and Standards, for purposes of this conversion we assume
that the average molecular weight of gasoline vapors is 110.
As indicated on table 3-3, for some compounds there appears to be insuffi-
cient data at this time for an adequate evaluation of their potential carcino-
genicity. While there is limited evidence that some lead compounds induce
tumors in experimental animals, CAG has not yet made a determination as to its
potential carcinogenicity. For other compounds, no data were located in the
available literature regarding their potential carcinogenicity. For these
reasons, a number of compounds included in this report are not not analyzed as
carcinogens.
To examine the possible chronic non-cancer health effects of the chemicals
we are analyzing, we compare monitored or modeled concentration levels with
estimated human thresholds, below which adverse effects are assumed not to
occur. These thresholds are based on both EPA-developed reference doses (RfDs?
-------
3-13
also referred to as Acceptable Daily Intakes or ADIs) and additional health
effects thresholds estimated by IEMD toxicologists and consultants for this
study. The IEMD thresholds were derived using the same procedures used in
estimating RfDs; however, these thresholds have not been subjected to EPA
review.
Possible non-cancer health effects, estimated quantitative thresholds, and
information sources are presented on table 3-4. Where evidence for an effect
exists but review of the literature or quantification of the threshold is not
complete, the effect is listed and no threshold estimate presented. *
* A limited number of studies have been conducted to evaluate the potential
for adverse effects on the fetus from exposure to TCA. These studies are sum-
marized in U.S. EPA 1984, Health Assessment Document for 1,1,1- Trichloroethane
(methyl chloroform), Final Report, EPA-600/8-82-003F, Washington, D.C., and in
IEMD, "Health Score Evaluations for Pollutants in the Santa Clara Valley
Integrated Environmental Management Program: 1,1,1-Trichloroethane," updated
November 1985. In its Health Assessment Document, U.S. EPA indicates that it
is not possible, on the basis of limited data, to define the full potential of
TCA to produce teratogenic effects. Each of the available mammalian studies
had methodological drawbacks that do not allow for conclusive evaluation of the
ability of TCA to produce a teratogenic response over a wide range of doses.
All of the studies performed in mammals at the time the Health Assessment
Document was published had been done in rats and mice. On the basis of these
studies, it does not appear that short- or long-term exposure to TCA results in
teratogenic effects in rats or mice. Therefore, the lEMP's base-case analysis
assumes that exposure to TCA poses no risk of fetal effects.
An unpublished study, which has not undergone scientific peer review and
which was completed subsequent to the publication of EPA's final Health Assess-
ment Document for TCA, reports fetotoxic effects (cardiac malformations) in rat
pups exposed jln utero to TCA (Dapson et al., 1984). Given that this study was
extremely limited (using only one dose level) and that the resultsif
reproduciblecould have significant implications, the National Toxicology
Program (NTP) has commissioned a project to repeat the Dapson study using
multiple dose levels. Results of the NTP study are expected in Fall 1986.
In order to assess the importance to Santa Clara Valley residents of
further research on this issue, we need to examine whether local exposures to
TCA may be of concern if the Dapson findings are validated. For this reason,
IEMD toxicologists have calculated a threshold of 16 ug/m^ based on the Dapson
study to be used in this report as a sensitivity case pending the outcome of the
NTP study. The sensitivity analysis examines the possible impact of TCA under
the alternative assumption that exposures above the estimated threshold could
pose the risk of cardiac malformations. This analysis, which appears in foot-
notes to the text, is not part of our base-cases analysis. THE SENSITIVITY
RESULTS SHOULD NOT BE INTERPRETED AS INDICATING WHETHER OR NOT A RISK IN FACT
EXISTS; EPA RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK MANAGEMENT DECISION-
MAKING OR REGULATORY ACTION pending the outcome of the follow-up study by NTP-
-------
Table 3-4
Non-Cancer Health Effects and Presumed Human Thresholds for
Substances in the Santa Clara Valley IEMP Air Analysis
Substance
Health Effect
Presumed Human
Threshold (ug/m^)
Source (1)
Scientific Evidence (2)
Metals:
Arsenic
multiple organ effects
liver
neurobehavioral
skin
arsenic poisoning
reproductive
fetal effects
13.3
(3)
13.3
13.3
(3)
(3)
(3)
lEMD/RfD
IEMD
IEMD
Heywood & Sortel, 1979
Clement Iron and Steel
(I&S) study
Clement I&S
Perry et al '48
(in NIOSH '75)
Zaldivar 1977
Hood et al 1977
Hood et al 1977
u>
Barium
Beryllium
Cadmium
hypertension &
card iovascular
decreased body weight
& growth
lung disease
kidney
liver
respiratory
fetal effects
reproductive
mutagenicity
175.0
1.9
(4)
0.24
0.42
2.0
21.0
119.0
(5)
lEMD/RfD (v)
lEMD/RfD (v)
lEMD/RfD (v)
IEMD
IEMD
IEMD
IEMD
Perry et al 1983;
Brenniman et al 1979a,b
Schroeder & Mitchner, 1975
Kjellstron et al 1977
Friberg 1950
Comtn. European
Communities 1978
Schroeder & Kitchener 1971
Scharpf et al 1972
-------
Table 3-4, cont.
Substance
Health Effect
Presumed Human
Threshold (ug/nP)
Source
Scientific Evidence
Chromium +3
Chromium +6
Lead
Nickel
(6)
liver
k idney
fetal effects
(6)
liver
nasal
reproductive
fetal effects
(7)
fetal effects
respiratory
cardiovascular
skin
reproductive
3500.0
3500.0
3500.0
3500.0
17.0
17.0
17.0
17.0
17.0
3.5
3.5
3,5
3.5
3.5
ZEMD/RfD
IEMD
IEMD
IEMD
lEMD/RfD
IEMD
IEMD
IEMD
IEMD
lEMD/RfD (v)
IEMD
IEMD
IEMD
IEMD
Ivankovic & Preussman
1975
Clement I & S
Clement I & S
Clement I & S
ADL-Petroleum
MacKenzie et al 1958
NIOSH 1975
NIOSH 1975 ,
Gale 1978
Gale 1978 l
Ambrose et al 1976
Ottolenghi et al 1974
Schroeder et al 1974
U.S. EPA 1985b; Nriagu 1980
Ambrose et al 1976
U)
Zinc
reproductive
1540.0
IEMD
Schlicker & Cox 1968
-------
Table 3-4, cont.
Substance
Health Effect
Presumed Human
Threshold (ug/m3)
Source
Scientific Evidence
Organic Gases:
Benzene
CFC-113
Carbon Tetrachloride
Chloroform
Dichlorobenzene
blood effects
fetal effects
psychonotor
impairment
liver
neurobehavioral
kidney
reproductive
fetal effects
liver
kidney
reproduct ive
fetal effects
neurobehavioral
mutagenicity
liver
neurobehavioral
respiratory
cardiovascular
kidney
blood effects
2.45
4.1
105000.0
2.4
2.4
108.
430.
24.2
35.0
22.5
2.43
2.43
11.7
(8)
315.0
315.0
315.0
315.0
315.0
315.0
lEMD/RfD
IEMD
lEMD/RfD (v)
lEMD/RfD
IEMD
IEMD
IEMD
IEMD
lEMD/RfD
IEMD
IEMD
IEMD
IEMD
lEMD/RfD (v)
IEMD
IEMD
IEMD
IEMD
IEMD
Snyder et al 1980
Kuna & Kapp 1981
Imbus & Adkins 1972
EPA 1980 & 1984
Smyth 1935 & 1936
Moller 1973
EPA 1980
Adams et al 1952
Schwetz et al 1974
Heywood et al 1979
Heywood et al 1979
Schwetz et al 1974
Schwetz et al 1974
Challen et al 1958
USEPA 1985
NTP 1982
Hollingsworth et al 1956
Hollingsworth et al 1956
NTP 1982
NTP 1982
Varashavskaya 1967;
NTP 1982
-------
Table 3-4, cont.
Substance
Health Effect
Presumed Human
Threshold (ug/m^)
Source
Scientific Evidence
1 , 2-Dichloroethane
1,1 Dichloroethylene
Ethylene Dibromide
Ethylene Oxide
liver
neurobehavioral
gastrointestinal
kidney
liver
kidney
mutagenicity
reproductive (male)
reproductive (female)
liver
kidney
respiratory-nasal
mutagenicity
embryotoxicity
neurobehavioral
blood
respiratory
reproductive
kidney
mutagenicity
26.0
26.0
26.0
26.0
31.0
2.5
(5)
1.75
11.9
12.8
12.8
944.
(5)
(9)
(9)
(9)
(9)
(9)
(9)
(9) (5)
lEMD/RfD
IEMD
IEMD
IEMD
lEMD/RfD (v)
IEMD
IEMD
IEMD
IEMD
IEMD
IEMD
Kozik 1957
Kozik 1957
USEPA 198 2a
Heppel et al
Hoffman et
McKenna et al
NTP 1982
Hurtt & Zenick
NTP 1982
1946
al 1971
1972
1985
Short et al 1977
NTP 1982
NTP 1982
NTP 1982
Gasoline Vapors
Glycol Ethers:
2-e thoxyethanol
2-tnethoxyethanol
2-bu toxye thanol
(10)
(11)
(11)
(11)
I
t'
^1
-------
Table 3-4, cont.
Substance
Methylene Chloride
Perchloroethylene
Phenol
1,1, 1-Trichloroethane
Tr i chloroe thylene
Health Effect
liver
blood
fetal effects
neurobehavioral
kidney
mutagenicity
liver
kidney
fetal effects
liver
liver
respiratory
neurobehavioral
fetal effects
liver
neurobehaviora 1
kidney
Presmed Human
Threshold (ug/m3)
210.0
8600.0
210.0
8600.0
699.0
(5)
69.9
69.9
909.0
350.0
97.9
1050.0
5500.0
(13)
26.0
26.0
3770.0
Source
lEMD/RfD (v)
IEMD
IEMD
IEMD
IEMD
lEMD/RfD (v)
IEMD
IEMD
lEMD/RfD (v)
lEMD/RfD
IEMD
IEMD
lEMD/RfD
IEMD
IEMD
Scientific Evidence
Burek et al 1980 & 1984
EPA 1984
Burek et al 1984
NIOSH 1976
Schwetz et al 1975
USEPA 1985a
NIOSH 1976
NTP 1985 Draft
Coler & Rossmiller 1953
NCI 1977
Nelson et al 1979
Dow Chemical 1976
McNutt et al 1975
Torkelson et al 1958
NIOSH 1976
EPA 1984
Grand jean et al 1955
Tucker 1982
CO
-------
Table 3-4, cont.
Presumed Human
Substance
Toluene
Health Effect
blood
liver
neurobehavioral
Threshold (ug/rtr>)
1050.0
1010.0
580.0
Source
lEMD/RfD (v)
IEMD
IEMD
Scientific Evidence
CITT 1980
USEPA 1983
Hanninen et al 1976
Xylene
respiratory
kidney
reproductive
fetal effects
liver
neurobehavioral
respiratory
cardiovascular
blood
kidney
reproductive
fetal effects
1010.0
1010.0
500,0
476.0
215.0 (14)
215.0
215.0
215.0
215.0
215.0
52.8
52.8
Seppalainen et al 19
IEMD Vbn Dettingen et al 19
Bruckner & Peterson 1
IEMD USEPA 1983
IEMD Matsumoto et al 1971
IEMD Hudak and Ungvary 1978
lEMD/RfD Bowers et al 1982
IEMD USEPA 1984
Savolainen et al 197
IEMD Hipolito 1980
Smyth & Smyth 1928
IEMD Hipolito 1980
Hirsch 1932
IEMD Browning 1965
Hipolito 1980
IEMD USEPA 1984
IEMD Ungvary et al 1980
IEMD Ungvary et al 1980
Organic Particulate;
Benzo(a)pyrene
reproductive
fetal effects
rautagenicity
(4)
(4)
(5)
Clement HEIS
Clement HEIS
-------
FOOTNOTES TO TABLE 3-4
(1) "lEMD/RfD" indicates an inhalation threshold derived by IEMD toxicologists based on a route-to-route extra-
polation from an EPA "Reference Dose" (RfD) or "Acceptable Daily Intake" (ADI) level for ingestion. RfDs and
ADIs are estimated no-effect thresholds that are intended to protect an individual from the most potent non-
cancer health effect.
"IEMD" indicates an inhalation threshold derived by IEMD toxicologists and consultants using existing litera-
ture and following EPA procedures. These thresholds have not been subjected to EPA review.
"(v)" indicates that an RfD is verified. A verified RfD represents a consensus value agreed upon by represen-
tatives of various EPA program offices. The verification process does not constitute peer review; it is an
internal EPA review. An unverified RfD is one which has been calculated and is used by an EPA program office,
but which has not yet been examined by the internal review group.
(2) For sake of brevity, negative evidence (i.e., where a laboratory test or epidemiological study has
been performed, but no evidence was found of a health inpact) is not reported. See the IEMD Health Score
Evaluation Documents for each chemical for a full discussion of the scientific evidence. u>
K)
(3) There is evidence of these effects via ingestion. However, no data were located to indicate whether or not °
these effects would be associated with exposure via inhalation. Therefore, at this time it is not appropriate
to perform a route-to-route extrapolation to estimate such effects via inhalation.
(4) Quantification of the threshold is not complete at this time.
(5) Mutagenic effects cannot be meaningfully expressed in quantitative terms for purposes of evaluating human
health risk. Therefore, we note positive evidence of mutagenicity by listing it as an effect.
(6) This is the highest dose given at which no adverse effect was seen; therefore this RfD-derived
threshold has no particular adverse health effect as its basis. (Based on NOEL, but not LOEL;
see chapter 2.)
(7) There is strong scientific evidence that airborne lead can have adverse health effects; however, the
evaluation of effects of chronic lead exposure is complex since the effects depend on the total body burden
from air, water and dust exposure. No EPA reference dose identifying a no-effect threshold has yet been
estimated. Information on the health impacts of lead is summarized in the Criteria Document ("Air Quality
Criteria for Lead," External Review Draft EPA 600/8-83-028B, 1984). Because of the issue's complexity and
the lack of an estimated no-effect threshold, we have not estimated health effects from lead in this report.
See text.
-------
FOOTNOTES TO TABLE 3-4, cont.
(8) On the basis of presently available data, no definitive conclusion can be reached concerning the mutagenicity
of this substance.
(9) These health effects have been tentatively identified as potential adverse effects associated with exposure
to ethylene oxide. Review of the literature and quantification of the thresholds are not complete at this
time.
(10) Evaluation of potential non-cancer health effects is not complete at this time. There is currently no RfD.
(11) There is no RfD or IEMD derived threshold for glycol ethers. See text for a discussion of the available
information on non-cancer health effects.
(12) Review of the literature regarding non-cancer health effects is not complete at this time.
(13) The IEMP conducted sensitivity analysis on TCA for possible fetal effects. See footnote to text.
U)
(14) For xylene there are two unverified RfD values for all health effects except reproductive and fetal effects. N>
One value is listed above; the other is 35 ug/m^. In addition to comparing exposures to the listed value,
we will use this second value in sensitivity analyses.
-------
3-22
As noted on the table, there is no RED or IEMD derived threshold for
glycol ethers. However, EPA's Environmental Criteria and Assessment Office
(ECAO) has calculated Acceptable Intake Chronic (AIC) values for glycol ethers
for use by EPA's Superfund program. The AIC is similar in concept to ADIs and
RfDs. It is an estimate of an exposure level that would not be expected to
cause adverse effects when exposure occurs for a significant portion of the
lifespan. The AIC is route-specific and estimates acceptable exposures for a
given route with the implicit assumption that exposure by other routes is
insignificant. The AIC values are presented in Appendix C of the "Draft
Superfund Public Health Evaluation Manual," December 18, 1985. Full supporting
documentation for the derivation of the AIC values for glycol ethers is found
in the Draft "Health Effects Assessment for Glycol Ethers," prepared for the
Office of Emergency and Remedial Response by ECAO, EPA/540/1-86-052, September
1984,
The estimated thresholds, based on the AIC, are 170 ug/m^ for 2-ethoxy-
ethanol (this is based on a NOEL; health effects of possible concern at higher
doses may include upper respiratory irritation, blood effects, and fetal effects);
84 ug/m^ for 2-methoxyethanol (this is based on a NOEL; health effects of
possible concern at higher doses may include testicular degeneration, blood
effects, and fetal effects); and 56 ug/m^ for 2-butoxyethanol (this is based
on a NOEL; health effects of possible concern at higher doses may include
blood effects, decreased body weight, and fetal effects). These values may
provide some insight into the possible risks from estimated exposures pending
the calculation of an RfD or IEMD threshold.
No non-cancer health effects information for lead is included in
table 3-4. There is, in fact, strong scientific evidence that airborne lead
can have adverse health impacts, particularly for children. These health
effects include blood-related problems and neurological dysfunctions, including
potential impairment of intelligence as measured in IQ tests.
Evaluation of effects from chronic lead exposure is complex since effects
depend on the total body burden accumulated from air, water and dust exposure.
No EPA reference dose identifying a no-effect threshold has yet been estimated.
Information on the health impacts of lead is summarized in the Criteria Document
("Air Quality Criteria for Lead," External Review Draft, EPA-600/8-83-028B,
1984). Because of the issue's complexity and the lack of an estimated no-effect
threshold, we were unable to estimate health effects from lead exposure in this
report. The IEMP hopes to analyze such effects as part of its Stage II analysis
of criteria pollutants.
-------
3-23
Methodology and Findings; Metals
Our analysis of metals in Santa Clara Valley considered eight compounds:
arsenic, barium, beryllium, cadmium, chromium, lead, nickel, and zinc. (Arsenic
is technically a metalloid, not a metal; but we refer to all eight substances
throughout this discussion as "metals").
The discussion of metals is in four parts. The first section describes
the levels of metals monitored in Santa Clara Valley's outdoor air. The second
section discusses the possible sources of metals in the Valley. In the third
part, we discuss the way in which we estimated exposure to metals. Finally,
we present the estimated health risks from exposure to metals. The risk results
are presented in two ways: first by pollutant and then by source.
Monitored Levels of Metals in Santa Clara Valley's Outdoor Air
To estimate the ambient levels of the eight metals in the outdoor air in
Santa Clara Valley, we used long-term monitoring data collected by the Bay
Area Air Quality Management District (AQMD) between 1977 and 1982. The AQMD
operates a network of monitoring stations in the Bay Area, some of which have
samples analyzed for metals levels. However, only one of the AQMD stations
whose samples are analyzed for metalsthe Fourth Street monitoring station in
San Joseis in the IEMP study area. Thus the metals data come from this one
station only.
In estimating ambient levels of metals from the AQMD data, we used annual
average values, which are the average of many samples (usually about thirty)
taken over the course of a given year. Only averages which met summarization
criteria set by EPA's Office of Air Quality Planning and Standards were included.
Annual average values which are based on only a small number of data points do
not meet those criteria. The values we used were from the years 1977-1982; more
recent monitoring data did not meet the criteriathere were too few observations.
However, except in the case of arsenic and beryllium, the more recent observations
were within the range of annual averages presented for previous years. In the
case of arsenic and beryllium, the more recent observations were outside the
range of annual averages presented, but were still within the range of single
values observed in previous years.
Table 3-5 presents the annual average concentrations and the single highest
values observed for the eight metals based on the Fourth Street monitoring
data. As noted, the range of arsenic levels and the lower end of the range of
levels for barium, beryllium and zinc were below the detection limits for the
analytical techniques used. The detection limit for arsenic is 0.0055 ug/m3;
for barium it is 0.003 ug/m3; for beryllium it is 0.00006 ug/m3; and for zinc it
is 0.0133 ug/m3.
When the metals concentrations were below the detection limit and there
was reason to believe the metal is present in Santa Clara County's anbient air
(for instance, if it had been detected in other samples), EPA used the conveneion
of assigning a value of one-half the detection limit to that pollutant. The
lower end of the range of concentrations for arsenic, barium, beryllium, and
zinc is, in each case, one-half of the detection limit for that pollutant. For
arsenic, the higher end of the range of concentration is also below the detection
limit of 0.0055 ug/m3, but it is higher than one-r.alf the detection limit
-------
3-24
Table 3-5
Ambient Levels of Metals at the Fourth Street
Monitoring Station in Santa Clara Valley
All values are in micrograms/cubic meter
Substance
Range of Annual Average
Concentrations, 1977-1982(1)
Highest Measured
Concentration, 1976-1983
Arsenic
Barium
Beryllium
Cadmium
Chromium
Lead
Nickel
Zinc
0.0027 to 0.0041 (2)
0.0015 to 0.0358 (3)
0.00003 to 0.00014 (3)
0.001 to 0.003
0.0126 to 0.0138
0.46 to 1.62
0.0058 to 0.009
0.0066 to 0.1075 (3)
0.017
0.067
0.0006
0.012
0.036
6.56
0.04
0.26
1 Only annual average values which met the summarization criteria of EPA's
Office of Air Quality Planning and Standards are included. Annual average
values which are based on only a small number of data points do not meet
those criteria.
2 This range of values is below the detection limit for the analytical equipment
usedl See text for a discussion of this issue.
3 The lower end of this range is below the detection limit for the analytical
equipment used. See text for a discussion of this issue.
-------
3-25
indicating that for some samples collected, arsenic was in fact detected. Those
levels of detection were then averaged with the other, lower values, putting the
annual average below the detection limit.
Of the eight metals, beryllium was found in the lowest concentrations,
with annual average levels of 0.00003 to 0.00014 ug/m3- Lead had the highest
annual average monitored concentrations, ranging from 0.46 to 1.62 ug/m-*- An
examination of the annual summaries of the monitoring data (available in
Appendix 3-A) reveals a trend of decreasing lead concentrations over the years
with a high annual average of 1.62 ug/m3 in 1977 to 0.54 ug/m3 in 1982.
Table 3-6 compares annual average concentrations of selected metals
monitored at the Fourth Street station in San Jose to levels found in other
industrialized cities. In most cases, the annual average ambient levels of
metals monitored at Fourth Street are comparable to, or slightly below, the
annual average monitored levels of metals found in heavily industrialized
cities studied by EPA. For example, monitored levels of cadmium in Santa
Clara Valley which ranged from 0.001 to 0.003 ug/m3, were comparable to the
monitored values of about 0.004 ug/m3 found in Philadelphia and Los Angeles.
Monitored levels of chromium in Santa Clara Valley were substantially
lower than those levels found in some of the other cities EPA studied. Annual
average chromium levels from 1977 to 1982 at the Fourth Street station ranged
from 0.0126 to 0.0138 ug/m3. EPA researchers found 0.06 ug/m3 of chromium in
the industrialized section of Philadelphia. Baltimore, which has a number of
large chromium sources, had even higher levels of 0,093 ug/m3. Similarly,
monitoring data for nickel from northern New Jersey, Philadelphia, Los Angeles
and Baltimore all showed levels around 0.02 ug/m3, two to three times as high
as the concentrations monitored in Santa Clara Valley which ranged from 0.0058
to 0.009 ug/m3-
Possible Sources of Metals in Santa Clara Valley
Small amounts of metal participates are present in all urban air, largely
because they are by-products of the combustion that occurs in motor vehicles,
home and commercial heating, and industrial processes. In some areas, metals
concentrations are elevated because of other large industrial sources, such as
smelters, utility plants, and iron and steel manufacturing plants, that emit
large quantities of metals directly to the air. Metals may also come from earth
crustal matter and be present in windblown and resuspended dust.
EPA has set emissions limits for some sources of some of these pollutants,
and recently lowered the allowable level of lead in gasoline. Both EPA and the
California Air Resources Board are considering further limitations on metals
emissions. In addition, other studies suggest that regulations designed to
control emissions of criteria pollutants, i.e., particulates, may indirectly
reduce emissions of toxic metals since these are generally in particulate form
(Office of Policy Analysis, "The Air Toxics Problem in the United States," USEPA,
May, 1985).
Neither EPA nor the ACMD has compiled an emissions inventory for metals in
Santa Clara County as the ACMD has for organic gases. The IEMP Draft Stage I
Report (October, 1985) identified the lack of local data on sources and emissions
-------
Table 3-6
Substance
Beryllium
Cadmium
Chromium
Nickel
Comparison of Monitored Levels of Metals
in Santa Clara Valley and in Other Cities
All values are annual average ambient levels
in micrograms/cubic meter
Santa Clara Elizabeth Philadelphia Downtown
County New Jersey Naval Hospital Los Angeles
0.00003 to 0.00014 0.00003 0.0004 0.0002
0.001 to 0.003 0.007 0.004 0.004
0.0126 to 0.0138 0.013 0.06 0.017
0.0058 to 0.009 0.02 0.023 0.019
Baltimore
(Guilford)
0.00007
0.001
0.093
0.018
r
Source: Bill Hunt, Bob Faoro, Tom Curran and Jena Muntz, "Estimated Cancer Incidence Rates for
Selected Toxic Air Pollutants Using Ambient Air Pollution Data," July 1984.
-------
3-27
of metals as a significant data gap. As pact of its recently adopted air toxics
program, the AQMD, with assistance from EPA, is planning to compile an emissions
inventory for the District, which includes Santa Clara County. However, such
information is not currently available.
In order to get a preliminary notion of the potential sources, the IEMP
made some rough estimates of emissions of metals from some selected sources:
residential heating, motor vehicles, and a cement plant. We did not consider
every possible source of metals, but rather, attempted to identify the more
important sources in terms of emissions and possible exposures.
To estimate emissions from residential heating and motor vehicles, we used
fuel consumption data from NEDS, a national data source for emissions. For
residential heating estimates, NEDS provides information on fuel consumption by
county, taking into account local factors such as the number of "heating degree
days" per year in the county (local weather conditions), the average number of
roans per dwelling unit in the county (from U.S. census data), total wood use
in the state (from Department of Energy studies), and the number of dwelling
units in the county using wood as the primary heating fuel (from U.S. census
data). For motor vehicles, NEDS has county-specific data provided by the state
on the type and number of vehicles driven (based on motor vehicle registration)
and the estimated number of miles driven per year per vehicle. By using the.-e
data and applicable emission factors, we estimated the release of the pollutants
of concern as a function of the area source category. (Memorandum to Eileen
Softer [IEMP] from Dennis Hlinka [Versar, Inc.], "Updated Santa Clara Valley
Study Report," May 12, 1986.)
The estimates of emissions from the cement plant were based on source
testing of the plant conducted by the AQMD. In addition to providing information
on emissions, the AQMD analysis provides evidence that the cement plant is a
highly controlled sourcethe removal of heavy metals by the baghouses ranges
from 99.79 to 99.99 percent efficiency, representing the Best Available Control
Technology presently available for the removal of heavy metals. (Memorandum
to Chairperson Silver and members of the Board of Directors of the BAAQMD from
Milton Feldstein, Air Pollution Control Officer, "Potential Toxic Air Contaminant
Emissions from Kaiser Permanente Cement Company," March 26, 1986.)
Table 3-7 presents the emissions estimates for metals from the source
categories we considered. Table 3-8 presents the relative contribution of
each of these source categories to emissions of the eight metals based on our
emissions estimates. Bear in mind that the information in this table is based
on estimates of emissions from only some of the possible sources of metals in
Santa Clara Valley. Although we think we have identified the more important
sources, we did not consider all possible sources; further data collection,
such as the compilation of an emissions inventory for metals as planned by the
AQMD, would be required for a comprehensive identification of potential
sources of metals. In addition, the emissions estimates for residential heating
and motor vehicles are based on rough calculations that have not been confirmed
by testing of sources. In sum, these findings are preliminary and uncertain.
-------
Table 3-7
Estimated Air Eknissions of Metals
in Santa Clara Valley by Source Category
All values are in metric tons (kkg) per year
Source Category
Jtesidential Heating (1)
Distil la ted Oil
Fireplaces
Woods toves
Heating Total
Cement Plant (2)
Road Vehicles (1)
Light Duty
Heavy Duty
Arsenic
0.00003
0.0013
0.001
0.002
0.002
ND
ND
Beryllium
0.00004
0.000001
0.000001
0.00004
0.02
ND
ND
Cadmium Chromium
0.0002 0.00002
0.00035 0.009
0.00026 0.006
0.0008 0.015
0.02 0.04
0.004 ND
0.009 ND
Nickel
0.002
0.016
0.012
0.03
0.02
ND
ND
Lead
0.0002
0.0047
0.0035
0.0084
0.02
26.0
0.14
oo
Vehicles Total
ND
ND
0.013
ND
ND
26.0
(1) Source: Memorandum to Eileen Soffer (IEMP) from Dennis Hlinka (Versar, Inc.), "Updated Santa Clara Valley
Study Report," May 12, 1986. These estimates have not been confirmed by source testing. See text.
(2) Source: Memorandum to Chairperson Silver and members of the Board of Directors of the Bay Area Air Quality
Management District (BAAQMD) frcm Milton Feldstein, Air Pollution Control Officer, BAAQMD, "Potential Toxic
Air Contaminant Emissions from Kaiser Permanente Cement Company," March 26, 1986.
ND: Not Determined
-------
3-29
Table 3-8
Possible Sources of Metals in Santa Clara Valley
Substance
Possible Source(s)
Arsenic
Barium
Beryllium
Cadmium
Chromium
Lead
Nickel
Zinc
Residential heating and the cement plant
contribute almost equally to emissions of arsenic.
Within the residential heating category, wood
burning, as opposed to oil, accounts for the bulk
of emissions.
No estimates are available.
Cement plant emissions are the primary source of
beryllium with residential heating, specifically
oil combustion, as a lesser source.
Motor vehicles and cement plant emissions are the
primary source of cadmium; residential heating
can be considered a secondary source.
The major sources of chromium are cement plant
emissions and residential heating, particularly
wood burning.
Motor vehicles are the greatest single source of
lead. Relatively small increments are added from
cement plant emissions and residential heating.
Both cement plant emissions and residential
heating, particularly wood burning, are sources
of nickel.
No estimates are available.
NOTE: The information presented in this table is based on emissions estimates
from selected sources provided in the previous table. Bear in mind
that although we think we have identified the more important sources of
metals in Santa Clara Valley, we did not consider all possible sources.
-------
3-30
Estimated Exposures to Metals
The only data we have on ambient levels of metals in Santa Clara County
come from the Fourth Street monitoring station in San Jose. Although the
AQMD operates other monitoring stations within the Valley, samples are not
analyzed for metals at these other stations. As such, there is uncertainty as
to what the ambient levels and hence exposures are elsewhere in the Valley.
If the levels of metals monitored at Fourth Street are generally representative
of levels throughout the Valley, then it is possible to estimate exposures and
resultant risks to Valley residents using the Fourth Street data.
EPA has found in its Philadelphia and Baltimore studies that unless large
point sources of metals are present, metals levels tend to be fairly constant
across any particular urban area. This makes intuitive sense since the other
sources of metals (including home heating and motor vehicles) are distributed
more or less uniformly across the area. Compared to other urban areas, Santa
Clara County is notable for the absence of large industrial sources that
contribute to airborne metals; thus, it is plausible that area sources account
for most of the metals monitored at Fourth Street.
However, the emissions estimates presented in tables 3-7 and 3-8 suggest
that at least one point sourcethe cement plantcontributes to the emissions
of metals. In terms of analyzing exposure and risk, the question is how these
estimated emissions relate to ambient concentrations of the metals included in
our analysis. If the cement plant emissions contribute substantially to ambient
levels for some of the metals, then there is greater uncertainty in applying
the Fourth Street monitoring data for those metals to the entire County.
Presumably, persons living nearer to the plant would have higher exposures and
hence higher individual risks than indicated by the Fourth Street data. Further
analysis, such as dispersion modeling and ambient monitoring would be necessary
to make better estimates of exposures. The AQMD is currently conducting a
modeling exercise to estimate exposures to metals for populations near the
cement plant. No results were available at the time of this writing. However,
based on the AQMD's knowledge of the location of the plant and the surrounding
terrain and meteorology, their understanding is that the plant's impact on
ambient levels is likely to be quite localized. As a result, on a County-wide
basis, dispersed area sources are likely to be the most significant contributors
to ambient levels of metals.
If metals in the Valley are, for the most part, from dispersed area sources,
then it is reasonable to assume that metals levels tend to be fairly constant
and apply data from the Fourth Street monitoring station to the entire County.
One piece of evidence supporting this assumption is that ambient levels of total
suspended particulates (of which metals are trace constituents and which we
likewise assume are from dispersed area sources), are relatively consistent
across monitoring stations located in different parts of Santa Clara County and
the adjacent areas. Although relatively consistent, levels of total particulates
elsewhere in the County are marginally lower than at the Fourth Street station.
As such, our risk estimates for metals based on Fourth Street data may be
marginally higher than they would be if we had metals monitoring data for the
entire County. However, using conservative estimates is consistent with the
conservative approach of this screening exercise.
For purposes of the current analysis, we decided to estimate exposures
-------
3-31
and health risks by applying Fourth Street monitoring data to the entire County.
This decision was based in part on our current knowledge of sources and the
monitored levels of TSP as discussed above. In addition, on a more practical
level, the Fourth Street monitoring data are the only data available on metals
in Santa Clara County's ambient air. Despite their limitations, these data
provide the best available basis for a rough approximation of exposure to
metals in the County. As new data become available on either ambient levels
of metals or source characterization, the analysis can be refined. Therefore,
the data presented in table 3-5 serve as the basis of our exposure estimates
for metals on a Valley-wide basis.
Estimated Health Risks from Exposure to Metals
This section discusses the estimated individual risks and the increased
incidence in disease which may result from exposures to metals in Santa Clara
County air based on the Fourth Street monitoring data. Results are presented
first by pollutant and then by source category.
An important source of uncertainty in estimating risks due to metals in
air is that the toxicity of some metals apparently varies widely according to
the particular compound of which the metal is a part, or according to the
valence state of the metal. Estimating exposures and risks is therefore
difficult because the techniques of chemical analysis which are used for the
samples of ambient air often do not distinguish among these compounds or valence
states; in any case, those compounds and valence states may change before or
during the analysis. As a result, we are forced to rely on assumptions about
the percentages of the metal which fall into the various categories of toxicity,
and this introduces considerable uncertainty into the findings.
Chromium is one of those metals that presents the problems described above.
The hexavalent form of chromium (also called hexachrome) is considered a potent
carcinogen, whereas non-hexavalent chromium is not considered a carcinogen on
the basis of human or animal studies. Unfortunately, the analytical techniques
usually used, including those used to analyze the Fourth Street samples, do not
distinguish between these forms. Thus, although we have data on chromium
levels similar to those for the other metals, our risk estimates are subject to
considerably more uncertainty.
As a result of the uncertainties, we first calculated a range of risks
assuming, at the low end, that none of the chromium found in Santa Clara Valley
air is hexavalent and, at the high end, that all of the chromium is hexavalent.
In order to narrow this range, we relied on information from other EPA analyses
on the valence state of chromium. In these studies, EPA found that the chromium
emitted by utility boilers is less than 1% hexavalent. The chromium emitted
from cement plants, on the other hand, may be up to 10% hexavalent. (Memorandum
from Bill DeVfees, ENTROPY Task Manager to Dennis Holzschuh, EPA Task Manager,
"Analytical Results of Hexavalent Chromium and Total Chromium for Samples
Submitted by EMB," May 29, 1985.) Given the uncertainties over the source of
chromium in Santa Clara Valley, and given that there is a cement plant in the
County that appears to emit at least some amount of chromium to the air, we
assumed that 10% of the chromium in Santa Clara County air is hexavalent and
calculated risks accordingly. This set of calculations, which assumes 10%
hexavalent chromium, is our base case for the analysis.
-------
3-32
Analyzing the cancer risks fron nickel poses similar difficulties. Oily
two relatively rare subspecies of nickel (nickel subsulfide and nickel carbonyl)
are considered carcinogenic. It appears that the fraction of nickel emitted in
carcinogenic form (e.g., subsulfide or carbonyl) is extremely low, perhaps on
the order of 1% nationally and potentially less than this level in Santa Clara
County. (Determination of Nickel Species in Ambient Air," Radian Corporation,
April, 1984; Tom Lahre, "Characterization of Available Nationwide Air Toxics
Emissions Data," June 13, 1985; "Locating and Estimating Air Emissions from
Sources of Nickel," March 1983.) While it is possible that other chemical
forms of nickel may be weakly carcinogenic, their potency will be much less
than the nickel subsulfide. EPA's Draft Health Assessment Document (May 1983)
discusses the complex toxicology of nickel, but further work is needed to
develop potency values for the various chemical forms of nickel. Preliminary
work along these lines is underway in EPA's Office of Research and Development
("Nickel Speciation Research Proposal," May 9, 1984).
As a result of the uncertainties and complexities associated with the
carcinogenicity of nickel, we have used a cancer potency value derived from
an epidemiological study. The value is for the mix of nickel found in refinery
dust, rather than the higher potency value which applies to pure nickel subsulfide.
The assumption is that the mix found in the refinery dust may be more represen-
tative of the mix of nickel found in ambient air than pure nickel subsulfide.
In addition, we assumed that the lower end of the range of risk of cancer from
nickel exposures is zero since it is possible that none of the nickel monitored
in Santa Clara County's ambient air is carcinogenic. As a result of the
difficulties in analyzing cancer risks from nickel, the estimates are subject
to more uncertainty than those for other metals.
Lead is one other metal included in our report which deserves special
comment. While there is limited evidence that some lead compounds induce tumors
in experimental animals, CAG has not yet made a determination as to its potential
carcinogenicity. Therefore, we do not consider lead as a carcinogen in this
analysis.
In addition, we do not analyze the non-cancer effects of lead in this
report. There is, in fact, strong scientific evidence that airborne lead can
have adverse health impacts, particularly for children. These health effects
include blood-related problems and neurological dysfunctions, including partial
impairment of intelligence as measured in IQ tests. However, evaluation of
effects from chronic lead exposures is complex since effects depend on the
total body burden accumtnulated from air, water and dust exposure. No EPA
reference dose identifying a no-effect threshold has yet been estimated.
There are, however, Federal regulations which have been adopted to control
human exposures to lead. For example, lead is a criteria pollutant for which
EPA has established a national ambient air quality standard of 1.5 ug/ro3- This
standard is currently under review by EPA.
In addition, EPA has recently promulgated a rule to reduce the amount of
lead in gasoline from its previous limit of 1.10 grams per leaded gallon (gplg)
to 0.10 gplg effective January 1, 1986. The primary objective of this rule is
to minimize the adverse health and environmental effects associated with lead
in gasoline. Previous EPA restrictions on lead levels in gasoline resulted in
measurable decreases in the levels of lead in ambient air. The most recent
-------
3-33
action should result in further declines in ambient lead levels over time,
with accompanying reductions in the risks due to exposures to the lead.
Other Federal activities designed to control exposures to lead include EPA's
national primary drinking water standard for lead and the Department of Housing
and Urban Development's goal to eliminate lead-based paint in federally funded
housing.
Because of the issue's complexity and the lack of an estimated no-effect
threshold, we were not able to estimate health effects from lead in this report.
The IEMP hopes to analyze such effects as part of its Stage II analysis of
criteria pollutants.
Results by Pollutant
Tables 3-9 and 3-10 present the individual risk of cancer and the increased
aanual incidence of cancer, respectively, attributable to lifetime exposures
to the levels of metals monitored at the Fourth Street station.
Bear in mind that because of significant uncertainties in the underlying
data and assumptions, these estimates of individual risk and disease incidence
are only rough approximations of actual risk. They are based on conservative
estimates of exposure and potency, and are more likely to overestimate risks
than to underestimate them.
As shown in the tables, the highest individual lifetime risk and aggregate
annual incidence of cancer, according to our estimates of exposure and potency,
comes from chromium. Assuming that 10% of the chromium is hexavalent, the
individual lifetime risk of cancer is estimated to be 2 x 10~^, or twenty
in a million, and the aggregate annual incidence is estimated at 0.4 cases, or
one additional case every 2.5 years.
If none of the chromiun monitored in Santa Clara Valley's air is actually
hexavalent, cancer risk from chromium exposures should be zero. If, on the
other hand, it is all hexavalent, individual lifetime risks due to chromium
exposure may be as high 2 x 10~^, or 200 in a million. Similarly, if none
of the chromiun is hexavalent, we expect no increase in annual cancer incidence.
If, on the other hand, all of the chromium is hexavalent, then the conservative
estimate for expected increase in cancer rates could be up to 4 cases a year.
This range of uncertainty is much greater than it is for the other metals; the
magnitude of the health risk that chromium poses clearly depends on the assump-
tions one makes about the proportion of chromium that is hexavalent.
The level of risk from chromium (using the 10% hexavalent assumption) is
followed closely by the risk from arsenic. The range of individual lifetime
risk of cancer from air exposure to arsenic is 1 x 10~5 to 2 x 10~5, or between
ten and twenty in a million. The range of aggregate increased incidence of
cancer is 0.2 to 0.4 cases annually, or roughly one case every 2.5 to 5 years.
The average of this range is 0.3 cases annually, or one case every 3.3 years.
Of all the metals that we studied and that are thought to be carcinogenic,
berylliun contributes least to the risk of cancer. The range of individual
lifetime risk of cancer from beryllium is 7 x 10~8 to 3 x 10~7, or 0.07 to 0.3
-------
Table 3-9
Implications of Metals Monitored at the Fourth Street Station
for Individual Lifetime Risk of Cancer in Santa Clara Valley
Substance
Arsenic
Bariun
Berylliun
Cadmium
Chromium (2)
Lead
Nickel (2)
Zinc
Estimated Range of
Individual Risk of
Cancer from Lifetime Exposures
1 x 1(T5 to
(1)
7 x 10~8 to
2 x 10~6 to
0 to
(4)
0 to
(1)
2 x icr5
3 x 10~7
5 x 10~6
2 x 10~4
3 x 1(T6
Average Individual Risk of Cancer
From Lifetime Exposure
1.5 x 1CT5 or 15 in a million
(1)
2.0 x 10~7 or 0.2 in a million
3.5 x 10~6 or 3.5 in a million
2.0 x 10~5 or 20 in a million (3)
(4)
1.5 x 10~6 or 1.5 in a million
(1)
Level of
Evidence *
A
B2
Bl
A
A
u>
U)
Total
1 x 10~5 to 2 x 10~4
4.0 x 10 5 or 40 in a million (2)
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF
INDIVIDUAL RISK AND DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON
CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN TO
UNDERESTIMATE THEM. SEE TEXT.
1 At this time there appears to be insufficient data for an adequate evaluation of the potential carcinogenicity
of this compound. Therefore, we do not analyze it as a carcinogen in this analysis.
2 The lower end of this range is zero because it is possible that none of the monitored metal in Santa Clara
County's air is a carcinogenic species. See text.
3 We assume that 10% of the chromium is hexavalent in deriving these estimates. See text.
4 While there is limited evidence that some lead compounds induce tumors in experimental animals, CAG has not
yet made a determination as to the potential carcinogenicity of lead. Therefore, we do not analyze lead
as a carcinogen in this analysis.
* See Table 3-3
-------
Table 3-10
Implications of Metals Monitored at the Fourth Street Station
for Increased Annual Incidence of Cancer in Santa Clara Valley
Substance
Arsenic
Barium (2)
Beryllium
Cadmium
Chromium (3)
Lead (4)
Nickel (3)
Zinc (2)
Range of Aggregate
Annual Increase in
Cancer Incidence
0.2 to 0.4
0
0.001 to 0.005
0.04 to 0.1
0 to 4.0
0
0 to 0.06
0
Aggregate Annual Aggregate
Increase in Increase in
Cancer Incidence Cancer Incidence Level of
(Average of Range) (1) (Average of Range) Evidence *
0.3
0
0.003
0.07
0.4
0
0.03
0
one case every 3.3 years A
one case every 333 years B2
one case every 14.3 years Bl
one case every 2.5 years A
one case every 33 years A
Ul
U)
Total
0.2
to 4.6
0.8
roughly one case every 1.3 years
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF
INDIVIDUAL RISK AND DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON
CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN TO
UNDERESTIMATE THEM. SEE TEXT.
1 The one exception is that we assume that 10% of the chromium is hexavalent in deriving these estimates. See text.
2 At this time, there appears to be insufficient data for an adequate evaluation of the potential carcinogenicity
of this compound. Therefore, we do not analyze it as a carcinogen in this analysis.
3 The lower end of this range is zero because it is possible that none of the metal monitored in Santa Clara
County's air is a carcinogenic species. See text.
4 Vfriile there is limited evidence that seme lead compounds induce tumors in experimental animals, CAG has not yet
made a determination as to the potential carcinogenicity of lead. Therefore, we do not analyze it as a carcinogen
in this analysis.
* See Table 3-3
-------
3-36
in a million (70 to 300 in a billion). The aggregate incidence is 0.003 cases
annually, or one case every 333 years.
The total individual lifetime risk of cancer to the average individual
from outdoor air exposure to all of the metals at the levels monitored at
Fourth Street ranges from 1 x 10~5 to 2 x 10~4; our point estimate for individual
lifetime risk (incorporating the assumption of 10% hexavalent chromium) is
roughly 4 x 10~5 or forty in a million. Our estimates of aggregate annual
increased incidence from exposure to all the metals at the monitored levels
range from 0.2 to 4.6 cases. The point estimate (incorporating the 10%
hexavalent chromium assumption) is 0.8 or roughly one case each 1.3 years.
Table 3-11 compares the monitored ambient levels (both annual average
concentrations and highest single observed concentration) with the lowest
presumed human threshold for non-cancer health effects for each metal. (See
table 3-4 for a full listing of thresholds and non-cancer health effects.) For
all metals for which we have quantified thresholds (which excludes lead), the
maximum level of these metals detected in monitoring of ambient air was far
below the levels which we estimate would be required to produce any increased
risk of non-cancer health effects. Average concentrations were, of course,
lower than the maximum values. Unless our estimates of the thresholds for
non-cancer effects for these chemicals are very significantly in error, exposure
to ambient levels of these metals is unlikely to increase the risk of the
non-cancer health effects we examined. Exposures to lead could pose non-cancer
health risks in Santa Clara Valley, but for reasons discussed above, we were
unable to estimate such potential health impacts in this report. We hope to
examine the potential effects of exposure to lead in Stage II as part of our
analysis of criteria pollutants.
Results by Source Category
In addition to analyzing risks by pollutant, it is desirable to examine
risk by source so that risk management activities can be directed at the sources
of those pollutants posing health risks. However, our estimates of ambient
levels of metals, and the risk they pose, are based on monitoring information.
The monitoring information provides measures of the ambient levels of pollutants,
but does not provide information on the sources of the measured pollutants.
As a result, although we can estimate risk by pollutant, we do not have enough
information to make definitive statements regarding the sources of the estimated
risk.
One way in which to estimate risk by source is to compile an emissions
inventory and relate the emissions to monitored ambient levels, for instance
through modeling techniques. As mentioned, however, there is currently no
emission inventory for metals for Santa Clara County.
Given the lack of information relating ambient levels, and thus risks, to
sources, the IEMP decided to use the limited emissions information that is
available to help identify potential sources of metals that may warrant follow-up
research. Using the emissions estimates provided in table 3-7, we apportioned
the estimated cancer risk from metals to the source categories identified.
Bear in mind that by apportioning the risk among these source categories, we
are implicitly assuming that they are the only sources contributing to ambient
-------
3-37
Table 3-11
Comparison of Levels of Metals Monitored
at the Fourth Street Station in Santa Clara Valley
with Lowest Thresholds for Non-Cancer Health Effects
Except as noted, values are annual averages
in micrograms per cubic meter
Substance
Monitored
Concentrations (1)
Highest
Monitored Lowest Presuned
Concentration (2) Hunan Threshold (3)
Arsenic
Barium
Beryllium
Cadmium
Chromium:
Chromium +3
Chromium +6
Lead
Nickel
Zinc
0.0027 to 0.0041
0.0015 to 0.0358
0.00003 to 0.00014
0.001 to 0.003
0.0126 to 0.0138
0.46 to 1.62
0.0058 to 0.0090
0.0066 to 0.1075
1. See note one, table 3-5.
2. Highest single value recorded in
3. See table 3-4 for all thresholds
4. See text for a discussion of why
0.170
0.0665
0.00063
0.012
0.0363
6.56
0.043
13.3
175.0
1.9
0.24
3500.0
17.0
(4)
3.5
0.263 1540.0
ambient data collected by AQMD, 1976-83.
and non-cancer health effects.
health effects from lead are not included
in this report.
-------
3-38
levels of metals in the Valley; as stated earlier, although we think we have
identified the more significant sources, we did not consider all potential
sources. As a result, the estimates may overstate the contribution of these
sources for the metals examined.
In addition, the emissions estimates themselves are, in general,
approximationsthe estimates for residential heating and motor vehicles are
based on rough calculations and were not confirmed by testing of sources.
A key assumption we make in estimating risks from metals by source
category is that the contribution of a source to ambient levels is proportional
to the emissions of metals from that source. That is, we have not modeled the
estimated emissions to determine their impact on ambient levels. Instead, we
have assumed, for example, that if one source emits twice as much arsenic as
the other sources combined, it is responsible for twice as much risk as the
others.
Estimating risk in this way does not mean we assume that the emissions
from all sources, including the cement plant, necessarily affect ambient values
measured at Fourth Street; rather, it assumes that the Valley-wide risk estimates,
based on Fourth Street data, reflect the contribution of these sources to
ambient levels and exposures. That is, the Fourth Street data represent an
average that accounts for local variation in terms of levels and sources.
However, because the cement plant emissions are likely to have only a
localized impact on ambient levels, and the extent of exposure to these levels
is uncertain, we conducted sensitivity analysis. As an alternative to the
assumption that risks from the cement plant are proportionate to its emissions,
we assumed that the plant's emissions have no measurable impact on ambient
levels to which people are actually exposed and thus, the plant poses no health
risk.
Clearly, there is a great deal of uncertainty in making estimates based
on the assumptions discussed above. However, for purposes of the screening
analysis, we have decided to use the limited information that is currently
available on sources of metals in this way, if only to identify possible areas
for further research. The information presented here should be considered
preliminary and uncertain; more work would need to be done to characterize the
sources of metals in the County, such as the compilation of a metals emission
inventory as planned by the AQMD, before more certain statements can be made
about the sources of the estimated cancer risk.
In the remainder of this section, we present estimates of the average
individual lifetime risk of cancer and the aggregate increased incidence of
cancer attributable to the different sources of metals considered. The total
estimated cancer risks discussed in this section are the same as in the previous
section. Essentially, we are presenting the same information in a different
way: instead of describing risks as a function of individual pollutants, we
describe these risks as a function of pollutant source categories. Each source
category examined emits more than a single pollutant.
Vfe did not analyze non-cancer health effects by source category since
estimated ambient levels of the metals for which we have toxicological
information (which excludes lead) are below the assumed human threshold for
non-cancer effects.
-------
3-39
Table 3-12 presents the estimated cancer risk from airborne metals by
source category. Under the assumption that risk is proportionate to emissions
from all three source categories, we estimate that exposure to metals emitted
frcm residential heating may pose an individual lifetime risk of cancer of
1.5 x 10~5 (or 15 in a million) and an aggregate annual incidence of 0.28, or one
additional case of cancer every 3.5 years. If we assume that the cement plant
emissions do not result in exposures and thus risks, residential heating may
pose an individual lifetime risk of cancer of 4 x 10~5 (or 49 in a million)
and an aggregate annual incidence of 0.73, or one additional case of cancer
every 1.4 years. Thus, the combustion of fuel for residential heating may
account for between 35% and 91% of the estimated cancer risk from exposure to
airborne metals.
Assuming that risks are proportionate to its estimated emissions, the
cement plant may pose an average individual lifetime risk of 2.7 x 10~^ (or
27 in a million), and an aggregate annual incidence of 0.49, or one additional
cancer case every 2 years. Thus, this source may account for up to 61% of the
total estimated cancer risk from airborne metals. Under the alternative assump-
tion that the plant emissions do not result in measurable exposures, the cement
plant is estimated to pose no measurable risk.
Motor vehicles accounted for less of the estimated cancer risk from metals
than the other two categories.
-------
Table 3-12
Implications of Metals Monitored at the Fourth Street Station
for Risk of Cancer (by Source) in Santa Clara Valley
Source Category
Residential Heating
Distillated Oil
Fireplaces
Woods toves
Heating Total
Ctattent Plant
Motor Vehicles
Light Duty
Heavy Duty
Vehicles Total
TOTAL (1)
Risk Apportioned by
from Three Soui
Average Individual
Risk of Cancer from
Lifetime Exposure
<2.0 x 1(T8
8.3 x 1CT6
7.2 x 1CT6
1.5 x 1(T5
2 "? v i n~5
5.5 x 10-7
1.1 x 1(T6
1.7 x 10-6
4.4 x 1CT5
Estimated Emissions
rce Categories
Aggregate Annual
Increase in Cancer
Incidence (% of total)
<0.0004
0.15
0.13
0.28 (35%)
OAQ t £.~\&\
0.01
0.02
0.03 ( 4%)
0.8 (100%)
Risk Apportioned by
Emissions Excluding
Average Individual
Risk of Cancer from
Lifetime Exposure
5.5 x 10~7
2.3 x 10~5
1.7 x 10~5
4.1 x KT5
1.1 x 10-6
2.8 x 10-6
3.9 x 10-6
4.4 x 10~5
Estimated
Cement Plant
Aggregate
Increase in
Incidence (%
0.01
0.42
0.3
0.73
0.02
0.05
0.07
0.8 (
fuinual
Cancer
of total)
(91%)
( 9%)
100%)
Ul
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF INDIVIDUAL
RISK AND DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF
EXPOSURE AND POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
NOT ALL POSSIBLE SOURCES OF METALS WERE CONSIDERED IN APPORTIONING RISK; THEREFORE, THESE ESTIMATES MAY OVER-
STATE THE CONTRIBUTION OF THESE SOURCES FOR THE METALS EXAMINED.
THESE RISK ESTIMATES ARE BASED SOLELY ON ESTIMATED EMISSIONS OF METALS. THESE EMISSIONS ESTIMATES WERE NOT MODELED
TO DETERMINE THEIR IMPACT ON AMBIENT LEVELS. THE ESTIMATED CANCER RISK FROM EXPOSURE TO AMBIENT LEVELS OF METALS WAS
APPORTIONED TO EACH SOURCE CATEGORY IN PROPORTION TO THE ESTIMATED EMISSIONS OF EACH. FOR RESIDENTIAL HEATING AND
MOTOR VEHICLES, EMISSIONS ESTIMATES WERE BASED ON ROUGH CALCULATIONS NOT CONFIRMED BY SOURCE TESTING.
(1) Totals may differ from those presented in table 3-9 due to rounding.
-------
3-41
Methodology and Findings; Organic Gases
Our analysis of organic gases includes consideration of eleven specific
compounds and gasoline vapors, which is a mixture of compounds. The eleven
compounds include benzene, carbon tetrachloride, chloroform, ethylene dibromide,
ethylene oxide, methylene chloride (dichloromethane), perchloroethylene (tetra-
chloroethene), toluene, 1,1,1-trichloroethane (TCA, or methyl chloroform),
trichloroethylene (trichloroethene, or TCE), and xylene. In addition, we
examined a number of substances from particular point sources for their impact
on most exposed individuals. These substances include CFC-113 (often referred
to as Freon 113), dichlorobenzene, dichloroethane (ethylene dichloride,) dichloro-
ethylene (vinylidene chloride), glycol ethers (2-ethoxyethanol, 2^nethoxyethanol,
and 2-butoxyethanol; often referred to as Cellosolves), isopropyl alcohol, and
phenol. (All of these substances are listed on table 3-1.)
The discussion of organic gases is in four parts. The first discusses
the estimates of emissions of organic gases from various source categories.
The second section describes the dispersion modeling of the emissions estimates
and presents the resultant exposure estimates. The third part discusses the
implications of the modeled exposure levels in terms of health risks. We
present the risk results in two ways: first by pollutant and then by source.
The final section on organic gases describes the results of the short-term
ambient monitoring program.
Emissions of Organic Gases From Different Source Categories
For most of the organic gases,* the AQMD provided estimates of the amount
released into the air in Santa Clara Valley from a variety of sources including
large point sources, such as major manufacturing facilities, and smaller, more
dispersed sources such as dry cleaners, gasoline stations, and motor vehicles,
which we refer to as "area sources." ** EPA supplemented the AQMD data with
estimates of emissions from three sources: publicly owned treatment works
(POTWs or sewage treatment plants), groundwater aeration facilities, and
municipal landfills.
For the most part, the AQMD estimates are not based on actual testing of
the emissions of a particular source. The reason is that, since 1970, EPA has
set national ambient air quality standards for a relatively small number of
* The AQMD did not provide emissions estimates for EDB. See the section on
"Modeled Exposures to Organic Gases" foe a discussion of the methodology used
to estimate exposures to EDB. Emissions estimates for gasoline vapors came
from a variety of sources, including the AQMD, and are discussed below.
** For purposes of this study, we will consistently distinguish between point
sources and area sources as defined above in the text. Be aware, however, that
the AQMD, the Air Resources Board (ARB) and other regulatory agencies often
categorize sources in a slightly different manner. For instance, the AQMD divides
sources into stationary sources (generally referring to permitted point sources),
mobile sources (including on and off road vehicles and such sources as airplanes),
and area sources (smaller and more dispersed). As such, our area source category
includes the AQMD mobile source category and some of the smaller, more dispersed
permitted point sources.
-------
3-42
pollutants, called criteria pollutants; air pollution agencies at all levels of
government have concentrated their attention on those pollutants. The criteria
pollutants include ozone, sulfur oxides, nitrogen oxides, carbon monoxide,
total suspended particulates, and lead, but not individual organic substances.
The AQMD and other state and local agencies regulate emissions of precursor
or reactive hydrocarbons, because hydrocarbons as a class react with oxides of
nitrogen to produce ozone, one of the criteria pollutants. Because different
organics vary in their propensity to form ozone, the AQMD has compiled some
information about the makeup of the hydrocarbon emissions from different kinds
of sources in the Bay Area including Santa Clara County. This information,
however, was developed to estimate impacts on ozone levels and not for the
specific purpose of estimating exposures to the organic gases themselves.
Because of increasing concern that some organic gases may themselves
present health risks, both EPA and AQMD have recently paid more attention to
the emissions of specific organics. The AQMD has compiled information
indicating that implementation of its rules to reduce ozone precursors has
reduced emissions of a number of organic gases. Other studies also suggest
that regulations designed to control criteria pollutants may indirectly control
toxic contaminants. (Office of Policy Analysis, "The Air Toxics Problem in the
United States," USEPA, May 1985.)
The AQMD's estimates of emissions of each of the organic gases studied in
the IEMP were derived by multiplying the estimates of total hydrocarbons
estimates which are thought to be fairly reliableby "speciation factors" that
represent the proportion of each organic gas to total hydrocarbon emissions.
This speciation factor obviously varies from one source category to another and
from one pollutant to another.
For example, the AQMD has estimated that benzene accounts for 4% of the
hydrocarbons in automobile exhaust, and that toluene accounts for 11.3% of
those emissions. For exhaust from diesel trucks, the AQMD estimates that 2.2% of
hydrocarbon emissions consist of benzene, and 2.1% are toluene. Since AQMD has
estimated the total hydrocarbon emissions from automobile and diesel truck exhaust
in Santa Clara County, it is a straightforward matter to estimate emissions of
benzene, toluene, and other organic gases from those source categories.
For the point source emissions estimates, we have data from a Toxics Use
Survey conducted by the AQMD in 1984.
The AQMD has apportioned its estimates of hydrocarbon emissions from each
source category geographically within the County. For large point sources,
this exercise is easy, since we know where those sources are located. The
distribution of area source emissions across the County is somewhat more diffi-
cult. For mobile sources, the AQMD used a transportation model which estimates
the number of vehicle miles traveled in each part of the County, uses these
figures to estimate total hydrocarbon loadings in each area, and then scales
back the hydrocarbon loadings according to the emissions factors to obtain
estimates of specific compounds. The emissions from the remaining area source
categories were distributed using some surrogates, such as population.
As noted above, the emissions estimates for gasoline vapors did not come
exclusively from the AQMD. Wfe estimated emissions of gasoline vapors (evapora-
-------
3-43
tive emissions from unbumt fuel) in Santa Clara Valley from three general
source categories: (on-highway) mobile sources, off-highway mobile sources, and
the fuels distribution process.
For the first two categories, we relied on a 1983 California Air Resources
Board (ARB) emissions inventory of sources of total organics. (Most of the
data for Santa Clara Valley in the ARB inventory, with the exception of on-road
mobile sources, originates with the AQMD. The data that the ARB generates on
mobile sources is, in turn, supplied to the AQMD for inclusion in its inventory.)
This inventory, which reflects what is believed to be current control status,
provided an estimate of evaporative emissions of gasoline vapors from mobile
sources. The inventory also provided information on total organics emissions
from off-road vehicles. Although the inventory did not distinguish between
evaporative and exhaust emissions for this category, we were able to apportion
the emissions using EPA emission factors, information obtained through interviews,
and engineering judgement.
Finally, for the fuels distribution category, which includes both point
and area sources, our estimates were based on benzene emissions data provided
by the ARB and the AQMD. For instance, the AQMD has estimated that benzene
emissions from service station spills account for 1.8% of gasoline vapor
emissions from spillage. Since we have the estimates for benzene from spillage,
it is a straightforward matter to scale the estimate to reflect gasoline vapor
emissions. (For a more detailed discussion of estimating gasoline vapors
emissions, see Versar Inc., "Emissions and Exposures to Gasoline Vapors in
Santa Clara County, California," May 1986.)
Once we generated the emissions estimates, we were able to analyze
gasoline vapors in the same way in which we analyzed the other organic gases
whose emissions estimates came directly from the AQMD inventory. For instance,
the estimated emissions of gasoline vapors were apportioned geographically
within the County in the same manner as the AQMD emissions estimates. Thus,
unless otherwise noted, discussions of the use of AQMD emissions estimates,
such as the one that follows, apply to gasoline vapors as well.
Emissions Estimates From the AQMD Inventory
The AQMD estimated emissions from 25 specific point sources and about a
hundred different area source categories. Table 3-13 lists the 25 point sources.
In comparison to the point sources, which include major manufacturing facilities,
the individual area sources are smaller and more dispersed. Table 3-14 presents
the 12 categories into which the area sources fall and describes those categories.
We developed these categories by consolidating the 100 emission source categories
analyzed by the AQMD. In some cases, the area source categories correspond to
a single industry, such as drycleaning, or a single type of emitter, such as
motor vehicles (mobile sources). However, in other cases, the area source
category cuts across industries and firms. For instance, degreasing operations
are found in many different industries including, but not limited to, metal
finishing firms and the semiconductor industry. In addition, emissions from
degreasing operations from within one of the 25 point sources will be accounted
for in the point source total, but not in the degreasing category since point
source emissions and area source emissions are mutually exclusive.
-------
Table 3-13
25 Point Sources for Air Emissions
of Organic Gases in Santa Clara Valley
Facility
Location
Advanced Micro Devices
Air Station, Moffett Field
AMI Gould
Barnes-Hind Pharmaceuticals
Burke Industries
Chevron USA
CTS Printex
Del Monte
Dysan
Fairchild Camera & Instrument
FMC (San Jose Ordnance)
Good Samaritan Hospital
Great Western Chemical
Hewlett-Packard
Hewlett-Packard, Manufacturing
Hewlett-Packard, Santa Clara
IBM
Lockheed Missiles & Space
Memorex
Monolithic Memories
National Semiconductor
Raytheon
Signetics
Southern Pacific Pipelines
Xidex
Sunnyvale
Sunnyvale
Santa Clara
Sunnyvale
San Jose
San Jose
Mountain View
San Jose
Santa Clara
Mountain View
San Jose
San Jose
Milpitas
Palo Alto
Palo Alto
Santa Clara
San Jose
Sunnyvale
Santa Clara
Santa Clara
Santa Clara
Mountain View
Mountain View
San Jose
Sunnyvale
SOURCE: AQMD Inventories
-------
3-45
Table 3-14
List and Description of Area Sources Categories
for Air Emissions of Organic Gases in Santa Clara Valley
Area Source Category (1)
Description
Burning of Waste Materials
Combustion of Fuels
Includes burning of waste materials by
commercial and industrial sources
Includes burning of oil and gas for
domestic, industrial, and commercial
purposes
Degreasers
Drycleaners
Includes solvents used by industry to remove
grease from metal surfaces
Fuels Distribution
Industrial Solvents Coatings
Mobile Sources
Off-Highway Mobile Sources
Other Chem/Industry
Other Organic Compounds
Evaporation
Pesticides Usage
Includes air emissions from fuel storage
tanks, loading trucks, delivery, and filling
stations
Includes evaporation of organic compounds
from structures coating, industrial coating,
surface coating clean up solvents, and
coatings manufacture
Includes emissions from on-road cars, trucks,
and motorcycles
Includes construction equipment, locomotives,
and off-road vehicles
Includes air emissions from sources such as
oil and gas field operations, and unspecified
chemical manufacturing and agricultural
processing
Includes evaporation of organic compounds from
sources such as storage tanks, asphalt paving,
fiberglass products manufacturing, and
rubber/plastic products manufacturing
Photoresist
A process in the semiconductor industry
(1) Based on categories in the AQMD inventories.
-------
3-46
Table 3-15 presents total estimated emissions of organic gases from the
AQMD inventories divided between point and area sources. Largest total emis-
sions are for gasoline vapors, xylene, toluene, and perchloroethylene. For
seme compounds, such as carbon tetrachloride, chloroform and ethylene oxide,
point sources account for all emissions in the inventory. For methylene chlor-
ide and 1,1,1-trichloroethane, both point and area sources contribute to
total emissions. For still others, most notably gasoline vapors, benzene,
perchloroethylene, toluene, and xylene, the overwhelming bulk of emissions
comes from area sources.
Table 3-16 presents estimates of emissions from the 12 area source
categories. According to the AQMD inventories, motor vehicles (mobile sources)
are the largest source of gasoline vapors, benzene, and toluene; motor vehicles
also contribute, to a lesser degree, to xylene emissions. The largest area
source emitting xylene, however, is photoresist operations, a process in the
manufacture of semiconductor chips. Drycleaners emit the bulk of perchloro-
ethylene. Degreasers also contribute substantially to perchloroethylene
emissions as well as being the sole area source in the inventory for 1,1,1-
trichloroethane, trichloroethylene, and methylene chloride. The category of
industrial solvent coatings is not the primary source of any individual substance
but, according to the AQMD inventory, it does contribute to the total area
source emissions of benzene, perchloroethylene, toluene, and xylene.
These emissions estimates do not take into account recently enacted AQMD
regulations affecting the emissions of organic gases from semiconductor manu-
facturing facilities (Regulation 8, Rule 30). The AQMD projects that the
regulation may reduce emissions of chlorinated solvents from semiconductor
facilities by 20% to 40%. In addition, by encouraging positive photoresist
operations over negative photoresist in facilities emitting at least 15 pounds
per day of precursor organic compounds, it may reduce xylene emissions by 90%.
Negative photoresist typically uses xylene formulated resin and developer
solutions whereas positive photoresist typically uses glycol ethers for primer
and resin carrier.
We have not attempted a detailed analysis of the regulation for a number
of reasons. First, it will not be fully implemented until January 1987. In
addition, it is for total reactive organic gases rather than the individual
species we are studying, making it more difficult to assess what the reduction
will be for any given compound, except perhaps for xylene from non-exempt
facilities. Because the focus of the regulation is on precursor organic
compounds (those organics that contribute to ozone formation), it does not
cover all the organic compounds analyzed in this report. For instance, it
does not include methylene chloride or 1,1,1-trichloroethane. Finally, the
objective of this air analysis is to estimate exposures and risk using current
emissions as a baseline in order to identify and compare potential toxic problems.
As is generally true, issues identified in this screening as being of concern
may warrant further analysis, at which time the assessment could be refined to
address any nutiber of expected changesin industrial practices, in regulations,
in demographicsif appropriate.
Similarly, these emissions estimates do not take into account possible
changes in benzene emissions. Benzene is present in gasoline as a component of
crude oil and as a product of the octane enhancing process. Due to the phase-out
of lead in gasoline, the California Air Resources Board (ARB) estimates that
-------
3-47
Table 3-15
Estimated Air Bnissions of Organic Gases
in Santa Clara Valley
Estimates are in metric tons (kkg)/year
Total Estimated
Compound Emission
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Di bromide
Ethylene Oxide
Gasoline Vapors
Methylene Chloride
Perchloroethylene
1 , 1 , 1-Trichloroe thane
Trichloroethylene
Toluene
Xylene
667
0.09
1
it
7
11193 **
63
1004
361
41
1484
1720
Total Estimated
Emissions from
25 Point Sources
(% of total)
22 (3%)
0.09 (100%)
1 (100%)
7 (100%)
265 ( 2%)
35 ( 56%)
28 ( 3%)
134 ( 37%)
7 ( 17%)
30 ( 2%)
70 ( 4%)
Total Estimated
Emissions from
Area Sources
(% of total)
645 (97%)
0
0
0
10929 (98%)
28 (44%)
976 (97%)
227 (63%)
34 (83%)
1454 (98%)
1650 (96%)
* The AQMD did not provide emissions estimates for EDB. See section on
"Modeled Exposures, to Organic Gases" for a discussion of the methodology
used to estimate exposures to EDB.
** Estimated emissions of gasoline vapors came from a variety of sources.
See text.
SOURCE: AQMD Emissions Inventories
-------
Table 3-16
Estimated Air Emissions of Organic Gases in Santa Clara Valley
by Area Source Category
Estimates are in
Area Source Category(l)
Burning Vfeste Materials
Combustion of Fuels
Degreasers
Drycleaners
Fuels Distribution
Indust. Solvents Coatings
Mobile Sources
Off-Highway Mobile Sources
Other Chem/Industry
Other Organics Evap.
Pesticides Usage
Photoresist
Total Area Sources
Total 25 Point Sources
Benzene
1.6
56.5
0
0
21.0
125.5
374.3
8.5
6.4
15.0
36.5
0
645
22
Gasoline
Vapors(2)
0
0
0
0
935.0
0
9867.0
126.7
0
0
0
0
10929
265
Methylene
Chloride
0
0
28.0
0
0
0
0
0
0
0
0
0
28
35
metric tons/year
Per-
chloro-
ethylene
0
0
281.0
573.0
0
103.0
0
0
0
19.0
0
0
976
28
Toluene
0
157.7
0
0
84.0
315.7
826.6
16.1
3.0
36.1
14.9
0
1454
30
1,1,1-
Trichloro-
ethane
0
0
227.0
0
0
0
0
0
0
0
0
0
227
134
Tri-
chloro-
ethylene
0
0
34.0
0
0
0
0
0
0
0
0
0
34
7
Xylene
0
46.5
0
0
116.3
233.5 w
288.6 °°
4.3
0
28.1
44.5
888.5
1650
70
(1) See table 3-14 for a definition of these area source categories.
(2) Estimated emissions of gasoline vapors came from a variety of sources. See text.
SOURCE: AQMD Emissions Inventories
-------
3-49
the average benzene content of gasoline may increase by 31% between 1984 and
1990. The ARB projects that this may increase benzene emissions from the
evaporation and combustion of gasoline by 10% in state by the year 2000.
(California Air Resources Board, Staff Report, "Proposed Benzene Control Plan,"
May 1986.) However, in January 1985, the ARE identified benzene as a Toxic Air
Contaminant (TAG). State law requires that once a substance is identified as a
TAG, the ARE and the Districts prepare a report on the need for and appropriate
degree of control for the substance. In its report, the ARB staff outlines
specific control measures for benzene. If the plan is approved and fully
implemented, it may result in a 50% reduction in benzene emissions statewide by
the year 2000.
Emissions Estimates Frcm Three Additional Sources
EPA supplemented the AQMD emissions data with independent analyses of
three source categories which were not included in the AQMD's inventories:
groundwater aeration facilities, publicly owned treatment works (POTWs), and
municipal landfills. The estimates of emissions fron those sources were
prepared by the Association of Bay Area Governments (ABAG). The results are
subject to considerable uncertainty; we felt, however, that it was important to
obtain some estimate of emissions from those sources to ensure that potentially
significant emissions were not overlooked.
Groundwater aeration facilities are a potential source of air emissions
as a result of their pumping and discharge of contaminated water. Several
industrial facilities in Santa Clara County currently pump groundwater that
is contaminated with low levels of organics and release this water to storm
sewers, from which it eventually reaches the San Francisco Bay. In some cases,
this water is treated to remove most of the organics before it is discharged;
in other cases, where contamination levels are low, it is not treated. It is
reasonable to assume that most of the organics, unless they are removed by
some form of treatment besides simple aeration, end up in the air, since the
contaminants in question are all highly volatile (i.e., they tend to evaporate
quickly).
ABAG estimated the emissions to air of several contaminants from the
aeration of contaminated groundwater. (Association of Bay Area Governments
[ABAG] , "Air Emissions Associated with Pumping of Contaminated Groundwater in
Santa Clara Valley," February 1985.) The total estimated air emissions of
1,1,1-trichloroethane fron all six facilities combined were 0.7 metric tons
per year. This compares to 361 metric tons per year of total estimated emissions
of TCA from the sources in the AQMD's inventory. ABAG estimated emissions of
l,l,2-trichloro-l,2,2-trifluoroethane (CFC-113, often referred to as Freon 113)
from the six facilities to be slightly higher than one metric ton per year; no
comparable figures are available on CFC-113 emissions from other source
categories. Emissions for each of the other substances ABAG examined (including
trichloroethylene, perchloroethylene, methylene chloride, toluene, xylene, and
dichloroethylene) were estimated to be 0.2 metric tons per year or less.
Detailed figures are presented in table 3-17.
-------
3-50
Table 3-17
Estimated Air Emissions of Organic Gases
from Groundwater Aeration in Santa Clara Valley
Estimates are in metric tons/year;
they are conservative estimates designed for screening purposes
Substance Total Emissions (1) (2)
Dichloroethane 0.096
Dichloroethylene (3) 0.084
Dichlorobenzene 0.003
CFC-113 1.033
Isopropyl Alcohol 0.121
Methylene Chloride 0.018
Perchloroethylene 0.003
Toluene 0.001
1,1,1-Trichloroethane 0.698
Trichloroethylene 0.117
Xylene 0.200
1 Estimates are for combined emissions from groundwater aeration at Fairchild,
Hewlett Packard, IBM, and Signetics.
2 Estimates assume 100% volatilization of organics present in discharge water.
3 Estimates are for 1,1-Dichloroethylene and 1,2-Dichloroethylene combined.
SOURCE: Association of Bay Area Governments (ABAC), "Air Emissions Associated
with Pumping of Contaminated Groundwater in Santa Clara Valley,"
February 1985.
-------
3-51
Publicly owned treatment works (POIVte), or sewage treatment plants, can
be an important source of organic emissions to air because some industrial
facilities release significant quantities of those substances to the POTWs in
their wastewater. Because the substances are volatile, they can enter the air
during treatment of the wastewater. ABAC estimated the resultant emissions to
air from each of Santa Clara County's three major POTWs: Palo Alto, San Jose/Santa
Clara, and Sunnyvale. (Association of Bay Area Governments [ABAC], "Air Emissions
Associated with Publicly Owned Treatment Works in Santa Clara Valley," May, 1985.)
The San Jose/Santa Clara plant, operated by the City of San Jose, is by far
the largest of the three. Its service area includes a population of approximately
1.1 million and encompasses the cities of San Jose, Santa Clara, Campbell, Monte
Sereno, Saratoga, Milpitas, DDS Gatos, and five sanitary districts. In their
analysis, ABAC calculated that 120 million gallons of wastewater per day are
treated at the San Jose/Santa Clara plant,* 30 million gallons per day at the
Palo Alto plant, and 20 million gallons per day at the Sunnyvale plant. All
three discharge the treated water to the southern part of the San Francisco
Bay or to streams that flow directly into the Bay.
All three plants treat not only household sewage but also industrial
wastewater, including toxic metals and organic chemicals. In many other urban
areas, industrial facilities treat their own wastewaters and discharge them
directly to rivers or other bodies of water, but this is not the case in Santa
Clara County. Here, all industrial facilities discharge their wastewaters to
one of the municipal water pollution control plants for treatment. (Most large
industrial plants treat their wastes to some extent before discharging them to
the sewer system; these operations are called "pretreatment." Chapter 5 provides
a more complete discussion of municipal and industrial wastewater discharges
and their effect on water quality.)
Although pretreatment removes much of the contamination in the wastewater,
the wastewater still contains organic substances and other pollutants, most of
which are removed from the water at the water pollution control plants. Some
are broken down into simpler, innocuous substances during treatment; others
became part of the sewage treatment plants' sludges, which are generally
disposed of on land. A small fraction of the contaminants in the wastewaters
pass through the plants and enter the Bay; these loadings, and their possible
impacts, are discussed in chapter 5. The remainder of the pollutants, including
most of the volatile organic compounds discharged to the plant, probably enter
the air before or during treatment at the POTWs. (Memorandum to IEMP from
Versar, Inc., "Technical Background and Estimation Methods for Assessing Air
Releases from Sewage Treatment Plants," October 11, 1984.)
ABAC estimated those air emissions in a way designed to yield conservative
estimates. It is difficult to measure directly the emissions from POTWs, because
those emissions enter the air over a large area within the plants; they are not
concentrated and sent up through a stack or vent as are many industrial emissions.
ABAC'S estimates are therefore based on monitoring of the wastewater which enters
and leaves the plant rather than on actual monitoring of the air at or near the
* In 1984, the San Jose/Santa Clara POTW actually treated 114 MGD. Bear in
mind that the estimates used by ABAC are intended to represent typical
throughput, not the exact amount processed in any given year.
-------
3-5?
plant. We caution that the data were sparse and ABAC'S analysis involves a number
of important assumptions, some of which are discussed below. For a more complete
discussion of the methodology, see the ABAC and Versar documents referenced
above.
Our strategy was to measure the concentrations of contaminants in both the
influent (that is, the incoming water) and the effluent and to assume that the
balance of the contaminants are volatilized. This monitoring effort is complicated
by the fact that concentrations of contaminants in both the influent and the
effluent fluctuate widely over relatively short periods of time, depending on
who is discharging what substances to the sewer system. When conducting monitoring
of this sort, it is necessary to time it so that one is examining the same water
in the influent as in the effluent. That is, one tries to estimate the time
that the water stays in the plant, to sample the influent, and then to wait to
sample the effluent until the water's "residence time" has elapsed. Estimates
of that time, however, are bound to be inexact. It can happen that the effluent
concentration is actually higher than the influent concentration; this is not
because the POTW adds pollutants to the water but is simply a product of
fluctuations in the waste stream.
Because of those fluctuations, it is desirable to have considerable volumes
of data on the influent and effluent concentrations if one is going to try to
estimate air emissions in this way. Our data for the plants in Santa Clara
County were not as extensive as we would have liked; because the data are
sparse, we cannot be confident that they are representative of typical levels.
The relatively small number of data points does not invalidate them for the
kind of screening exercise we are conducting; it does, however, increase the
uncertainty attached to the results and the range of reasonable estimates one
can draw from the data.
In addition to these problems associated with the monitoring strategy,
there is also some uncertainty about the fate of the organics in the plant.
However, we believe that our assumption that the most volatile synthetic
organics enter the air rather than being broken down into other compounds is
reasonable, given the literature on this subject. (D.F. Bishop, "The Role of
Municipal Wastewater Treatment in Control of Toxics," presented at the NATO/CCMS
meeting, Bari, Italy, September, 1982; E.D. Pellizzari, "Volatile Organics in
Aeration Gases at Municipal Treatment Plants," Contract No. 68-03-2780. USEPA,
Cincinnati, OH, 1981; and A.C. Petrasek, et. al., "Removal and Partitioning of
Vblatile Organic Priority Pollutants in Wastewater Treatment," presented at the
Ninth U.S.-Japan Conference on Sewage Treatment Technology, Tokyo, Japan,
September, 1983.)
Using the method described above, ABAC estimated a range of emissions
from each of the three large POTWs in the northern part of the county. The
results are presented in table 3-18. Estimates of emissions from wastewater
treatment are higher than those for aeration of contaminated groundwater because
the volumes of water treated are much greater and because the concentrations of
contaminants in the wastewater tend to be higher.
Comparison of the results of the analysis of air emissions from POTWs
with the AQMD inventory (see table 3-15) indicates that the contribution of the
plants to total loadings of most substances is rather small. In the case of
methylene chloride, our upper-bound estimate of emissions from the Palo Alto
-------
Table 3-18
Estimated Air Emissions of Organic Gases
fron POTWs in Santa Clara Valley
Except where noted, estimates are in tons/year; ( 1)
they are conservative estimates designed for screening purposes.
Substance
Benzene
Bromod i chlorome thane
Brccnoform
Chloroform
Dibromochlorome thane
1 , 2-Dichloroethylene
Methylene chloride
Perchloroethylene
Toluene
1,1, 1-Trichloroe thane
Tr ichloroethylene
San Jose
Santa Clara
(120 mgd)
0.4 to 0.8
0.7 to 1.1
0 to 0.8
3.6
0 to 0.8
0.7 to 1.1
16.8 to 17.2
4.7 to 5.1
25.6
1.5 to 1.9
2.2 to 2.6
Sunnyvale
(20 mgd)
0 to 0.8
1.5 to 1.9
0 to 0.8
1.9
1.5 to 1.9
0 to 0.8
1.1 to 1.5
1.1 to 1.5
0.4 to 0,8
0.7 to 1.1
0.4 to 0.8
Palo Alto
(30 mgd)
0.4 to 0.8
0.4 to 0.8
0.4 to 0.8
1.5
0.4 to 0.8
0 to 0.8
25.7
0.4 to 0.8
4.4 to 4.8
1.5 to 1.9
1.1
Totals
(tons/yr)
0.8 to 2.4
2.6 to 3.8
0.4 to 2.4
7
1.9 to 3.5
0.7 to 2.7
43.6 to 44.4
6.2 to 7.4
30.4 to 31.2
3.7 to 4.9
3.7 to 4.5
Totals
(kkg/yr)(l)
0.7 to 2.2
2.4 to 3.5
0.4 to 2.2
6
1.7 to 3.2
0.6 to 2.5
39.5 to 40.3
5.6 to 6.7
27.6 to 28.3
3.4 to 4.4
3.4 to 4.1
Ul
1 Total emissions estimates have been converted to metric tons per year (kkg) to facilitate comparison with other
emissions estimates. One U.S. ton is roughly equivalent to 0.9 metric tons.
SOURCE: Association of Bay Area Governments (AB/^3), "Air Emissions Associated with Publicly Owned Treatment Works
in Sani.a Clara Valley," May
-------
3-54
plant is significant compared to other sources of the substance. Further
testing would certainly be needed to verify this emissions estimate before any
regulatory action would be warranted. Estimates of chloroform emissions from
PCTWs exceeded those from all other sources in the AQMD's inventories.
Municipal landfills are the third category of sources for which ABAG
estimated emissions. Santa Clara County has no facilities licensed to accept
hazardous wastes, but eleven landfills in the county are licensed to accept
garbage and other municipal wastes. Because these wastes may contain organic
compounds which can enter the air from the landfill, ABAG made some conservative
estimates of the possible loadings from municipal landfills. (Mae Clark and Gary
Silverroan, "Evaluation of Air Emissions, Runoff to Surface Water, and Leachate
to Groundwater from Sanitary Landfills," Association of Bay Area Governments
[ABAG], July, 1985.)
Although landfills are potentially an important source of hydrocarbon
emissions, very little data are available on the emissions for specific
organic compounds from landfills, and no data of this sort were available for
Santa Clara County. ABAG's estimates are based on studies conducted in other
areas and on assumptions which allow crude extrapolations of the data from
those other areas to Santa Clara County. These estimates are summarized in
table 3-19.
ABAG's estimates of emissions from landfills, using these admittedly rough
techniques, were low. For example, ABAG estimated that benzene emissions from
all Santa Clara County landfills totaled under 0.2 metric tons per year, compared
to the annual estimate of roughly 645 metric tons per year of benzene from area
sources in the AQMD inventory.
The exception to this general tendency is vinyl chloride, a human carcino-
gen. AQMD's inventories contained no information on emissions of vinyl chloride
from landfills. ABAG's conservative estimates of vinyl chloride emissions
from all sanitary landfills in the County came to almost 2 metric tons per
year. This estimate was based on extrapolation of data from landfills in the
southern part of California, some of which may, in the past, have received
wastes from the production of vinyl chloride or from other processes involving
similar chlorinated organic chemicals. Since no such wastes are known to be
present in Santa Clara County landfills, the estimates are more likely to be
too high than too low.
We planned to model the estimated emissions from groundwater aeration
facilities, POTWs and landfills on a county-wide basis, just as we do for the
emissions from the AQMD inventory to determine their impact on population
exposures. The contribution of these three sources to total emissions was
estimated to be so low, however, that we decided not to do so. We did
model exposures to those people living directly downwind from groundwater
aeration facilities and POTWs. The results of these analyses are presented
later in this chapter.
-------
Table 3-19
Estimated Air Bnissions of Organic Gases
from Sanitary Landfills in Santa Clara Valley
Except where noted, estimates are in tons per year; (1)
they are conservative estimates designed for screening purposes
Facility
Benzene
Perchloro-
ethylene
1,1,1-Tri-
chloroethane
Trichloro-
ethylene
Toluene
Vinyl
Chloride
City of Palo Alto
Refuse Disposal Area
City of Mountain View
Shoreline Regional Park
Stierlin Road Disposal Company
City of Sunnyvale Sanitary Landfill
Marshland Development Co.
Disposal Site
City of Santa Clara
All Purpose Landfill
Ov*sns-Corning Solid Waste
Disposal Site
Nine-Par Conpany Solid Waste
Disposal Facility
Newby Island Sanitary Landfill
San Jose Municipal Landfill
Guaoalupe Rubbish Conpany
0.01
0.04
0.01
0.01
0.03
0.02
0.05
0.01
0.04
0.13
0.04
0.04
0.10
0.08
0.15
0.02
0.01
0.02
0.01
0.01
0.02
0.01
0.02
<0.01
0.02
0.05
0.02
0.02
0.04
0.03
0.06
0.01
0.08
0.24
0.07
0.08
0.19
0.16
0.29
0.04
0.13
0.43
0.13
0.13
0.34
0.27
0.50
0.07
Total (tons/yr)
Total (kkg/yr) (1)
0.18
0.16
0.60
0.54
0.11
0.10
0.25
0.23
1.15
1.04
2.0
1.8
1 Totals have been converted to metric tons per year (kkg/yr) to facilitate comparison with other emissions
estimates. One U.S. ton is roughly equivalent to 0.9 metric tons.
SO'jRfR: Association of Bay Area Governments, "Evaluation ot Air Emissions, Runoff to Surface Water, and Leachate
to Groundwater from Sanitary i.andf ills,'1 July 1985.
-------
3-56
Modeled Exposures to Organic Gases
To calculate exposures to organic gases, the EPA and AQMD used a computer
model that generates estimates of ambient conditions in different parts of the
Valley based on the emissions estimates from the different sources discussed
above. The modeled estimates of ambient conditions were then combined with
population information to calculate exposures.
Figure 3-2 illustrates the way in which the computer model, which estimates
the dispersion of a given pollutant, provides estimates of ambient conditions
in 48 grid cells, with each cell representing a 5 kilometer square. This
grid system includes data provided by ABAC on population density within each
5 kilometer square. Thus, for each of the 48 grid cells, the number of people
exposed to any given level of pollutant can be estimated. A more detailed
description of the modeling, as well as estimated concentrations of the other
pollutants by grid cell, is available in Appendix 3-B.
This modeling approach is chiefly useful for estimating average or aggregate
exposures and risks over a relatively large area. We also, however, used the
model to estimate exposure to the maximum exposed individuals (MEIs); that is,
we estimated the concentrations to which people living near major point sources
of pollution (chiefly industrial facilities) might be exposed. (This modeling
is described in Versar, Inc., "Follow-up Air Quality Analysis in Support of the
Integrated Environmental Management Division's Santa Clara Projectf" June 12,
1985.)
Reliable, long-term ambient monitoring data for the organic gases were not
available to check the results of our model. The IEMP and AQMD did conduct
some short-term monitoring of the organic gases in the Fall of 1984. However,
because the sampling period was so short, we could not, in general, use the
data to estimate long-term ambient levels with any confidence, nor could we
use the monitoring data formally to validate the model. (The results of the
short-term monitoring program are discussed at the end of the section on organic
gases.) As such, the modeling results are subject to more uncertainty than
they would be if they had been based on both modeling and monitoring data.
One must be aware of the uncertainties and limitations of results based only
on modeling of this kind.
The most important limitation is that the model results are clearly
dependent on the emissions inventories provided by the AQMD. These inventories
were not developed for the purpose of modeling specific organic pollutants,
and may be incomplete. In addition, EPA has found in the Philadelphia IEMD
study that the type of dispersion model we used tends to underestimate ambient
levels, typically by a factor of two to three, for some pollutants. Under-
estimates are particularly likely for very persistent compounds like carbon
tetrachloride and benzene. Our short-term monitoring found benzene levels of
up to 6 ug/m3. in addition, more recent monitoring conducted by the ARE
between March and December of 1985 found average levels at the AQMD's San Jose
station to be roughly 10 ug/m3 compared to our modeled estimate of 2.6 ug/m3
in that general area. Neither of these monitored estimates are annual averages
as our estimates are; however, they do suggest that the model may be under-
estimating ambient benzene levels.
One possible explanation for this underestimate is that the background
-------
FIGURE 3-2 MODELED CONCENTRATIONS OF PERCHLOROETHYLENE
(in ug/m3)
'07
4145
4140
SANTA CLARA COUNTY
590 595 600 605 610
(Each grid square is 5km x 5km
total area shown is
30 km x 40 km)
4115
570 575 580
Ul
I
Ul
-J
-------
3-58
levels of the persistent pollutants such as benzene may be higher than antici-
pated. Since the background levels are not known they are not included in our
modeling of emissions. Thus, the higher the background level, the more a
model such as this will underestimate ambient levels. Another possible
explanation is that the model spreads the pollutant plume too much too quickly,
thereby lowering the estimated concentration.
A related limitation is that the model begins with estimates of emissions
and tracks pollutant dispersion assuming no chemical reactions. Because of
this simplifying assumption, the model cannot appropriately be used to estimate
exposures to highly reactive compoundsthose formed or rapidly degraded by
reactions in the atmosphere. For example, according to the AQMD's inventories,
1361 metric tons per year of formaldehyde are emitted in Santa Clara Valley,
almost entirely from motor vehicles. Such emissions are large relative to the
emissions of the other organic gases we studied. However, there are complex
mechanisms by which formaldehyde forms other compounds and by which other
compounds form formaldehyde in the atmosphere. For this reason, we were unable
to model exposures and resultant risks attributable to formaldehyde, which is
considered a carcinogen.
Table 3-20 presents the modeled range of annual average ambient levels
of organic gases in Santa Clara Valley. These estimates are in micrograms
per cubic meter (ug/m3). Table 3-21 provides factors for converting these
estimates into parts per billion (ppb). For almost all of the modeled
concentrations, levels near the higher end of the range were found in the
northern, industrialized, heavily traveled, and densely populated part of the
Valley; levels near the lower end of the range were in the mountains to the
south and west of the Valley. This pattern is illustrated in figure 3-2.
As noted in table 3-20, the AQMD did not provide data on emissions of EDB,
so we could not use the model directly as we did for other organic gases. In
1983, EPA severely restricted the use of EDB as a pesticide and soil fumigant
because of concern over its potential human health effects. EDB's other main
use is as an additive to leaded gasoline. When vehicles bum leaded gasoline,
deposits form in the engine and exhaust system. To reduce combustion chamber
deposits, EDB is added to "scavenge" the lead. Because of this use, EDB is
likely to be present in small amounts in the ambient air where leaded gasoline
is used. To estimate exposures to EDB, we estimated the ratio of benzene to
EDB in the exhaust of light-duty vehicles using national figures on EDB emissions.
Vfe used this ratio to estimate a range of ambient EDB levels and exposures
from exhaust. The reasonable assumption here is that the ratio of EDB to
benzene in exhaust is about the same in Santa Clara County as it is nationally.
Using this method, we estimated that ambient concentrations of EDB range
from 0.17 to 1.9 nanograms per cubic meter (0.00017 to 0.0019 ug/m3), the lowest
estimated concentration of any of the organic gases studied. As the concentration
of lead in gasoline is reduced in accordance with the lead phase-down regulations,
there will be a proportional reduction in EDB use and an accompanying decrease
in emissions and ambient concentrations.
For carbon tetrachloride, the AQMD's inventory included data on only one
point source and none on area sources. Resultant modeled exposure levels were
extremely low, with a maximum modeled concentration of one nanogram per cubic
meter (0.001 ug/m3).
-------
3-59
Table 3-20
Modeled Anbient Levels of Organic Gases
in Santa Clara Valley
All values are annual averages in micrograms/cubic meter
Substance
Geographic Range of
Modeled Concentrations
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Dibromide
Ethylene Oxide
Gasoline Vapors
Methylene Chloride
Perchloroethylene
1,1,1-Tr ichloroethane
Trichloroethylene
Toluene
Xylene
0.2 to 2.6
0.2 to 1.2 (1)
0.001 to 0.014
0,00017 to 0.0019 (2)
0.002 to 0.04
3.4
0,02
0.4
0.1
0.02
0.5
0.6
to
to
to
to
to
to
to
43.0
0.4!
4.0
2.8
0.2
6.2
7.6
1 This range of concentration is based on the snort-term monitoring conducted
for the IEMP in the Fall of 1984 rather than on modeling. See text for a
discussion of the modeling and monitoring results.
2 We did not model emissions estimates for EDB directly. See text for a
discussion of the methodology used for estimating concentrations.
-------
3-60
Table 3-21
Factors for Converting Estimates of Organic Gas Concentrations
from micrograms per cubic meter (ug/m3) to parts per billion (ppb)*
One microgram per One part per
Substance
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Di bromide
Ethylene Oxide
Gasoline Vapors **
Methylene Chloride
Perchloroethylene
1,1, 1-Trichloroethane
Trichloroethylene
Toluene
Xylene
cubic meter equals:
0.31 ppb
0.16 ppb
0.20 ppb
0.11 ppb
0.56 ppb
0.22 ppb
0.29 ppb
0.15 ppb
0.18 ppb
0.19 ppb
0.27 ppb
0.23 ppb
billion eguals:
3.2 ug/m3
6.3 ug/m3
5.0 ug/m3
7.7 ug/m3
1.8 ug/m3
4.6 ug/m3
3.4 ug/m3
6.7 ug/m3
5.5 ug/m3
5.3 ug/m3
3.8 ug/m3
4.3 ug/m3
* ug/m3 = (ppb)(0.04087) (molecular weight)
ppb = (ug/m3)(24.47/molecular weight)
** Assumes average molecular weight of 110 for the mix of compounds
in gasoline vapors.
-------
3-61
Carbon tetcachloride is extremely persistent in ambient air and has
repeatedly been found in ambient monitoring studies in other cities at levels
of 0.5 to 1 ug/ro^- These levels are found in urban areas in which, as in Santa
Clara Valley, there are few or no known sources of carbon tetrachloride. In
addition, the short-term monitoring results for Santa Clara Valley, discussed
below, show carbon tetrachloride to be present at average levels ranging from
0.2 to 1.2 ug/m^. Furthermore, studies suggest that there may be a global
background level of carbon tetrachloride of roughly 0.7 ug/m3. (P.G. Simmonds,
et al.r "The Atmospheric Lifetime Experiment: Results for Carbon Tetrachloride
Based on Three Years Data," Journal of Geophysical Research, October 20, 1983.)
Although in general we do not regard the short-term monitoring data as
a sufficient basis for exposure estimates, we believe that for the reasons
discussed above, in the case of carbon tetrachloride they are a better
indicator than our dispersion model of actual levels present in Santa Clara
Valley's ambient air.
Table 3-20 reveals that, based on the dispersion modeling of estimated
emissions, gasoline vapors, xylene, toluene, and perchloroethylene have the
highest resulting concentrations. Chloroform and ethylene oxide were modeled
to have the lowest ambient concentrations. These results are consistent with
the emissions estimates presented in table 3-15.
According to the AQMD, chloroform is emitted to air in relatively small
quantities from several industrial facilities in the Valley. Modeling of those
estimates yields estimates of ambient exposures well under 0.1 ug/m^. In the
Los Angeles area, however, the California Air Resources Board (ARB) has observed
that model results based on inventoried sources of chloroform do not predict
monitoring results well. As a result, the ARB has embarked on a research
project to identify possible uninventoried sources of chloroform in the South
Coast Air Basin. It is quite possible that the findings in the South Coast Air
Basin will have implications for Santa Clara Valley. However, results of the
short-term monitoring conducted in the Valley for the IEMP are consistent with
the model results in finding that chloroform is present in the Valley's outdoor
air at extremely low concentrations. Thus, for purposes of the present study,
we rely on our modeling results.
Emissions data on ethylene oxide, the other compound for which we estimate
low ambient concentrations, are also limited. Ethylene oxide is used in medical
equipment manufacture and as a sterilizer for medical equipment in hospitals.
The AQMD inventory had data on emissions from only two point sourcesone
industrial facility and one hospital.
Table 3-22 compares the concentrations of selected organic gases in Santa
Clara Valley's outdoor air with annual average levels monitored in other cities.
One must exercise caution in comparing these levels since different methodologies
were used: in Santa Clara County, the results are based primarily on modeling
(except for the estimated concentrations of carbon tetrachloride which are
based on short-term monitoring), whereas the data from the other cities is
based on long-term monitoring.
In most cases, the annual average modeled levels of gases in Santa Clara
County are comparable to or slightly lower than the annual average monitored
levels in the other cities. For example, modeled levels of benzene in Santa
-------
Table 3-22
Comparison of Ambient Levels of Organic Gases
in Santa Clara Valley with Levels in Other Cities
All values are measured in micrograms/cubic meter
Except where noted, all values are annual averages based on ambient monitoring
Substance
Benzene
Carbon
Tetrachloride
Perchloro-
ethylene
1,1,1-TCA
Sar_a Clara
Valley
0.2 to 2.6 (1)(2
0.2 to 1.2 (3)
0.4 to 4.0 (1)
0.1 to 2.8 (1)
Elizabeth
New Jersey
) 22.3
0.2
2.9
0.4
Philadelphia
Naval Hospital
10.4
4.0
5.4
28.8
Eowntown
Los Angeles
15.2
0.2
9.0
6.8
Baltimore
(Guilford)
13.9
8.5
4.1
15.7
u>
CTl
K)
1 These ranges are based on dispersion modeling, not ambient monitoring; as such, one must
exercise caution in comparing these results with the monitored levels from the other cities
since the methodologies were different.
2 The dispersion model may be underestimating ambient benzene levels in Santa Clara Valley.
3 This range is based on short term monitoring; as such, one must exercise caution in comparing these
results with the annual average monitored levels from the other cities since the methodologies were
different.
Source: Bill Hunt, Bob Faoro, Tom Curran and Jena Muntz, "Estimated Cancer Incidence Rates for
Selected Toxic Air Pollutants Using Ambient Air Pollution Data," July 1984.
-------
3-63
Clara Valley are between 5 and 10 times lower than the levels monitored else-
where. (Bear in mind, however, that benzene levels based on the dispersion
model may be too low, perhaps by a factor of two or more.) Short-term monitored
levels of carbon tetrachloride are similar to or below levels found in long-term
monitoring elsewhere. Modeled levels of perchloroethylene and trichloroethane
were also similar to or below levels found in the long-term monitoring in other-
cities.
In addition to estimating average concentrations and exposures over the
entire study area, we used the computer model to estimate concentrations to
which people living near major sources of pollution might be exposed. This
modeling provides estimates of exposures to the most exposed individual (MEI).
Table 3-23 presents the MEI concentration estimates for each compound modeled
and the type of source that was modeled. Except as noted, the estimated MEI
concentration includes background concentrations from other sources as well as
emissions from the modeled source.
In addition to the eleven gases modeled for average exposures, we included
glycol ethers and phenol in our analysis of most exposed individuals. The
emissions data on these additional substances are extremely limited; the AQMD
provided emissions estimates for the largest source of each identified in its
inventory, and those estimates were then modeled to calculate exposure to the
most exposed individual. Because the inventory does not differentiate between
the various glycol ethers, the estimate presented is for 2-ethoxyethanol,
2-methoxyethanol, and 2-butoxyethanol combined (often referred to as Cellosolves)
Vfe had hoped to include arsine and phosphine in this analysis, but the
AQMD does not have an inventory of emissions for either. Although it has
collected usage information on these compounds recently, the data are not
compiled in such a way as to be usable for our modeling purposes. The limited
source testing performed by AQMD staff has indicated that routine emissions of
these compounds are below the detection limits. The detection limits, however,
are relatively high118 ug/m3 for arsine, and 208 ug/m3 for phosphine.
To estimate the MEI for gasoline vapors, we examined exposures to persons
from self-service pumping of gasoline. The ARB provided an estimate of benzene
concentrations at the pump which we scaled up to reflect gasoline vapors, much
as we did for emissions from other sources. The ARB estimate reflects the
Stage II vapor recovery system at the pump. The average ambient concentration
of gasoline vapors at the pump was estimated to be 49000 ug/m3 at any given
moment. We assumed that the highly exposed individual pumps 40 gallons of
gasoline per .veek and that it requires 1.25 minutes to pump ten gallons.
(These assumptions were provided in USEPA, "Evaluation of Air Pollution
Regulatory Strategies for the Gasoline Marketing Industry," July 1984.) Given
these assumptions, the most exposed individual is exposed for five minutes each
week to the estimated concentration. If this exposure were spread over a year,
the annual average would be 17 ug/m3.
Finally, in addition to modeling AQMD emissions estimates for MEIs, we
modeled ABAG's estimates of emissions of organic compounds from POTVte and
groundwater aeration facilities. These compounds include CFC-113, dichloro-
benzene, dichloroethane, dichloroethylene, and isopropyl alcohol. Bea,.- in
mind that these compounds were modeled only from these specific facilities,
which may not be the most significant sources. The MEI estimates for these
-------
3-64
Table 3-23
Modeled Ambient Levels of Organic Gases
for Most Exposed Individuals in Santa Clara Valley
All values are annual averages in micrograms/cubic meter
Substance
Source Modeled
Modeled MEI
Concentration (1)
Benzene
Carbon Tetrachloride
CFC-113
Chloroform
Dichlorobenzene
Dichloroethane
Dichloroethylene
Ethylene Dibromide
Ethylene Oxide
Gasoline Vapors
Glycol Ethers
Isopropyl Alcohol
Methylene Chloride
Perchloroethylene
Phenol
Toluene
Trichloroethane
Trichloroethylene
Xylene
Traffic Intersection
(2)
Groundwater Aeration
POTW
Groundwater Aeration
Groundwater Aeration
Groundwater Aeration
Traffic Intersection
Hospital
Service Station Pump
Computer Mfg. Facility
Groundwater Aeration
Industrial Facility
Drycleaning Facility
Photo Equip. & Supplies Mfg.
Pipeline
Industrial Facility
Semiconductor Facility
Pipeline
27.6
1.4 (3) (4)
0.1
0.004 (3) (4)
0.1 (3) (4)
0.1 (3) (4) (5)
0.03
2.0
17.0 (6)
55.4 (3) (7)
0.2 (3) (4)
7.6
30.0
21.0 (3)
11.8
44.2
14.0
16.0
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
3-65
FOOTOOTES TO TABLE 3-23
(1) Except as noted, nodeled concentration includes emissions from listed point
source as well as background concentrations from other sources.
(2) Analysis of carbon tetrachloride was based on nonitoring rather than modeling;
thus no MEI calculations were modeled.
(3) Estimate is based solely on modeled maximum emissions fron the source listed.
Total ambient concentration could be higher due to background levels from
other sources.
(4) Emissions of this compound were modeled only from groundwater aeration
facilities which may not be the most significant source. The estimate
presented is thus the highest exposure from the source modeled, but may
not represent the highest exposure in the Valley.
(5) Estimate is for 1,1-dichloroethylene and 1,2-dichloroethylene combined.
(6) At any given moment of pumping gasoline, an individual may be exposed to
49,000 ug/m-* of gasoline vapors. Assuming that the most exposed individual
pumps forty gallons of gasoline a week, he/she may be exposed for five
minutes each week. If this exposure were spread over a year, the annual
average would then be 17 ug/m^.
(7) Estimate is for 2-ethoxyethanol, 2-methoxyethanol, and 2-butoxyethanol
combined.
-------
3-66
substances are thus the highest exposures from the source modeled, but may
not represent the highest exposure to the compound in the Valley.
The MEI concentration estimates are typically five to one hundred times
higher than the average concentration levels. The difference between the
average and MEI exposure was greater for chemicals such as chloroform and
ethylene oxide whose emissions were dominated by a few point sources, and less
for chemicals such as toluene and xylene that are emitted by many dispersed
sources.
Estimated Health Risks from Exposure to Organic Gases
This section discusses the estimated average individual risks, the
increased incidence in disease, and the risks to the most exposed individual
that may result from exposures to organic gases in Santa Clara Valley. The
estimates are based on the modeled exposure levels presented in the previous
section. The results are presented first by pollutant and then by source
category.
Bear in mind that our analysis of toxic air pollutants includes several
species of chlorinated hydrocarbons. These substances include carbon tetra-
chloride, chloroform, dichloroethane, dichloroethylene, methylene chloride,
perchloroethylene, 1,1,1-trichloroethane, and trichloroethylene. For these
compounds, as for the others in the analysis, we present estimates of the risk
of cancer and other diseases. However, the evidence on carcinogenicity for
most of the chlorinated hydrocarbons is not as strong as it is for benzene or
for many of the metals. Considerable debate exists about whether some of
these chlorinated hydrocarbons are carcinogenic in humans. In this analysis,
we use the strength-of-evidence criteria presented in table 3-3.
EPA's current policy is that 1,1,1-trichloroethane (TCA) should not be
considered a human carcinogen, and we follow that policy as a base case. EPA
formerly considered it a possible carcinogen, but has suspended that classifi-
cation because a key implicating study is currently under review. Because of
the uncertainty on this issue, and pending the outcome of the review, we have
performed sensitivity analysis of the possible impact of TCA if it were a
carcinogen. The value of this analysis is that it can indicate the importance
of further research on this issue in terms of local risk assessment. These
estimates, which appear in footnotes to the text and tables, should be regarded
as an extreme upper-bound of possible potency. They are not part of our base-
case analysis.
There is also uncertainty over the carcinogenicity of gasoline vapors.
EPA's Carcinogen Assessment Group (CAG) has calculated an upper-bound cancer
potency value for gasoline vapors presented in table 3-3. Much of the evidence
of carcinogenicity comes fron a two-year animal study sponsored by the American
Petroleum Institute (API). Although the study was well conducted, its relevance
to human risk assessment is uncertain. One issue is that the animal models
used in.the study may be particularly susceptible to the observed effect,
thereby diminishing the relevance of the findings to humans. For purposes of
this analysis, we follow EPA policy and rely on the strength-of-evidence
criteria presented in table 3-3.
-------
3-67
Another uncertainty centers on the question of vapor content. In the API
study, the gasoline was wholly vaporized for animal exposure; that is, the
inhaled mixture was identical to that in the liquid phase. However, this
mixture is not representative of the evaporative mixture found in ambient
situations. Some of the larqer hydrocarbon molecules comprising the liquid
gasoline mixture are less volatile and will therefore be present in lower
proportions in ambient vapors. The importance of this distinction is that
certain subsets of the higher molecular weight compounds are the ones that
appear most likely to be responsible for toxic effects observed in the male
rats in the API study. The API experiment may therefore overstate the toxic
potential of gasoline vapors in the ambient environment. Studies suggest that
the wholly vaporized mixture may overestimate concentrations of these higher
molecular weight compounds by a factor of three to five relative to measured
concentrations in the ambient environment. (For example, see Health Effects
Institute, "Gasoline Vapor Exposure and Human Cancer: Evaluation of Existing
Scientific Information and Recommendations for Future Research," September
1985.) As a result, based on such findings and the professional judgement
of EPA's Office of Air Quality Planning and Standards, for our base-case
analysis we have divided our risk estimates by a factor of four. However,
when appropriate, we also indicate what the full range of risk might bewhich
ranges from zero to the undivided estimate based on the upper-bound CAG value.
Results by Pollutant
Tables 3-24 and 3-25 present estimates of average individual lifetime risk
and aggregate increased incidence of cancer, respectively, attributable to
exposure to the modeled levels of the organic gases studied. Bear in mind that
because of significant uncertainties in the underlying data and assumptions,
these estimates of individual risk and disease incidence are only rough approxi-
mations of actual risk. They are based on conservative estimates of exposure
and potency and are more likely to overestimate risks than underestimate them.
According to our estimates, of the organic gases we studied, benzene poses
the greatest risk of cancer from exposures to annual average concentrations.
Based on our modeling of benzene concentrations, the estimated individual life-
time risk of cancer from exposure to benzene is between 2 x 10~6 and 2 x 10~^
with an average individual lifetime risk of 2 x 10~5, or twenty in a million.
(The average takes into account the number of people exposed to different
modeled concentrationsi.e., it is population weightedand thus is not a
simple average of the range.) Expressed in terms of aggregate increased inci-
dence of disease, the conservative estimate for benzene is roughly 0.3 cases
annually, or one case every 3.3 years.
There is some evidence, discussed earlier, that the dispersion model we
used may be underestimating benzene levels by a factor of two to four. If
ambient benzene levels are four times greater than we modeled, then the average
individual risk could be as high as 8 x 10-5 (80 in a million) and the aggregate
incidence could be up to 1.2 cases per year. More work would be needed to
determine actual exposure levels.
In its recently released report (cited above), the ARE staff indicated
its belief that the risk from exposure to benzene in California is significant
and not protective of public health. If the plan they outline to control benzene
-------
Table 3-24
Implications of Modeled Ambient Levels of Organic Gases
for Individual Lifetime Risk of Cancer (by Pollutant) in Santa Clara Valley
Substance
Estimated Geographic
Range of Individual Risk
of Cancer From Lifetime Exposures (1)
Average Individual Risk of Level of
Cancer from Lifetime Exposures (2) Evidence*
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Dibronide (5)
Ethylene Oxide
Gasoline Vapors
Methylene Chloride
Perchloroethylene
1,1,1-Trichloroethane
Trichloroethylene
Toluene
Xylene
Total
2 x 1CT6 to 2 x 10-5 (3)
3 x 10~6 to 2 x 10-5 (4)
2 x 10~8 to 3 x 10~7
4 x 10~8 to 4 x 10-7
2 x 10-7 to 4 x 10-6
6 x ID"7 to 8 x 10~6 (6)
8 x 10~8 to 2 x 10~6
2 x 10~7 to 2 x 10~6
(7)
3 x 10~8 to 3 x 10-7
(8)
(8J
6 x 10~6 to 6 x 10-5
2 x 10-5
1 x 10~5
6 x 10~8
2 x 10-7
2 x 10~6
6 x 10~6
6 x ID'7
2 x 10~6
1 x 10-7
or 20 in a million A
or 10 in a million (4) B2
0.06 in a million B2
0.2 in a million B2
2 in a million Bl
or
or
or
or
or
or
(7)
or
(8)
in a million (6) B2
0.6 in a million
2 in a million
0.1 in a million
4.0 x ID"5 or 40 in a million
B2
B2
B2
Ul
Ol
00
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF INDIVIDUAL RISK
AND DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF
EXPOSURE AND POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
* See table 3-3
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
FOOTNOTES TO TABLE 3-24
1 Except where noted, the range of individual risk is based on the lowest and highest modeled concentrations
to which people are exposed in the County. The range therefore reflects geographic variations in ambient
concentrat ions.
2 Except where noted, the average takes into account the number of people exposed to the different modeled
concentrations in the County (i.e., it is a population-weighted average), and thus is not a simple average of
the range.
3 This geographic range does not reflect the full range of possible individual risk. The individual lifetime
risk of cancer may be as high as 8 x 10 ~^ if the model is underestimating ambient levels of benzene by a
factor of 4. See text.
4 The risk estimates are based on exposures estimated from the short-term monitoring rather than on modeling
results; therefore, average risk is calculated as a simple average of the range of individual risk.
5 Estimates are based on a cancer potency value from ingestion studies. Inhalation value may be lower.
6 This geographic range does not reflect the full range of possible individual risk. The full range varies from
0 to 3 x ICT^ depending upon assumptions regarding the carcinogenicity of gasoline vapors. See text.
7 If 1,1,1-Trichloroethane (TCA) is carcinogenic, then the individual risk ranges between 5 x 10"^ and
1 x 10~6, average individual risk is 5 x 10~7, or 500 in a billion. Because the risks from TCA are
relatively low/ the totals are unchanged under this alternative.
8 At this time, there appears to be insufficient data for an adeguate evaluation of the potential carcinogenicity
of this compound. Therefore, we do not analyze it as a carcinogen in this analysis.
-------
Table 3-25
Implications of Modeled Ambient Levels of Organic Gases
for Aggregate Annual Increased Incidence of Cancer (by Pollutant) in Santa Clara Valley
Substance
Benzene
Carbon Tetrachloride (2)
Chloroform
Ethylene Dibronide (3)
Ethylene Oxide
Gasoline Vapors
Methylene Chloride
Perchloroethylene
1,1, 1-Tr ichloroethane
Tr ichloroethylene
Ibluene
Xylene
Total
Aggregate Annual Increase
in Incidence of Cancer
0.3 (1)
0.2
0.001
0.004
0.03
0.1 (4)
0.01
0.03
(5)
0.002
(6)
(6)
0.7
Aggregate Increase in
Incidence of Cancer
one case every 3.3 years
one case every 5 years
one case every 1,000 years
one case every 250 years
one case every 33 years
one case every 10 years
one case every 100 years
one case every 33 years
(5)
one case every 500 years
(6)
(6)
one case every 1.4 years
Level of
Evidence *
A
B2
B2
B2
Bl
B2
B2
B2 i
c
B2
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF INDIVIDUAL RISK
AND DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF
EXPOSURE AND POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
1 Based on certain assunptions regarding benzene levels. Aggregate incidence may be as high as 1.2. See text.
2 Based on the exposures estimated from the short-term monitoring rather than modeling results; see text
3 Based on cancer potency value from ingestion studies. Inhalation value may be lower.
4 This estimate is based on certain assumptions regarding the carcinogenicity of gasoline vapors.
The full range of possible aggregate annual increase in incidence of cancer is 0 to 0.4. See text
5 If 1,1,1-Trichloroethane (TCA) is carcinogenic, then the increased annual incidence is 0.01 cases. Because the
risks from TCA are relatively low, the totals are unchanged under this alternative.
6 At this time, there appears to be insufficient data for an adequate evaluation of the potential carcinogenicity of
this compound.
See table 3-3
Therefore, we do not analyze it as a carcinogen in this analysis.
-------
3-71
is approved and fully implemented, the ARE estimates a reduction in statewide
cancer risk of 39% between 1984 and 2000. However, some of this estimated
reduction in risk would result from statewide implementation of Stage I and
Stage II vapor recovery. Such controls are already in place in Santa Clara
Valleythe AQMD has required them in the Bay Area since the mid-seventies.
The next highest risk we estimated was from carbon tetrachloride with an
individual lifetime risk of cancer ranging from 3 x 10~6 to 2 x 10~^, and an
average of 1 x 10~^, or ten in a million. (These risk estimates were based
on short-term monitoring and thus we do not have information on geographic
variability; as such, the average is calculated as a simple average of the
range.) For carbon tetrachloride, the aggregate increased incidence is roughly
0.2 cases annually, or an additional case every five years.
Gasoline vapors were estimated to pose an individual lifetime risk of
cancer ranging from 6 x 10~^ to 8 x 10~* with an average of 6 x 10~6, or
six in a million; the estimated aggregate increased incidence is 0.1, or one
additional case every 10 years. As noted earlier, there is uncertainty over
the carcinogenicity of gasoline vapors; the full range of possible individual
risk reflecting these uncertainties is 0 to 3 x 10~~5 (zero to thirty in a
million). The full range of possible increased annual incidence is 0 to 0.4
(zero cases to one case every 2.2 years). At the high end of the full range,
gasoline vapors pose the greatest risk of any of the organic gases studied.
Thus, the magnitude of the health risk that gasoline vapors may pose depends
on the assumptions made regarding the carcinogenicity of ambient vapors.
Of all the organic gases that we studied and that are thought to be
carcinogenic, chloroform contributes least to the risk of cancer, probably
because the modeled levels are so low. The range of individual lifetime risk
of cancer from chloroform is 2 x 10~8 to 3 x 10~7, or 0.02 to 0.3 in a million
(20 to 300 in a billion). The aggregate annual incidence is estimated to be
0.001, or one case every 1,000 years. Chloroform and its close chemical
relatives, called trihalomethanes, are of concern as drinking water contaminants,
and are discussed in that context in the next chapter. Our analysis of chloroform
in air indicates fairly strongly that chloroform is primarily a drinking water
problem.
The total individual lifetime risk of cancer from air exposure to all the
organic gases at the levels estimated, primarily through modeling, ranges from
6 x 10~6 to 6 x 10~5, with an average of 4 x 10~5, or 40 in a million. Our
estimate of aggregate annual increased incidence from exposure to all the
organic gases at the estimated concentrations is 0.7 cases, or roughly one
case every 1.4 years.*
* We conducted sensitivity analysis examining the impact of TCA if it were a
carcinogen. The results of this analysis indicate that the risk of cancer
from airborne exposures to modeled annual average ambient levels of TCA are
relatively low; as a result, the total estimated risk would be unchanged under
this alternative.
-------
3-72
Table 3-26 compares the modeled annual average ambient level of each
organic gas with the lowest presumed human threshold for non-cancer health
effects for that substance. (See table 3-4 for a full listing of thresholds
and non-cancer effects.) Except in the case of benzene, the maximum modeled
annual average level of the pollutants were below the levels which we conserva-
tively estimate would be required to produce any increased risk other than
cancer. *
In the case of benzene, we modeled annual average ambient concentrations
ranging from 0.2 to 2.6 ug/m3. The higher end of this range is at the threshold
at which one might expect increases in adverse blood effects. We estimate
that roughly 100,000 people may be at some risk of blood effects from benzene
exposures. "Blood effects" usually involve decreased cell counts of one or
more blood cell types as the most common sign of toxicity in humans and in
laboratory animals following sufficient inhalation or oral exposure to benzene.
Blood cell counts may be depressed in humans to an extent to cause anemia
(i.e., depressed red blood cell count) or to make an individual prone to
infection (i.e., leukopenia or depressed white blood cell count). The assumed
human threshold for blood effects from benzene is estimated to be 2.45 ug/m^.
Note that the threshold is computed in a manner designed to be highly conserva-
tive; it is not certain that exposures at or near the threshold actually will
result in increased risk. Further work may be appropriate to verify these
findings and better assess the magnitude of the potential risk from benzene,
particularly given that the model may be underestimating ambient levels by a
factor of two or more.
Table 3-27 presents the individual lifetime risk of cancer for the most
exposed individuals (MEIs) based on our modeling. As shown, one of the highest
risks of cancer we calculated for an MEI was from exposures to ethylene oxide
emitted by large hospitals. Using conservative assumptions about ethylene
oxide emissions from hospitals, we estimate that the individual increased life-
time risk of cancer to such maximum exposed individuals could be up to 2 x 10"^,
or 200 in a million. This estimate is based on a modeled exposure directly
downwind of a large hospital of 2 ug/m^. Ambient monitoring or further work
on estimating emission rates and the reactivity of ethylene oxide would be
necessary to verify this tentative finding. Unfortunately, monitoring techniques
for organic gases such as the one we used in conducting our short-term monitoring
do not identify or quantify ethylene oxide. However, the ARB, in conjunction
with the California Department of Health Services, is currently preparing a
report identifying levels of exposure to ethylene oxide statewide and evaluating
the health effects associated with exposures. If ethylene oxide is identified
by the ARB as a Toxic Air Contaminant, the report will serve as a basis for
future regulatory action. A draft of the report is expected to be available
for public comment in Summer 1986.
* TCA has not demonstrated any teratogenic potential in published studies
conducted using rodent species. Therefore, the lEMP's base-case analysis
assumes that exposure to TCA poses no risk of fetal effects. An unpublished
study, which has not undergone scientific peer review, reports fetotoxic effects
in rat pups exposed in utero to TCA (Dapson et al., 1984). In order to assess
the importance of further research on this issue, the IBMP uses the Dapson
study to examine the alternative assumption that exposures above a threshold of
16 ug/m3 could pose the risk of fetal effects. Annual average modeled concentra-
tions of TCA are below this estimated threshold. See the methodology section
in this chapter for a more complete discussion of the analysis.
-------
3-73
Table 3-26
Comparison of Modeled Anbient Levels of Organic Gases in Santa Clara Valley
with lowest Thresholds for Non-Cancer Health Effects
All values are annual averages in micrograms per cubic meter
Substance
Modeled Concentrations
Lowest Presumed
Human Threshold (1)
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Dibromide
Ethyl ene Oxide
Gasoline Vapors
Methylene Chloride
Perchloroethylene
1,1, 1-Tr ichloroe thane
Tr ichloroe thy lene
Toluene
Xylene
0.2
0.2
0.001
0.00017
0.002
3.4
0.02
0.4
0.1
0.02
0.5
0.6
to
to
to
to
to
to
to
to
to
to
to
to
2.6
1.2 (2)
0.014
0.00019
0.04
44.0
0.48
4.0
2.8
0.2
6.2
7.6
2.45 (blood effects)
2.4
2.43
1.75
(3)
(4)
210.0
69.9
97.9 (5)
26.0
476.0
52.8
1 See table 3-4 for all thresholds and non-cancer effects.
2 This range of concentration is based on the short-term monitoring rather than
modeling. See text for a discussion of the modeling and monitoring results.
3 Quantification of the threshold is not complete at this time.
4 Evaluation of potential non-cancer health effects is not complete at this time.
5 The IEMP conducted sensitivity analysis on TCA for possible fetal effects.
See footnote to text.
-------
3-74
table 3-27
Maximum Exposed Individuals (MEIs) to Organic Gases:
Implications of Modeled Concentrations
for Individual Lifetime Risk of Cancer in Santa Clara Valley
Substance
Source Modeled
for MEI
Concentrat ion
Individual Lifetime
Risk of Cancer
for MEIs (1)
Level of
Evidence
for Cancer*
Benzene
Carbon
Tetrachloride
CFC-113
Chloroform
Dichlorobenzene
Dichloroethane
Dichloroethylene
Ethylene Dibromide
Ethylene Oxide
Gasoline Vapors
Glycol Ethers
Isopropyl Alcohol
Methylene Chloride
Perchloroethylene
Phenol
1,1,1-Trichloroethane
Trichloroethylens
Toluene
Xylene
Traffic Intersection
(2)
Groundwater Aeration
POTW
Groundwater Aeration
Groundwater Aeration
Groundwater Aeration
Traffic Intersection
Hospital
Service Station Pump
Computer Mfg. Facility
Groundwater Aeration
Industrial Facility
Drycleaning Facility
Photo Equip. & Supplies Mfg.
Industrial Facility
Semiconductor Faci t. ity
Pipeline
Pipeline
2 x Iff"4
(3)
2 x 10-6
(3)
3 x 10-6 (4)
5 x 10-6 (4)(5)
6 x 10-6 (6)
2 x 10-4
3 x 10-6 (7)
(8)
(3)
3 x ID"5
1 x 10~5
(9)
(10)
2 x 10-5
(3)
(3)
B2
B2
B2
C
B2
Bl
B2
B2
B2
B2
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF RISK OF DISEASE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY, AND
ARE MORE LIKELY TO OVERESTIMATE RISKS THAT TO UNDERESTIMATE THEM. SEE TEXT.
* See table 3-3
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
3-75
FOOTNOTES TO TABLE 3-27
1 Risks are calculated using the modeled concentrations presented in table 3-23.
2 Analysis of carbon tetrachloride was based on monitoring rather than modeling
results; thus no MEI calculations were modeled.
3 At this time there appears to be insufficient data for an adequate evaluation
of the potential carcinogenicity of this compound; therefore it is not analyzed
as a carcinogen in this report.
4 Bnissions of this compound were modeled only from groundwater aeration
facilities which may not be the most significant source. The estimate presented
is thus the highest individual lifetime risk of cancer from this substance
based on the source modeled, but may not represent the maximum individual
lifetime risk of cancer from this substance in the Valley.
5 Exposure estimate is for 1,1-dichloroethylene and 1,2-dichloroethylene
combined; potency value is for 1,1-dichloroethylene.
6 This estimate is based on a cancer potency value derived from ingestion
studies. Inhalation value may be lower.
7 This estimate is based on certain assumptions regarding the carcinogenicity
of gasoline vapors as discussed in the text. The full range of possible
individual risk for the MEI is 0 to 1 x 10~5. There is additional uncertainty
in calculating risks to the most exposed individual since human exposure to
gasoline vapor when filling fuel tanks is typically brief and sporadic whereas
the exposure experiments to date have generally been sub-chronic or chronic.
8 No data regarding the potential carcinogenicity of these compounds were
located in the available literature.
9 No inhalation data regarding the potential carcinogenicity of this compound
were located in the available literature.
10 If 1,1,1-trichloroethane is carcinogenic, then the individual risk of cancer
to the MEI is 2 x 10~5.
-------
3-76
People living in the immediate vicinity of heavily traveled intersections
may likewise be at increased risk for cancer from exposure to benzene. We
estimate that the individual lifetime risk of cancer to such MEIs may be as
high as 2 x 10~4, or 200 in a million. This estimate is based on the modeled
benzene MEI concentration of 27.6 ug/m3. Other MEI estimates for cancer are
summarized in table 3-27.
Table 3-28 compares the modeled MEI concentrations to the lowest presumed
human threshold for non-cancer health effects for each substance. As shown,
the MEI concentration for benzene is modeled at 27.6 ug/m3 and the conservatively
estimated threshold for blood effects is 2.45 ug/m3. Thus, the maximum exposed
individual to benzene may be at increased risk of depressed red or white blood
counts (as described more fully above). In addition, the modeled level of
benzene at the intersection exceeds the presumed human threshold of 4.1 ug/m3
for fetal effects consisting of delayed fetal growth. Thus, the maximum exposed
individual to benzene may be at increased risk of this effect as well. Further
work on estimating actual exposure levels, such as with ambient monitoring of
benzene, would be needed to verify these findings. In most ambient monitoring
studies, the researchers try to put the monitors away from major traffic zones
because the high localized concentrations would interfere with their efforts to
get more representative values.
Although there is no RfD or IEMD derived threshold for the glycol ethers,
EPA's Environmental Criteria and Assessment Office has calculated AIC (Acceptable
Intake Chronic) values, which are similar in concept to RfDs. These values may
provide insight into possible health risks from exposure to glycol ethers
pending the calculation of an RfD or IEMD threshold. (See this chapter's
methodology section for a more complete discussion of this issue.) However,
the AICs are for the individual compounds whereas our exposure estimate is for
2-ethoxyethanol, 2-methoxyethanol, and 2-butoxyethanol combined (the AQMD
inventory does not differentiate between the different compounds), making a
health risk assessment more difficult.
As presented earlier, the thresholds based on AICs are 170 ug/m3 for
2-ethoxyethanol; 84 ug/m3 for 2-methoxyethanol; and 56 ug/m3 for 2-butoxyethanol.
We conservatively estimate that a person living directly downwind of the largest
point source identified in the AQMD inventory for emissions of glycol ethers
may be exposed to up to 55 ug/m3 for all three compounds combined. Therefore,
even if the entire modeled concentration were of 2-ethoxyethanol or 2-methoxy-
ethanol, the exposure to the most exposed individual would be below the AIC.
If the entire modeled concentration were 2-butoxyethanol, the exposure would
still be below the AIC, but would be very close to that estimated value.
Further work identifying emissions of the individual glycol ethers would be
useful in assessing exposures to persons in the immediate vicinity of large
point sources.
-------
3-77
Table 3-28
Maximum Exposed Individuals (MEIs) to Organic Gases:
Comparison of Modeled Concentrations in Santa Clara Valley
with Lowest Thresholds For Non-Cancer Health Effects
All values are annual averages in micrograms/cubic meter
Substance
Source Modeled
for MEI
Concentration
Modeled
Concentra t ion
for MEI
Lowest Presumed
Human Threshold(1)
Benzene
Carbon
Tetrachloride
CFC-113
Chloroform
Dichlorobenzene
Dichloroethane
Dichloroethylene
Ethylene Dibromide
Ethylene Oxide
Gasoline Vapors
Glycol Ethers
Isopropyl Alcohol
Methylene Chloride
Perchloroethylene
Phenol
Toluene
1/1,1-Trichloroethane
Trichloroethylene
Xylene
Traffic Intersection
(2)
Groundwater Aeration
POTW
Groundwater Aeration
Groundwater Aeration
Groundwater Aeration
Traffic Intersection
Hospital
Service Station Pump
Computer Mfg. Facil.
Groundwater Aeration
Industrial Facility
Drycleaning Facility
Photo Equip. Mfg.
Pipeline
Industrial Facility
Semiconductor Facility
Pipeline
27.6
2.45 (blood effects)
4.1 (fetal effects)
2.4
1.4 (3)(4) 105000.0
0.1
0.004 (3)(4)
0.1 (3)(4)
0.1 (3)(4)(5)
0.03
2.0
17.0 (7)
55.4 (3)(9)
0.2 (3H4)
7.6
30.0
21.0 (3)
11.8
44.2
14.0
16.0
2.4
315.0
26.0
2.5
1.75
(6)
(8)
(10)
(11)
210.0
69.9
350.0
476.0
97.9 (12)
26.0
52.8
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF RISK OF DISEASE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY, AND
ARE MORE LIKELY TO OVERESTIMATE RISKS THAT TO UNDERESTIMATE THEM. SEE TEXT.
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
3-78
FOOTNOTES TO TABLE 3-28
1 See table 3-4 for all thresholds and non-cancer health effects.
2 Analysis of carbon tetrachloride was based on monitoring rather than modeling
results and thus no MET modeling was done.
3 Estimate is based solely on modeled maximum emissions from the source listed
Total ambient concentration could be higher due to background levels from
other sources.
4 Emissions of this compound were modeled only from groundwater aeration
facilities which may not be the most significant source. The estimate presented
is thus the highest exposure from the source modeled, but may not represent
the highest exposure in the Valley.
5 Estimate is for 1,1-dichloroethylene and 1,2-dichloroethylene combined.
6 Table 3-4 lists non-cancer health effects that have been tentatively identified
from the literature. However, quantification of thresholds is not complete at
this time.
7 At any given moment of pumping gasoline, an individual may be exposed to
49,000 ug/m3 of gasoline vapors. Assuming that the most exposed individual
pumps forty gallons of gasoline a week, he/she may be exposed for five minutes
each week. If this exposure were spread over a year, the annual average would
then be 17 ug/m3.
8 Evaluation of potential non-cancer health effects is not complete at this time.
9 Estimate is for 2-ethoxyethanol, 2^methoxyethanol, and 2-butoxyethanol
combined.
10 There is no RfD or IEMD derived threshold for glycol ethers. See text for a
discussion of available information on non-cancer health effects.
11 Review of the literature regarding non-cancer health effects is not complete
at this time.
12 The IEMP conducted sensitivity analysis on TCA for possible fetal effects.
See footnote to text.
-------
3-79
For all other pollutants, the maximum modeled exposures were below the
levels which we conservatively estimate would be required to produce an increased
risk of non-cancer health effects.*
Results by Source Category
In this section, we present the estimated average individual lifetime
risk, the aggregate annual increased incidence of cancer, and the risk to the
most exposed individual attributable to different sources of organic gases in
Santa Clara County. The source categories considered were discussed above and
are listed in tables 3-13 and 3-14. We used the emissions estimates from the
various source categories presented in tables 3-15 and 3-16 to calculate annual
exposures and hence risks attributable to those sources.
The total estimated cancer risks discussed in this section are the same as
in the previous section. Essentially, we are presenting the same information
in a different way: instead of describing cancer risks as a function of individual
pollutants, we describe these risks as a function of pollutant sources. In
some cases, these sources emit more than one pollutant.
We did not analyze non-cancer health effects for annual average exposures
by source category since, except in the case of benzene, estimated ambient
concentrations of the organic gases (for which we have thresholds) are below
the assumed human thresholds for non-cancer effects. According to the ACMD's
inventories, motor vehicles and other area sources account for 97% of benzene
emissions in Santa Clara County (see tables 3-15 and 3-16). We do, however,
analyze non-cancer health effects by source for the most exposed individuals.
Table 3-29 presents the estimates of cancer risk from organic gases by
source category. Motor vehicles (mobile sources) are the greatest single
source of cancer risk from organic gases posing an average individual lifetime
risk of 1 x 10~5 (OL- ten in a million), and an aggregate annual incidence of
0.25. That is, we conservatively estimate one additional case of cancer every
* TCA has not demonstrated any teratogenic potential in published studies
conducted using rodent species. Therefore, the IEMP base-case analysis assumes
that exposure to TCA poses no risk of fetal effects. An unpublished study,
which has not undergone scientific peer review, reports fetotoxic effects
(cardiac malformations) in rat pups exposed ir± utero to TCA (Dapson et al.,
1984). In order to assess the importance to Santa Clara Valley residents of
further research on this issue, the IEMP uses the Dapson study to examine the
possible impact of TCA under the alternative assumption that exposures above an
estimated threshold of 16 ug/n\3 could pose the risk of fetal effects. THE
SENSITIVITY RESULTS SHOULD NOT BE INTERPRETED AS INDICATING WHETHER OR NOT A
RISK IN FACT EXISTS; EPA RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK
MANAGEMENT DECISION-MAKING OR REGULATORY ACTION. Under this alternative
assumption, the estimated exposure of 44 ug/m^ to the most exposed individual
(MEI) downwind of the modeled point source exceeds the estimated threshold.
This finding suggests that more research is appropriate, both on the actual
levels of exposure to MEIs and on TCA's potential adverse effects. The National
Toxicology Program has conmissioned a project to repeat the limited Dapson study;
results are expected in Fall 1986.
-------
3-80
Table 3-29
Implications of Modeled Ambient Levels of Organic Gases
For Risk of Cancer (by Source Category) in Santa Clara Valley
Average Individual Risk
of Cancer From
Aggregate Annual
Increase in Incidence
Source Category
Burning of Waste Material
Combustion of Fuels
Degreasers (1)
Drycleaners
Fuels Distribution
Indust. Solvents Coating
Mobile Sources
Off-Highway Mobile Sources
Other Chem/Indust.
Other Organ ics Evaporation
Pesticides Usage
Photoresist
AREA SOURCE TOTAL
25 Point Sources Total
Carbon Tetrachloride (2)
Lifetime Exposure
4 x 10~8
1 x ID"6
8 x 10-7
8 x 10-7
1 x 10-6
3 x 10-6
1 x 10-5
3 x 10-7
2 x 10-7
4 x 10-7
8 x 10-7
0
2 x 10-5
6 x ID"6
1 x 10~5
of Cancer (%
0.0007
0.02
0.015
0.015
0.02
0.05
0.25
0.005
0.003
0.007
0.015
0
0.4
0.1
0.2
of total)
(0.1%)
(3%)
(2%)
(2%)
(3%)
(7%)
(36%)
(0.7%)
(0.4%)
(1%)
(2%)
(0%)
(57%)
(14%)
(30%)
TOTAL
4 x 10-5
0.7 (100%)
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF INDIVIDUAL RISK AND DISEASE INCIDENCE ARE ONLY ROUGH
APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES
OF EXPOSURE AND POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN
UNDERESTIMATE THEM. SEE TEXT.
1 This estimate assumes that 1,1,1-Trichloroethane is not carcinogenic; however,
if it is a carcinogen, because the risk is relatively low, the total individual
risk and aggregate incidence is virtually unchanged.
2 Estimates for carbon tetrachloride are based on monitoring of background levels
and thus cannot be attributed either to point or area sources.
-------
3-81
four years attributable to motor vehicles. This source alone accounts for 36%
of the total estimated cancer risk from organic gases and 63% of the risk from
these compounds from area sources. (This estimate incorporates our base-case
assumptions discussed earlier regarding the carcinogenicity of gasoline vapors
and modeled levels of benzene.)
The twenty-five point sources combined are the next greatest source of
cancer risk from exposure to organic gases. We conservatively estimate an
individual lifetime risk of cancer from exposure to these sources of 6 x 10~°
(or six in a million) and an aggregate annual increased incidence of 0.1, or
one additional case of cancer every ten years. Together, the 25 point sources
account for roughly 14% of the total estimated cancer risk from organic gases
in Santa Clara Valley's air.
According to our estimates, one area source categoryphotoresist
operationsdoes not contribute at all to estimated cancer risks from exposure
to organic gases. The reason is that, according to the AQMD inventories, this
source category emits only one organic gas of those we studiedxylene; EPA
currently does not consider xylene to be carcinogenic.
Carbon tetrachloride poses more risk than any individual source category
except mobile sources, accounting for roughly 30% of the estimated cancer risk
fron organic gases. We estimate an individual lifetime risk of cancer from
exposure to carbon tetrachloride of 1 x 10~5 (ten in a million) and an aggregate
annual increased incidence of 0.2, or one additional case every five years.
Carbon tetrachloride cannot be attributed to either area or point sources
since our concentration estimates come from monitoring of background levels.
Overall, area sources account for roughly 57% of the total estimated
cancer risks from the organic pollutants we analyzed. Carbon tetrachloride
accounts for 30% of the estimated risk and the 25 point sources combined
account for about 14% of the risk.
Table 3-30 presents the individual lifetime risk of cancer to the most
exposed individual from different sources. In some cases, we mooeled a source
for emissions of only one carcinogenic substance. However, other sources
were modeled for emissions of more than one carcinogen, in which case the
risks were added to arrive at the total estimated risk to the most exposed
individual from the many substances emitted by the source. The risks were
estimated based on emissions from the modeled source only, and do not include
possible background levels of the substance from other sources. As such,
actual exposures and risks could be somewhat higher. In most cases, however,
background levels are minimal compared to the modeled maximum concentrations.
As shown, the highest individual lifetime risk of cancer we estimated is
attributable to a heavily congested traffic intersection. The lifetime cancer
risk to a person in the immediate vicinity of the intersection may be as high
as 3 x 10~^, or three hundred in a million. The substances modeled include
benzene (which poses most of the estimated risk at the intersection), ethylene
dibromide, cadmium, and benzo(a)pyrene. The latter two substances are not
organic gases, but were included in the estimate because they had been modeled
from the sane source.
Hospitals and pharmaceutical manufacturers emitting ethylene oxide were
-------
3-82
Table 3-30
Maximum Exposed Individuals (MEIs) to Organic Gases
Implications for Individual Lifetime Risk of Cancer
Modeled Source
Modeled
Substances/Level
of Evidence (1)
from Selected Sources:
in Santa Clara Valley
Individual Lifetime
Risk of Cancer for
MEI from Modeled Source
Computer Equipment
Manufacturer
Dry Cleaning Facility
Gas Station
Gas Station Pump
Groundwater Aeration
Facility
Heavily
Congested
Traffic
Intersection
Hospital
Industrial Facility
Pharmaceutical
Manufacturer
Pipeline
POTW
Semiconductor
Facility
Benzene/A
Methylene Chloride/82
(2)
Perchloroethylene/B2
Benzene/A
Benzene/A
Gasoline Vapors/82
Methylene Chloride/82
Trichloroethylene/B2
(4)
Benzene/A
Benzo(a)pyrene/B2 (5)
Cadmium/81 (6)
Ethylene Dibromide/B2
Ethylene Oxide/Bl
Benzene/A
Methylene Chloride/82
(7)
Ethylene Oxide/81
Benzene/A
Benzene/A
Chloroform/82
Methylene Chloride/82
Pe rchloroe thylene/B2
Trichloroethylene/B2
(8)
Benzene/A
Methylene Chloride/82
(9)
4 x icr5
1 x 10-5
8 x 10-8
4 x 10-6 (3)
2 x 10-7
3 x 10-4
2 x 10-4
3 x 10-5
1 x 10-4
2 x 10-5
5 x 10-6
2 x 10-5
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMP-
TIONS, THESE ESTIMATES OF INDIVIDUAL RISK ARE ONLY ROUGH APPROXIMATIONS OF
ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY,
AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
3-83
FOOTNOTES TO TABLE 3-30
(1) Only those substances that were both modeled for MEI estimates and analyzed
as carcinogens in this report are listed. Risks were estimated based on modeled
emissions of the substances from the source only and do not include possible
background levels of the substance from other sources. See table 3-31 for a
full listing of modeled substances, concentrations, and sources. See table 3-3
for an explanation of "level of evidence" for cancer.
(2) If 1,1,1-Trichloroethane is carcinogenic, then the source total would be
5 x 10-5.
(3) This estimate is based on certain assumptions regard ing the carcino-
genicity of gasoline vapors as discussed in the text. The full range of possible
individual risk for the MEI from gasoline vapors alone is 0 to 1 x 10~5. There
is additional uncertainty in calculating risks to the most exposed individual
since human exposure to gasoline vapor when filling fuel tanks is typically
brief and sporadic whereas the exposure experiments to date have generally been
sub-chronic or chronic.
(4) If 1,1,1-Trichloroethane is carcinogenic, the source total remains unchanged
because the added risk is relatively low.
(5) BaP is an organic particulate, not an organic gas; it is included in this
table since it was part of the modeling of the traffic intersection.
(6) Cadmium is a metal, not an organic gas; it was included in this table
since it was part of the modeling of the traffic intersection.
(7) If 1,1,1-Trichloroethane is carcinogenic, then the source total would be
5 x 10-5.
(8) If 1,1,1-Trichloroethane is carcinogenic, the source total remains unchanged
under this alternative because the added risk is relatively low.
(9) If 1,1,1-Trichloroethane is carcinogenic, then the source total would be
2 x 10~5-
-------
3-84
the next largest source of individual cancer risk to the most exposed individual
with estimates of 2 x 1CT4 (two hundred in a million) and 1 x 10~4 (one hundred
in a million) respectively.
Table 3-31 compares the MET concentration by source with the lowest presumed
human threshold for non-cancer health effects for the various compounds modeled
from different sources. The modeled exposure levels may be somewhat different
from the estimates presented in table 3-28 because the estimates in this table
do not include estimated background levels as noted above. Because non-cancer
effects vary for each individual pollutant and cannot be added, and because no
one source exceeds thresholds for more than one pollutant, most of the information
presented in this table has been discussed in the section on risks by pollutant.
One exception is that benzene emissions trom a computer manufacturing facility
are slightly above the non-cancer threshold for blood effects and are at the
threshold for fetal effects. (Because benzene emissions at the intersection
were so much greater than from this facility, only the exposure level at the
intersection was presented earlier.) Again, further work identifying the actual
levels of exposure is necessary to confirm this finding.
Results of the Short-Term Monitoring of Organic Gases
In the Fall of 1984, the EPA and AQMD collected samples of ambient air in
Santa Clara County over the course of a week and analyzed the samples for most
of the organic gases discussed in the preceding section. Because the sampling
period was so short, we could not, in general, use the data to estimate long-term
ambient levels with any confidence, nor could we use the monitoring data to
formally validate the models. The results can, however, serve as a trough check
on the model results.
Table 3-32 presents the results of the short-term monitoring for the
substances we studied. As noted on the table, the monitoring techniques we
used are unreliable for methylene chloride and TCA, and thus we have no monitoring
data for these substances. In addition, we do not have monitoring results for
EDB or ethylene oxide.
Benzene is one of the organic gases for which the model results are not
entirely consistent with the few measurements of ambient air which are available.
The monitored values are higher than the levels the model predicts. Of course,
the monitored levels were average levels over the course of a week, whereas the
modeled levels are annual average values so that we do not expect them to be
identical.
In the Philadelphia and Baltimore lEMPs, EPA used similar models to estimate
benzene levels and found that models of the sort we used in this analysis tend
to underestimate benzene levels by about a factor of two or three compared to
the ambient monitoring data from those cities. The reasons for this consis-
tent underestimation are unclear; it is possible that emissions inventories
for benzene in all three areas are incomplete or that the model spreads the
pollutant plume too much too quickly, thereby lowering the estimated concen-
tration. Based on the information from Philadelphia and Baltimore, we think
it is likely that ambient benzene levels in Santa Clara County may be higher
than the levels the model predicts, perhaps by a factor of two.
-------
Table 3-31
Maximum Exposed Individuals (MEIs)
Comparison
Source Modeled
Computer Equipment
Manufacturer
Computer Mfg. Facility
Dry Cleaning Facility
Gas Station
Gas Station Pump
Groundwater Aeration
Fac ill ty
Heavily
Congested
Traffic
Intersection
to Organic Gases from Selected
of Modeled Concentrations in Santa Clara Valley with Lowest
All values are annual aver
Substance
Modeled
Benzene
Methylene Chloride
1 , 1 ,1-Trichloroethane
Xylene
Glycol Ethers
Perchloroethylene
Benzene
Toulene
Xylene
Benzene
Gasoline Vapors
CFC-113
Methylene Chloride
Toluene
1,1, 1-Tr ichloroethane
Trichloroethylene
Xylene
Benzene
Benzo(a)pyrene (9)
Cadmium ( U)
Ethylene Dibrcmid'?
ages in micrograms/cubic meter
Modeled
Concentration
for MEI (1)
4.1
2.3
11.7
7.6
55.0 (4)
25.9
0.01
0.02
0.003
0.17 (6)
17.0 (7)
0.055
0.01
0.002
0.008
0.1
0.1
25.6
0.003
0.04
0.027
Sources
Thresholds
Lowest Presumed
Human Threshold (2)
2.45 (blood effects)
4.1 ( fetal effects)
210.0
97.9 (3)
52.8
(5)
LO
69.9 oo
LTl
2.45
476.0
52.8
2.45
(8)
105000.0
210.0
476.0
97.9 (3)
69.9
52.8
2.45 (blood effects)
4.1 (fetal effects)
(10)
0.24
1.75
-------
Table 3-31 cont.
Source Modeled
Hospital
Industrial Facility
Substance
Modeled
Ethylene Oxide
Benzene
Methylene Chloride
Toluene
1,1, l~Tr ichloroe thane
Modeled
Concentration
for MEI (1)
2.0
0.09
7.2
0.09
42.0
Lowest Presumed
Human Threshold
(10)
2.45
210.0
476.0
97.9 (3)
(2)
Pharmaceutical
Manufacturer
Photo Equip. & Supplies
Pipeline
POTW
Semiconductor Facility
Semiconductor Facility
Ethylene Oxide
Phenol
Benzene
Toluene
Xylene
Benzene
Chloroform
Methylene Chloride
Perchloroethylene
Toluene
1,1,1-Trichloroethane
Benzene
Methylene Chloride
1,1,1-Trichloroethane
Xylene
Methylene Chloride
Trichloroethylene
1.2
21.0
1.8
6.8
9.2
0.02
0.08
0.7
0.2
1.1
0.06
1.8
1.0
5.1
3.3
1.3
14.0
(10)
350.0
2.45
476.0
52.8
2.45
2.4
210.0
69.9
476.0
97.9
2.45
210.0
97.9
52.8
210.0
69.9
(3)
(3)
CO
cr\
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF INDLVILXJAL
RISK ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND
POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
FOOTNOTES TO TABLE 3-31 ARE ON THE FOLLOWING PAGE
-------
FOOTNOTES TO TABLE 3-31
(1) Estimate is based solely on modeled maximum emissions from the source listed. Total ambient concentration
could be higher due to background levels from other sources. See text.
(2) See table 3-4 for all thresholds and non-cancer effects.
(3) The IEMP conducted sensitivity analysis on TCA for possible fetal effects. See footnote to text.
(4) Estimate is for 2-ethoxyethanol, 2-rnethoxyethanol, and 2-butoxyethanol combined.
(5) There is no RfD or IEMD derived threshold for glycol ethers. See text for a discussion of available
information on non-cancer health effects.
(6) At any given moment of pumping gasoline, an individual may be exposed to 490 ug/m^ of benzene. Assuming the
most exposed individual pumps forty gallons a week, he/she may be exposed for five minutes each week. If this
exposure were spread over a year, the annual average would then be 0.17 ug/m^.
(7) At any given moment of pumping gasoline, an individual may be exposed to 49000 ug/m3 of gasoline vapors.
Assuming the most exposed individual pumps forty gallons of gasoline a week, he/she may be exposed for five
minutes each week. If this exposure were spread over a year, the lifetime annual average exposure would then w
be 17 ug/m^. co
^j
(8) Evaluation of potential non-cancer health effects is not complete at this time.
(9) BaP is an organic particulate, not an organic gas; it was included in this table since it was part of the
modeling of the traffic intersection.
(10) Quantification of this threshold is not complete at this time.
(11) Cadmium is a metal, not an organic gas; it was included in this table since it was part of the modeling of the
traffic intersection.
-------
3-88
Table 3-32
Results of Short-Term Monitoring
for Selected Organic Gases in Santa Clara Valley
All values are in micrograms/cubic meter
Highest Monitored
Substance Concentration (1)
Benzene 6.0
Carbon Tetrachlocide 1.2
Chloroform 0.1
Ethylene Dibronide (2)
Ethylene Oxide (2)
Methylene Chloride (2)
Perchloroethylene 6.2
1,1,1-Trichloroethane (2)
Trichloroethylene 1.6
Toluene 14,9
Xylene 8.1
1 These results are based on the short-term monitoring conducted by EPA and
the AQMD in the Fall of 1984.
2 No monitoring data available; the monitoring techniques we used are
unreliable for methylene chloride and TCA.
-------
3-89
The short-term monitoring values for toluene and TCE were also higher than
the model predicted, suggesting that the emissions data we used to model these
concentrations may be incomplete. (The analytical methods used do not address
TCE concentrations well and thus, the estimate from the short-term monitoring is
more uncertain for TCE than for other gases.) The monitoring results for xylene
and perchloroethylene were reasonably consistent with the model. The modeling
results for carbon tetrachloride and chloroform were discussed earlier.
In addition to measuring the ambient levels of compounds in which we were
particularly interested (benzene, perchloroethylene, and so forth), we asked
the laboratory chemists to attempt to identify all of the organic gases in some
of the samples collected during the short-term monitoring. They found a number
of organic compounds, some at high levels, about which we have little informatio'
on sources and still less on toxicology.
Table 3-33 summarizes these data. It should be stressed that the sampling
periods >vere short (24 hours) and that the samples were taken at only Live
locations; it is not possible to conclude from these data that the compounds
are present throughout the valley or that the levels we found in October 1984
are representative of long-term averages. In addition, the laboratory findings
are not definitive, but rather are tentative, since there was not sufficient
evidence to make positive identification of substances. Some substances listed
(with an asterisk.) are most likely present from gases or substances used in the
laboratory testing procedure rather than being from the air sample. FV.rther,
the listed concentrations are only estimates. The data do, however, provide
suggestive evidence that airborne exposures to a large number of compounds are
common.
Most of the identified organic gases are probably constituents of
gasoline and other petroleum-based fuels. They are unsubstituted aliphatics
(chain compounds which contain only carbon and hydrogen). Our analysis of
gasoline vapors may take seme of these compounds into account. Some of the
other compounds are substituted aliphatics. Compounds which contain a single
chlorine atom and otherwise consist solely of carbon and hydrogen appear to be
fairly common. Many of these substances may be attributed to some incomplete
combustion process, such as burning of fuels.
The laboratory identified a number of other organic gases which are not
simple aliphatics; these are also listed on table 3-33. Again, many of these
substances can be attributed to the burning of natural materials.
The laboratory was unable to identify some other substances which were
present in the ambient air samples. Each of the samples analyzed contained at
least one unidentifiable gas, at concentrations which ranged from 0.03 to
70 ug/m3.
The short-term monitoring program included analysis for glycol ethers,
but none were detected in any of the samples. The analytic equipment used,
however, is incapable of detecting glycol ethers except at levels over 5 ug/m3.
The monitoring equipment we used is incapable of detecting the inorganic gases
such as arsine and phosphine.
The IEMP analysis of ambient levels of organic gases was limited by the
data available in the AQMD emissions inventory. The short-term monitoring
results, tentative as they are, do provide some evidence that a wide variety
of organic gases besides the contaminants examined in this study are present
in Santa Clara Valley's outdoor air.
-------
3-90
Table 3-33
Tentatively Identified Organic Gases from
Short-term Ambient Monitoring in Santa Clara Valley
Substance Highest Concentration (ug/m3)
1. Unsubstituted Aliphatics
methylhexane 4
2,2,4-trimethylpentane 2
1-undecene 4
2,4,6-trimethyloctane 2
2,6,10,15-tetramethylheptane 4
pentane 25
3-methylhexane 7
3-dodecene 22
3,8-dimethylundecane 7
2-methylbutane 150
6-ethyl-2-methyloctane 26
6,6-dimethylundecane 10
3-methyl-5-propylnonane 9
4-nonene 2
7-methyloctene 4
1-octene 2
5-methyl-l-hexene 10
3-methylpentane 91
heptane 6
2,2,3,4-tetramethylpentane 44
3,7-diraethylundecane 63
2,6,7-trimethyldecane 58
3-methyldodecane 37
2,2,7-trimethyldecane 26
2,7,10-trimethyldcdecane 33
2,7,7-trimethyldecane 8
methylcyclohexane 6
cyclododecane 31
cyclotetradecane 29
bicyclo [4,2,0] octa-l,3,5-triene 26
octylcyclopropane 24
pentylcyclopropane 9
me thylcyclododecane 13
cyclopentane 11
butylcyclopropane 2
methylcyclopentane 17
-------
3-91
Table 3-33, cont.
Substance
Highest Concentration (ug/m3)
2. Substituted Aliphatics
1-c hlorononane
1-chlorododecane
3-buten-2-ol, 2-methyl
3-penten-2-one
nonanol
4-fnethyl-2-propyl-l-pentanol
1-c hlorooctane
isooctanol
1-chlorooctadecane
3-ethyl-2,7-dimethyloctone
1-chloroheptane
3-methyl-l-hexanol
2,4,6-triinethyloc tone
3-chlorodecane
1-e thylidene-1B-indene
2-cyclohexen-l-trimethylsilane*
pentamethyldisilane*
hexamethylcyclotrisiloxane*
trichlorofluorcmethane*
3. Other Organic Gases
methyl ester, benzole acid
1,2,4-trimethylbenzene
1,2,3-trimethylbenzene
3-methylphenol*
2,3-d imethylphenol*
benzaldehyde*
unknown
unknown alcohol
unknown hydrocarbon
9
25
3
12
20
16
27
31
10
4
12
12
9
35
26
14
4
4
794
24
7
14
22
24
21
70
3
28
*This denotes substances that are most likely present from gases or substances
used in the laboratory testing procedure. These substances are not found in
the actual air samples.
SOURCE: Versar Inc., Analysis of Anbient Air Samples, May 1, 1985.
-------
3-92
Methodology and Findings; Organic Particulates
Our analysis of organic participates in Santa Clara County focused on
polycyclic aromatic hydrocarbons (PAHs), often referred to as Products of
Incomplete Combustion (PICs). The chemical components of PAH can be organized
into three categories: naphthalene, the anthracene group, and the benzo(a)pyrene
group. The chemical constituents of these three groups are listed on table 3-34.
For our analysis of PAHs, we have neither site-specific ambient monitoring
data nor site-specific estimates of emissions. Although the data are extremely
sparse, we included organic particulates in the analysis because a nationwide
study of air toxics conducted by EPA had shown that they could be significant
contributors to airborne risk. (Office of Policy Analysis, "The Air Toxics
Problem in the United States: An Analysis of Cancer Risks for Selected
Pollutants," USEPA, May, 1985.) We decided that it was preferable to provide
rough estimates of potential risk from organic particulates than to ignore this
possible source of risk altogether.
In order to approximate ambient concentrations of PAH in Santa Clara
County, we used two different methods which are described below. With this
information, we then focused on the health effects of the benzo(a)pyrene (BaP)
group of compounds, since there is no known evidence of carcinogenicity for
naphthalene and relatively weak evidence for the anthracene group in comparison
with the BaP group.
Sufficient data on potency do not exist to characterize the health risks
associated with exposure to the BaP group on a compound by compound basis.
Thus, to facilitate our screening analysis, we applied the potency value for
BaP, a probable human carcinogen for which good toxicological data exist, to
the entire BaP group of compounds. Previous EPA analyses have taken a similar
approach, using BaP as a proxy to estimate the cancer risk from PAHs as a whole.
Sources of Organic Particulates
The combustion of wood and coal for residential heating accounts for the
vast majority of PAH air emissions in this country. Two other sources are
forest fires and agricultural burning. However, these Latter sources are, in
general, not applicable to urban areas such as Santa Clara County. Consequently,
the vast majority of air releases of PAH in the County can be attributed to
residential heating. The primary sources of BaP in Santa Clara County are
thought to be residential heating and motor vehicles. (Memorandum to Eileen
Soffer [IEMP] from Dennis Hlinka [Versar, Inc.], "Updated Santa Clara Valley
Study Report," May 12, 1986.)
Methodology for Estimating Exposures
As mentioned, we used two different methods to estimate ambient levels of
organic particulates in Santa Clara Valley. The first method involved relating
benzo(a)pyrene concentrations to potential PAH concentrations based on estimated
emissions of both BaP and PAH. The estimated range of BaP concentrations we
used is based on concentrations from 46 urban areas similar to Santa Clara
County. Emissions rates were estimated by scaling down national emissions data
-------
3-93
Table 3-34
The Chemical Constituents of Polycyclic Arcmatic Hydrocarbon (PAH)
1. Naphthalene
2. The Anthracene Group
anthracene
acenaphthene
fluorene
phenanthrene
pyrene
fluoranthene
3. The Benzo(a)pyrene Group
benzo(a)pyrene
acenaphthylene
benzo(a)anthracene
chrysene
dibenz(a,h)anthracene
benzo(b)f1uoranthene
benzo(k)fluoranthene
benzo(g,h,i)perylene
indeno(1,2,3-c,d)pyrene
-------
3-94
to reflect relevant local factors such as population and local fuel consumption.
Using this information, we estimated a range of ambient concentration of PAH of
1 to 60 ng/m3 (0.001 to 0.06 ug/m3). we want to emphasize that this range is
based on the total range in urban concentrations for the 46 areas but does not
include potential variances in the emissions estimates, factors for the different
reactivity of BaP and PAH, or variances to account for the different particle
size distribution of BaP and PAH. The range is intended to serve only as a
screening level estimate.
Our second method of estimating ambient concentrations of PAH in Santa
Clara County used monitoring ratios of Total Suspended Participates (TSP) to
PAH from other cities to derive crude approximations of potential PAH concen-
trations based on TSP monitoring data from Santa Clara County. We used ratios
of TSP to PAH fron three independent sets of monitoring data. The first two
ratios came from monitoring data taken from an area outside Wilrijk, Belgium,
in 1978. The third ratio came from monitoring data from a suburban site near
Birmingham, England. The TSP concentration we used for Santa Clara County was
64.7 ug/m3 (64,700 ng/m3)the 1983 annual average concentration from the Fourth
Street monitoring site. Based on the given ratios and this monitoring data,
the estimated concentration of PAH ranges from 18 ng/m3 to 125 ng/m3 (0.018 to
0.125 ug/m3).
Like the first method, this method of calculating PAH concentrations has
substantial uncertainties. There is no known correlation between TSP/PAH
ratios in Santa Clara County and the two cities monitored in Europe. Further-
more, the different monitoring techniques used, the time of year, and the
different number of compounds measured in the two cities add a considerable
amount of potential error to this technique. Again, the range is intended to
serve as a screening level estimate only.
Together, the two methods provide an estimated range of PAH concentrations
in Santa Clara County of 1 to 125 ng/m3 (0.001 to 0.125 ug/m3). (For a more
detailed discussion of this analysis, see Memorandum to Eileen Softer [IEMP]
from Dennis Hlinka [Versar, Inc.], "Updated Santa Clara Valley Study Report,"
May 12, 1986.)
Analysis of Health Risks
In order to gain some perspective on the health effects of these concen-
trations, it is useful to consider the three major groups of chemical compon-
ents of PAH individually. As mentioned, there is no evidence for carcinogeni-
city for naphthalene and relatively weak evidence for the anthracene group in
comparison to the benzo(a)pyrene group. As such, we considered the health
effects of the BaP group only.
Based on national emissions data, adjusted to consider only those sources
applicable to Santa Clara County, the BaP group comprises roughly 16% of total
PAH emissions. Therefore, by taking 16% of the PAH concentration in Santa
Clara County (1 to 125 ng/m3), we arrive at an estimated average concentration
of 0.2 to 20 ng/m3 (0.0002 to 0.02 ug/m3) for the BaP group in the County.
In order to determine the cancer risks associated with exposure to this
concentration of the BaP group, we multiply the concentration by a value which
-------
3-95
represents the plausible upper-bound estimate of the carcinogenic potency of
the chemicals. To facilitate such a calculation, we assumed that all of the
compounds in the BaP group have the same potency value as does BaPthe only
chemical in the group for which we have sufficient toxicological data. As
stated, previous EPA studies have used BaP in a similar way as a proxy to
estimate the cancer risk from PAHs a whole.
Following this approach, we conservatively estimate the lifetime individual
risk of cancer due to continuous exposure to these levels of the benzo(a)pyrene
group to be between 7 x 10~7 and 7 x 10"^, or between 0.7 in a million and 70
in a million. If the entire population of the County were exposed to these
estimated levels throughout their lifetimes, a conservative estimate for expected
annual increase in cancer rates would be between 0.01 and 1.3 annually, with an
average of 0.7 cases annually, or one case every 1.5 years.
BaP has been implicated in connection with other non-cancer health effects,
as presented on table 3-4. However, quantification of thresholds for such effects
is not complete at this time.
-------
3-96
Summary and Conclusions
1. The impact of air toxics on public health can be estimated for screening
purposes, although the estimates are subject to considerable uncertainty and
do not represent total airborne risk in the Santa Clara Valley.
We attempted to estimate, in a conservative way, the health effects due
to exposures to eight metals, eleven organic gases, the mixture of compounds in
gasoline vapors, and one class of organic particulates in Santa Clara Valley's
outdoor air. Every attempt was made to use the best available data. However,
existing data on the toxicology, emissions and ambient levels for air toxics
are limited and of varying quality. Because of uncertainties in the underlying
data and assumptions, estimates of individual risk and disease incidence are
only rough approximations of actual risk. (The key assumptions in the air
analysis are discussed in the relevant sections in this chapter and are outlined
in the following section.) As more data become available, both in terms of
exposure and toxicology, the risk estimates will quite likely change. The
portrait presented here should be regarded as a "snapshot" that represents the
best assessment given the available tine and resources.
Our estimates of health effects are designed to be conservative in several
ways. In examining the risks of non-cancer health effects, we use estimated
human thresholds as noted in table 3-4. The thresholds represent the level of
exposure below which adverse effects are assumed not to occur. Because these
thresholds incorporate safety factors, they are considered to be conservative
estimates of an actual threshold level.
In estimating cancer risks, we use cancer potency values developed by
EPA's Carcinogen Assessment Group (CAG) as noted in table 3-3. These values
are generally regarded as plausible upper-bound estimates rather than "best-
guess" estimates. That is, the actual potency is not likely to be higher, but
could be considerably lower. In addition, the weight of evidence of carcino-
genicity for the compounds examined varies as noted in table 3-3. For our
analysis we treated substances that are suspected carcinogens as human
carcinogens. Thus, for each of the pollutants for which we present cancer
risk estimates, the estimates are more likely to overestimate risks than to
underestimate them.
In this analysis, we assume that the cancer risk estimates can simply be
added to obtain an estimate of risk from exposure to several compounds. It
is not clear whether this assumption is conservative or not. If several
substances act in synergy, the risk of cancer from exposure to those substances
may be greater than the sum of the individual risks. This is known to be the
case for certain combinations of substances, such as asbestos and cigarette
smoke. It is also possible, however, that the cumulative effects of exposures
to several substances may be less than the sum of the individual risk estimates.
An important way in which this analysis is not conservative is that, due
to limitations in the data and in the scope of the analysis, we were unable to
analyze all of the substances and exposure pathways that may contribute to
chronic health effects in Santa Clara Valley. Table 3-35 outlines some of the
specific omissions of this analysis, which are discussed in greater detail else-
where in this chapter. Despite these specific omissions, we have tried to
-------
Table 3-35
Specific Substances and Exposure Pathways Not Included in the Santa Clara Valley IEMP
Air Analysis Due to Limitations in Data or in the Scone of the Analysis
Substance or
Exposure Pathway
Reason Not Included
Comments
Lead
Formaldehyde
Arsine, Phosphine
and other
compounds used
by local industry
Limitations in toxicology
assessment.
Dispersion model used is
not able to account for
complex atmospheric
transformations.
Lack of emissions
data.
Could pose health hazard in Santa Clara Valley,
particularly for children.
According to the AQMD inventories, emissions are
large relative to the other compounds studied.
Estimated emissions are almost entirely from motor
vehicles. Substance is considered carcinogenic.
More research would be needed for even a preliminary
assessment of their potential significance.
u>
i
Criteria Pollutants
Asbestos
Indoor Air
Episodic Releases
Outside scope of analysis.
Outside scope of analysis,
Outside scope of analysis.
Outside scope of analysis.
Except for some violations of the ozone and carbon
monoxide targets, the air in Santa Clara County
meets ambient standards set to protect public health.
Outdoor air exposures to asbestos have been raised as
a concern in Alviso following the finding of asbestos
wastes in the soil. EPA and the California Depart-
ment of Health Services have conducted some emergency
remedial work. EPA and the Army Corps of Engineers
will be overseeing long-term efforts to define the
extent of the problem and explore remedial actions.
Other research suggests that indoor air exposures
may exceed exposures in outdoor air for selected
compounds.
More research would be needed for even a preliminary
assessment of their mtential significance.
-------
3-98
focus on the more important substances, based on our current understanding of
toxic pollutants; however, urban air is characterized by hundreds, perhaps
thousands of substances for which little information is available. The short-
term monitoring conducted as part of the IEMP Stage I analysis provides some
evidence that a wide variety of organic gases besides those examined in this
study are present in ambient air in Santa Clara Valley. As such, it is possible
that analyses such as the IEMP air analysis, which focus on a select set of
compounds, may account for only a small percentage of total health risks from
airborne exposures.
Thus, the risk estimates we present tend to be conservative for the
substances studied, based as they are on plausible upper-bound estimates of
toxicological potency. However, the overall analysis of risks from air
pollutants may not be conservative insofar as we were not able to analyze
every potentially toxic air contaminant and every exposure pathway due to
limitations in the data and in the scope of the analysis. Given the limita-
tions, the results cannot be interpreted as an estimate of total airborne risk
in Santa Clara Valley.
Despite the uncertainties and limitations of the estimates, they are
valuable in that they assess risks to the extent possible so that appropriate
steps can be taken to reduce those risks that are identifiable. The risk
estimates presented in this analysis enable us to achieve the goals of the
IEMP Stage I analysis: to identify, evaluate, and compare potential human
health risks attributable to various pollutants, sources and exposure pathways;
and based on this evaluation, set priorities for further research and possible
control.
2. We estimate that exposures to the outdoor toxic air pollutants and sources
studied in Santa Clara Valley may be responsible for approximately two additional
cases of cancer per year.
According to EPA's models of carcinogenicity, exposures to the eight
metals, eleven organic gases, the mixture of compounds in gasoline vapors, and
the class of organic particulates included in this analysis may be responsible
for approximately 2 cases of cancer per year beyond what would be expected in
the absence of those contaminants in Santa Clara Valley. Each of the three
categories of air pollutantsmetals, organic gases, and organic particulates
account for roughly one-third of the estimated aggregate risk. These estimates
are presented in table 3-36.
The point estimate of about 2 cases is derived from an estimated range of
roughly 1 to 8 additional cases of cancer each year from exposure to all of
the substances we analyzed in outdoor air. The point estimate incorporates a
number of assumptions.* The most significant assumption centers on the risk
* In our base-case analysis we assume, in accordance with EPA's current policy,
that 1,1,1-trichloroethane (TCA) is not a carcinogen. As a sensitivity analysis,
however, we examine the potential impacts of TCA if it were a carcinogen. (See
the section "Overview of the Methodology" in this chapter for a more complete
discussion of this issue.) Assuming TCA is carcinogenic, the estimated cancer
risk from exposures to this compound is relatively low and would not change the
total estimated risk.
-------
3-99
Table 3-36
Estimated Annual Increase in Cancer Incidence
from Toxic Pollutants in Santa Clara Valley's Outdoor Air
Point Estimate of
Aggregate Annual
Increase in
Range of Aggregate
Cancer Incidence(l)
Substance (Percent of Total)
Metals:
Arsenic
Barium (4)
Beryllium
Cadmium
Chromium
Lead (5)
Nickel
Zinc (4)
Subtotal
Organic Gases:
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Dibromide (6)
Ethylene Oxide
Gasoline Vapors
Methylene Chloride
Perchloroethylene
1,1,1-Trichloroe thane (7)
Trichloroethylene
Toluene (4)
Xylene ( 4 )
0.3
0.003
0.07
0.4
0.03
0.8 (36%)
0.3
0.2
0.001
0.004
0.03
0.1
0.01
0.03
0.002
-.
Annual Increase in
Cancer Incidence (2)
0.2 to 0.4
0.001 to 0.005
0.04 to 0.1
0 to 4.0
0 to 0.06
0.2 to 4.6
0.3 to 1.2
0.2
0.001
0.004
0.03
0 to 0.4
0.01
0.03
0.002
Level of
Evidence (3)
A
-
B2
Bl
A
-
A
-
A
B2
B2
B2
Bl
B2
B2
B2
B2
Subtotal
Organic
Particulates
Benzo(a)pyrene
0.7 (32%)
0.7 (32%)
0.6
to 1.9
0.01 to 1.3
B2
TOTAL
2.2 (100%)
0.8
to 7.8
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND
ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
3-100
FOOTNOTES TO TABLE 3-36
1 The point estimates incorporate a number of important assumptions, particularly
for chromium, benzene, and gasoline vapors. See text.
2 For some of the organic gases, no range was calculated. The point estimate
for these compounds is a population weighted average based on the modeled
range of individual risk.
3 EPA classification for level of evidence for carcinogenicity: A = carcinogenic
in humans; Bl = probably carcinogenic in humans (limited human evidence);
B2 = probably carcinogenic in humans (insufficient human evidence but sufficient
animal evidence); C = possibly carcinogenic in humans.
4 At this time there appears to be insufficient data for an adequate evaluation
of the potential carcinogenicity of this compound. Therefore, we do not analyze
it as a carcinogen.
5 While there is limited evidence that .some lead compounds induce tumors in
experimental animals, GAG has not yet made a determination as to its potential
carcinogenicity. Therefore, lead is not analyzed as a carcinogen in this
report.
6 The estimates for EDB are based on a cancer potency value derived from
ingestion studies; inhalation value may be lower.
7 In our base-case analysis, we assume, in accordance with EPA's current
policy, that 1,1,1-trichloroethane (TCA) is not a carcinogen. As a sensitivity
analysis, however, we examine the potential impacts of TCA if it were a carcinogen.
(See the section "Overview of the Methodology" in this chapter for a more
complete discussion of this issue.) If TCA is carcinogenic, then the increased
incidence from exposures to this compound is 0.01 cases per year. Because
the estimated risk is relatively low, the totals are unchanged under this
alternative.
-------
3-101
posed by exposures to the toxic metal chromium. There is uncertainty regarding
the portion of the total chromium monitored in Santa Clara County's ambient air
that is hexavalent, the carcinogenic form. Our estimated range of risk reflects
this uncertainty: if none is hexavalent, we estimate no increased risk; if all
is hexavalent, we estimate up to 4 additional cases each year from exposure to
levels monitored in the Valley. Based on information from other studies and
limited information regarding the possible local sources of chromium, we assume
that 10% of the chromium in the Valley's ambient air is hexavalent. (See the
section on "Estimated Health Risks from Exposure to Metals" in this chapter for
a more complete discussion of this issue.) As such, our point estimate for
increased annual incidence from exposure to chromium is 0.4 cases. The magnitude
of the health risk posed by chromium clearly depends on the assumptions one
makes about the portion that is hexavalent.
Similarly, there are difficulties in estimating the risk from exposure to
gasoline vapors. First, there is uncertainty with respect to the relevance to
humans of existing evidence of carcinogenicity; in addition, there is evidence
that even if gasoline vapors are carcinogenic, CAG's potency value may over-
estimate the effects by roughly a factor of four. We estimated an aggregate
risk of cancer from gasoline vapors ranging from zero, if it is not carcinogenic
in humans, to 0.4, if one applies CAG's upper-bound estimate of carcinogenic
potency. For our base case we assumed, in accordance with EPA's current policy,
that gasoline vapors are carcinogenic; however, based on other studies and the
approach taken by EPA's Office of Air Quality Planning and Standards, we divided
the risk estimates by a factor of four to account for the lesser volatility of
some gasoline constituents suspected of toxic effects. (See the section on
"Estimated Health Risks from Exposure to Organic Gases" in this chapter for a
more complete discussion of this issue.) Thus, our point estimate for increased
incidence from exposure to gasoline vapors is 0.1 per year.
In addition, our air dispersion model predicts concentrations of benzene
ranging from 0.2 to 2.6 ug/m-*. On the basis of this exposure level, we estimate
an annual incidence of 0.3 attributable to benzene. However, EPA has found in
the Philadelphia IEMD study that the dispersion model may underestimate the
ambient level of persistent compounds such as benzene by a factor of two or
three. In addition, our short-term monitoring and monitoring conducted by the
ARB over a nine-month period also suggest that the estimated levels of benzene
may be too low by a factor of two to four. As such, the estimated incidence
from benzene could be as high as 1.2 cases per year. For our base case, which
the point estimate reflects, we have used the estimate of 0.3 cases per year
based on the dispersion model results; the higher estimate is reflected in the
estimated range of risk.
Finally, we estimated a wide range of risk from exposures to the benzo(a)-
pyrene group of organic particulates based primarily on studies conducted in
other cities. The estimated increase in annual cancer incidence from exposure
to benzo(a)pyrene ranged from 0.01 to 1.3 annually. This range reflects the
uncertainty involved in extrapolating information from other metropolitan areas
to Santa Clara Valley. The midpoint of this range, 0.7 annual cases, is the
estimate incorporated in our point estimate of a total annual cases.
For all other substances contributing to cancer risk, the point estimate
used was the simple average (for metals) or the population-weighted average
(for modeled organic gases) of the estimated range of risk.
-------
3-102
The average individual lifetime risk of cancer from exposure to all of the
air toxics we studied that corresponds to the point estimate of approximately 2
cases per year is roughly 100 in a million (or one in ten thousand). This is
the estimated lifetime risk of cancer for typical individuals exposed to all
the air pollutants we studied; individuals who are highly exposed to the pollu-
tants, by virtue of their proximity to a source, may face significantly higher-
than-average cancer risk. (Risks to such maximum exposed individuals are
discussed below.)
These estimates of individual risk and incidence are designed to be conser-
vative for those substances we analyzed. The estimate of 2 cases per year is a
statistical measure of increased cancer cases, not death. An estimated 3,600
cases of cancer occur annually in Santa Clara County.* Appendix A provides
other statistical measures of risk and a general discussion of interpreting
risk estimates.
3. We estimate that about 10% of the population in Santa Clara Valley may be
exposed to one chemicalbenzeneat levels high enough to pose a risk of
effects other than cancer. Annual average levels of all the other air toxics
appear to be below the thresholds at which one would expect an increase in the
risk of chronic non-cancer health effects.
In addition to estirnating the risk of cancer from selected air toxics, we
examined exposures to see if they exceeded thresholds for other chronic (non-
cancer) health effects. With the exception of benzene, amhient levels of the
air toxics monitored or modeled in Santa Clara County's outdoor air for which
we have quantified thresholds appear to be below the levels at which one would
expect an increase in the risk of non-cancer health effects.
We modeled annual average concentrations of benzene in Santa Clara Valley
ranging from 0.2 to 2.6 ug/m^. The higher end of this range, which was modeled
in the northern part of the County, is at the threshold for blood effects. We
estimate that roughly 100,000 people may be at some risk of blood effects from
benzene exposures. Blood effects usually involve decreased blood cell counts
to an extent to cause anemia (i.e., depressed red blood cell count) or to make
an individual prone to infection (i.e., leukopenia or depressed white blood
cell count.) The assumed human threshold for blood effects from benzene exposure
is 2.45 ug/m-^. Note that the threshold is computed in a manner designed to be
highly conservative; it is not certain that exposures at or near the threshold
actually will result in increased risk. Further work might be appropriate to
verify these findings and better assess the magnitude of the potential risk
from benzene, particularly given that the model may be underestimating ambient
benzene exposures by a factor of two or so.
Bear in mind that our estimates of benzene exposures and risk are based on
current conditions. Benzene exposures are likely to be affected by a number of
potential future activities. For example, benzene emissions from evaporation
and combustion of gasoline may increase by an estimated 10% in the state by
* Ratio of estimated cancer cases to estimated cancer deaths, from 1983 data
from the American Cancer Society: 1.93. Cancer deaths in Santa Clara County,
1984: 1,879. 1,879 cancer deaths x 1.93 cases/death = 3,626 estimated cancel-
cases in Santa Clara County in 1984.
-------
3-103
the year 2000 due to the phase-out of lead. On the other hand, emissions may
be reduced under a plan outlined by the ARB staff in a report released in May
1986. In the report, the ARB staff expressed its belief that the risk in
California from exposure to ambient benzene is significant and not protective
of public health. They outline specific control measures for benzene which, if
approved and fully implemented, may result in a 50% reduction in benzene emissions
statewide by the year 2000, with accompanying reductions in ambient levels.
Although we were unable to estimate the potential impacts of exposures to
lead in this report, such exposures in Santa Clara Valley could pose health
hazards. There is substantial evidence that lead can have adverse health
effectssuch as blood effects, neurological dysfunctions, and decreased IQ
particularly for children. The TEMP hopes to examine the potential for such
effects as part of its Stage II analysis.
It is important to note that exposures to toxic chemicals in the air may
pose some health risk at levels below estimated thresholds if exposures from
other sourcessuch as diet, drinking water, or occupationare significant.
Even in this instance, comparisons with estimated thresholds provide a useful
indication of the significance of the portion of exposure due to outdoor air.
4. Roughly 90% of the cancer risk we estimated from outdoor air exposures in
Santa Clara Valley is attributable to six of the pollutants studied.
The most important air toxic pollutants of those we studied, in terms of
their projected impacts on human health, are arsenic, benzene, benzo(a)pyrene,
carbon tetrachloride, chromium and the mix of compounds in gasoline vapors. We
estimate that these contaminants account for roughly 90% of the estimated
cancer risk from air toxics. That is, approximately 2 excess cases of cancer
each year in Santa Clara Valley may be attributed to these contaminants. These
estimates are presented in table 3-37.
Chronium is potentially the most important substance of all in terms of
its impact on risk, although its significance depends on what portion of the
chrcmium monitored in Santa Clara County's air is hexavalent. As a base case,
we assume that 10% of the chromium in the County's ambient air is hexavalent.
(See the section "Estimated Health Risks fron Exposure to Metals" in this
chapter for a discussion of the basis for this assumption.)
Benzene, another substance identified as being significant in terms of
health impacts, is currently being considered for control by the ARB. In a
report released in May, the ARB staff outlined specific control measures for
benzene. If the plan is approved and fully implemented, the ARB estimates a
reduction in statewide cancer risk of 39% between 1984 and 2000. However, some
of this estimated reduction in risk would result from statewide implementation
of Stage I and Stage II vapor recovery. Such controls are already in place in
Santa Clara County-'--the ACMD has required them in the Bay Area since the mid-
seventies.
-------
3-104
Table 3-37
Estimated Annual Increase in Cancer Incidence in Santa Clara Valley
Attributable to Selected Air Toxics
Point Estimate of
Aqgregate Annual Increase
in Cancer Incidence Percent Level of
Substance (Range of Risk)(l) of Total Evidence (2)
Arsenic 0.3 (0.2 to 0.4) 14% A
Benzene 0,3 (0.3 to 1.2) 14% A
Benzo(a)pyrene
(group of canoounds) 0.7 (0.01 to 1.3) 32% B2
Carbon Tetrachloride 0.2 9% B2
Chromium 0.4 (0 to 4.0) 18% A
Gasoline Vapors
(mix of compounds) 0.1 (0 to 0.4) 5% B2
Subtotal 2.0 (0.5 to 7.3) 91%
All Others 0.2 (0.3 to 0.5) 9% A-B2
TOTAL 2.2 (0.8 to 7.8) 100%
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS
OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND
POTENCY AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM.
SEE TEXT,
(1) The point estimates incorporate a number of important assumptions,
particularly for chromium, benzene, and gasoline vapors. See text.
(2) See previous table for a definition of the level of evidence for
carcinogenicity.
-------
3-105
5. More than 75% of the cancer risk we estimated from airborne exposures
cones from pollutants whose sources are poorly categorized.
There is a great deal of uncertainty over the sources of metals and
organic particulates (benzo(a)pyrene) in Santa Clara Valley. The AQMD does
not maintain emissions inventories for these substances as it does for most of
the organic gases we studied. The IEMP Draft Stage I Report (October 1985)
identified the lack of local data on sources and emissions of metals as a
significant data gap. As part of its recently adopted air toxics program, the
AQMD, with assistance from EPA, is planning to compile an emissions inventory
for the District, which includes Santa Clara County. However, such information
is not currently available.
In addition, there is uncertainty over the sources of carbon tetrachloride
in the County. Other studies suggest that there may be a global background
level of carbon tetrachloride of roughly 0.7 ug/m^. Our concentrations data
for this compound came from short-term monitoring of background levels and, as
such, cannot be attributed to any particular source category.
As a result, roughly 77% of the estimated cancer risk from the air toxics
we studied is from uncertain sources. Of the contaminants discussed above as
contributing most substantially to cancer risk, only twobenzene and gasoline
vaporsare from well categorized sources. Both are present in Santa Clara
Valley's ambient air primarily because of emissions from motor vehicles.
These findings are presented in table 3-38.
One important conclusion of this study is that, although the ambient
levels of most of these substances are comparable to or lower than those foun^
in other cities, and although the cancer risks appear low relative to the
overall cancer incidence in the County, more work should be done to characterize
the sources of the metals and BaP. Such source information is important in
estimating exposures and resultant risks. Better source information for chromium,
in particular, is important because estimates of the portion of chromium in the
ambient air that is hexavalent depend, to a certain extent, on the source of
the chromium. With more reliable source information, we may be better able to
assess the magnitude of the cancer risk from exposure to chromium. In addition,
reliable source information is important as a risk management tool in that it
allows us to estimate the impact of various pollution control options. The
metals emissions inventory that the AQMD plans to compile will fill a significant
data gap and will assist ongoing efforts to assess both risks and control options.
6. Preliminary estimates based on limited information (designed to guide
further research) suggest that most of the cancer risk from air toxics may be
attributable to residential heating, cars and other area sources.
As mentioned above, most of the estimated cancer risk from air toxics is
from sources that are poorly categorized. Benzene and gasoline vapors, which
together account for roughly 19% of the estimated risk, are present in Santa
Clara Valley's air primarily from motor vehicle emissions. Little information
is available on the sources of benzo(a)pyrene, carbon tetrachloride, arsenic,
and chromium the four other pollutants that have a significant impact on
estimated cancer risk.
-------
3-106
Table 3-38
Estimated Annual Increase in Cancer Incidence in Santa Clara Valley
Attributable to Uncertain vs Identified Sources
Substance
Point Estimate of Aggregate Annual
Increase in Cancer Incidence (1) Level of
(Percent of Total) Evidence;2)
Uncertain Sources
Arsenic
Benzo(a)pyrene
(group of compounds)
Carbon Tetrachloride
Chromium
Other Metals
0.3 (14%)
0.7
0.2
0.4
0.1
(32%)
( 9%)
(18%)
( 5%)
B2
B2
A
A-B2
All Uncertain Sources:
1.7 (77%;
Identified Sources
Benzene
Gasoline Vapors
(mix of compounds)
Other Organic Gases
All Identified Sources:
0.3 (14%)
0.1 ( 5%)
0.08 ( 4%)
0.5 (23%)
B2
B1-B2
TOTAL
2.2 (100%)
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND
ARE MORE LIKELY TO OVERESTIMATE THAN UNDERESTIMATE THEM. SEE TEXT.
(1) The point estimates incorporate a number of important assumptions,
particularly for chromium, benzene and gasoline vapors. See text.
(2) See table 3-36 for definition of the level of evidence for cancer. See text.
-------
3-107
Since source information is an important risk management tool, the IEMP
took the first steps toward identifying potential sources of metals and organic
particulates in Santa Clara Valley. The purpose of this preliminary assessment
is to identify sources that may warrant follow-up, not to make definitive state-
ments regarding the sources of risk. For this purpose, the IEMP used the
limited data available to calculate rough estimates of emissions from three
source categories that we thought might be the most significant in the Valley:
motor vehicles, fuel combustion sources, and a coal-burning cement plant. The
available information consisted of preliminary source testing of the cement
plant and information contained in the national literature on emissions of
organic particulates and metals from motor vehicles and combustion sources.
Using these emissions estimates, we apportioned the estimated cancer risk from
metals and organic particulates to these sources. (See section "Estimated
Health Risks from Exposures to Metals" for a more complete discussion of the
analysis.)
One of the greatest uncertainties in such an analysis centers on appor-
tioning the risk to a point source such as the cement plant. Based on the
AQMD's knowledge of the location of the plant and the surrounding terrain and
meteorology, they believe that the plant's impact on ambient levels may be
quite localized, and the extent of exposures to these levels is uncertain.
Thus, unlike the emissions from dispersed area sourcescars and fuel burning
the cement plant emissions may not result in measurable exposures. We took
this possibility into account by analyzing two casesone that includes the
cement plant and one that does not.
Table 3-39 presents the results of these analyses. In Case 1 we assume
that the cement plant emissions do result in exposures and risks. In Case 2
we assume that the cement plant emissions do not result in measurable exposures,
and thus exclude the plant as a source in apportioning the estimated cancer
risk. Based on these two cases, we tentatively estimate that between 29% and
50% of the cancer risk may be attributed to combustion of fuel, primarily
wood, for residential heating. Motor vehicles may account for roughly 30%
of cancer risk from air pollutants. Other area sources combined appear to
account for an additional 7% of the risk, so that roughly 64% to 86% of the
estimated risk can be attributed to area sources as a whole. The coal and
coke burning cement plant may account for zero to 22% of the risk depending
upon assumptions regarding exposures to its emissions; the other 25 point
sources together account for about 5%; and carbon tetrachloride (which we
cannot attribute to either point or area sources due to insufficient data) for
another 9%.
This preliminary finding that between 64% and 86% of the estimated cancer
risk may be attributable to area sources is important since control of dispersed
area sources is relatively difficult and expensive. Bear in mind, however, that
these estimates are preliminary and uncertain. They are rough approximations
based on limited data. The purpose of the preliminary identification of sources
is to identify areas that may warrant further research. More work would need
to be done to characterize the sources of metals, BaP and carbon tetrachloride
in the County, such as the compilation of a metals emissions inventory as
planned by the AQMD, before more certain statements can be made about the
sources of the estimated cancer risk.
-------
3-108
Table 3-39
Estimated Annual Increase in Cancer Incidence in Santa Clara Valley
Tentative Source Identification
Source Category
CASE #1(1)
Point Estimate of
Annual Increase
in Cancer Incidence (3)
(Percent of Total)
CASE #2 (2)
Point Estimate of
Annual Increase
in Cancer Incidence (3)
(Percent of Total)
Area Sources
Residential Heating
Motor Vehicles
Other Area Sources
Area Source Subtotal:
Point Sources
Cement Plant
25 Point Sources
Point Source Subtotal
Background
Carbon Tetrachloride
(background)
Background Subtotal
TOTAL
0.63 (29%)
0.63 (29%)
0.15 ( 7%)
1.4 (64%)
0.5 (22%)
0.1 ( 5%)
0.6 (27%)
0.2 ( 9%)
0.2 ( 9%)
2.2 (100%)
1.1 (50%)
0.67 (30%)
0.15 ( 7%)
1.9 (86%)
0.1 ( 5%)
0.1 ( 5%)
0.2 ( 9%)
0.2 ( 9%)
2.2 (100%)
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND
ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
THIS SOURCE IDENTIFICATION IS PRELIMINARY AND UNCERTAIN. IT IS BASED TO A
LARGE EXTENT ON ESTIMATED EMISSIONS FROM VARIOUS SOURCES WHICH, IN MANY CASES,
HAVE NOT BEEN CONFIRMED BY SOURCE TESTING. THIS TENTATIVE IDENTIFICATION IS
DESIGNED TO IDENTIFY AREAS THAT MAY WARRANT FOLLOW-UP WORK. SEE TEXT.
FOOTNOTES TO THIS TABLE ARE ON THE FOLLOWING PAGE
-------
3-109
FOOTNOTES TO TABLE 3-39
1 In Case #1, we have apportioned estimated risk from metals and organic
particulates to all the sources we examined, including the cement plant,
in proportion to their estimated emissions. Estimated risk from organic gases
was attributed to sources based on the AQMD inventory and modeling results.
See text.
2 In Case #2, we have apportioned estimated risk from metals and organic
particulates to all the area sources we examined, excluding the one point
sourcethe cement plant. Estimated risk from organic gases was attributed
to sources based on the AQMD inventory and modeling results. See text.
3 The point estimates incorporate a number of important assumptions, particularly
for chromium, benzene, and gasoline vapors. See text.
-------
3-110
7. Sane individuals who live near pollution sources face toxic health risks
that appear to be significantly higher than average in Santa Clara Valley.
Estimated exposures to the maximum exposed individual (MET) near sources
of air toxics were typically five to one hundred times higher than average
concentration levels. The difference between the average and MEI exposures was
greater for chemicals such as ethylene oxide and chloroform whose emissions were
dominated by a few point sources, and less for chemicals such as toluene and
xylene, which are emitted by many dispersed sources.
One of the highest individual lifetime risks of cancer we calculated was
for exposures to individuals living in the immediate vicinity of a heavily
traveled intersection. We estimate an individual lifetime risk of cancer for
such maximum exposed individuals of 300 in a million (or three in ten thousand).
The pollutants we modeled at the intersection that contribute to this risk
include benzene, benzo(a)pyrene, cadmium, and ethylene dibromide. Benzene
contributed most significantly to this estimate, accounting for an individual
lifetime risk of cancer of 200 in a million (or two in ten thousand). In
comparison, the average individual lifetime risk of cancer from exposure to
benzene is estimated to be 20 in a million.
Ws also estimated relatively high individual lifetime risks of cancer
for the maximum exposed individuals living immediately downwind of large
hospitals using ethylene oxide as a sterilant. Using conservative assumptions
about ethylene oxide emissions from hospitals, we estimate that the increased
risk of cancer to such individuals could be up to 200 in a million (or two in
ten thousand). In comparison, the estimated increased risk of cancer from
ethylene oxide to the average exposed person in the County is 2 in a million.
Ambient monitoring or further work on estimating emission rates and the
reactivity of ethylene oxide may be appropriate in order to make a better
assessment of the health risks to people living in the immediate vicinity of
point sources emitting ethylene oxide.
The ARB, in conjunction with the California Department of Health Services,
is currently preparing a report on the levels of exposure to ethylene oxide in
the state and the health effects associated with exposure to this compound. A
draft report should be available for public comment in Summer 1986. If the ARB
identifies ethylene oxide as a Toxic Air Contaminant, this report will serve as
a basis for future regulatory actions.
In addition to facing an increased risk of cancer, the maximum exposed
individual may be at some risk of non-cancer health effects from exposures to
benzene. Concentrations directly downwind of the largest industrial point
source of benzene identified by the AQMD were modeled at 4 ug/m3; concentrations
in the immediate vicinity of the major intersection were modeled at 25.6 ug/m3.
The assumed human threshold for blood effects from exposure to benzene is
conservatively estimated at 2.45 ug/m3 and the threshold for fetal effects
is estimated to be 4 ug/m3. As such, people living in the immediate vicinity
of the sources we modeled may be at some risk of these effects. As described
earlier, blood effects may involve decreased blood cell counts to an extent to
cause anemia (i.e., depressed red blood cell count) or to make an individual
prone to infection (i.e., leukopenia or depressed white blood cell count).
Fetal effects consist of delayed fetal growth. Further work on estimating
actual exposures would be necessary to verify these findings.
-------
3-111
Table 3-40 summarizes the risks from different pollutants faced by
individuals living in the immediate vicinity of a number of point sources we
modeled. The table compares the estimated maximum individual cancer risk with
the estimated average risk facing County residents.
We estimated risks for the maximum exposed individual from a number of
other sources and organic gases in addition to the ones discussed here (see
section on "Estimated Heath Risks from Exposure to Organic Gases" in this
chapter). These sources included groundwater aeration facilities, drycleaning
facilities, gas stations and other industrial facilities. For all of the
additional sources and pollutants we modeled, the estimated individual lifetime
risk of cancer was lower than the estimated risk from the sources and pollutants
presented above; in addition, for all of the other sources and pollutants, the
maximum modeled exposures were below the levels we conservatively estimate
would be required to produce a risk of non-cancer health effects.*
We were unable, in general, to calculate risks to the most exposed
individual from exposures to metals since, without reliable emissions data, we
cannot model concentrations from various sources. Our preliminary analysis of
possible sources of metals suggests that it would be useful to analyze risks
for metals exposures near traffic intersections and the cement plant. The AQMD
is currently conducting a limited modeling exercise to estimate exposures to
metals for populations near the cement plant. However, no results were available
at the time of this writing.
* TCA has not demonstrated any teratogenic potential in published studies
conducted using rodent species. Therefore, the IEMP base-case analysis assumes
that exposure to TCA poses no risk of fetal effects. An unpublished study,
which has not undergone scientific peer review, reports fetotoxic effects
(cardiac malformations) in rat pups exposed ir± utero to TCA (Dapson et al.,
1984). In order to assess the importance to Santa Clara Valley residents of
further research on this issue, the IEMP uses the Dapson study to examine the
possible impact of TCA under the alternative assumption that exposures above an
estimated threshold of 16 ug/m3 could pose the risk of fetal effects. THE
SENSITIVITY RESULTS SHOULD NOT BE INTERPRETED AS INDICATING WHETHER OR NOT A RISK
IN FACT EXISTS; EPA RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK MANAGEMENT
DECISION-MAKING OR REGULATORY ACTION. Under this alternative assumption, the
estimated exposure of 44 ug/m^ to the most exposed individual (MEI) living
directly downwind of the modeled point source exceeds the estimated threshold.
This finding suggests that more research is appropriate, both on the actual
levels of exposure to the MEI and on TCA's potential adverse effects. The
National Toxicology Program has commissioned a project to repeat the limited
Dapson study; results are expected in Fall 1986.
-------
Table 3-40
Individual Lifetime Health Risks in Santa Clara Valley (by Pollutant)
Maximum Exposed Individual (MEI) vs Average Individual
Substance/Level
of Evidence (1)
Source Modeled
for MEI Estimate(2)
Most Exposed Individual
Lifetime Risk of Cancer
(Chances in a Million)
Average Individual
Lifetime Risk of Cancer
(Chances in a Million)
Potential Non-
Cancer Health
Effects for MEI(3)
Benzene/A ( 4 )
Benzene/A ( 4 )
Chloroform/B2
EDB/B2 (5)
Ethylene Oxide/Bl
Perchloroethylene/B2
Methylene Chloride/B2
Tr ichloroethylene/B2
Traffic Intersection
Computer Equipment Mfg.
POTW
Traffic Intersection
Hospital
Dryc leaner
Industrial Facility
Semiconductor Facility
200
40
2
6
200
10
30
20
20
20
0.06
0.2
2
2
0.6
0.1
blood & fetal
blood & fetal
u>
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS, THESE ESTIMATES OF INDIVIDUAL
RISK ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE
AND POIENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. SEE TEXT.
1 See table 3-36 for a definition of level of evidence for carcinogenicity
2 The risks presented are not the total for the source, rather they are risks by pollutant. See text for risk
by source.
3 Reported only if exposure exceeds threshold.
4 The point estimate for average individual risk incorporates important assumptions. See text.
5 Risks are based on a cancer potency value derived from ingestion studies. Inhalation value may be lower.
-------
3-113
8. Santa Clara Valley's ambient air appears to contain the toxic pollutants
common to most urban air, at levels, in most cases, slightly below the levels
in other industrialized areas studied by EPA.
Table 3-41 presents annual average concentrations of selected air toxics
monitored or measured in Santa Clara Valley and levels monitored in other cities.
The values from the other cities are based on long-term ambient monitoring and
thus are not directly comparable to some of the values from Santa Clara Valley,
which are based on modeling or short-term monitoring. For the most part,
concentrations of air toxics modeled or monitored in Santa Clara Valley appeal-
to be similar to or lower than pollutant concentrations typical in other urban
areas.
Estimated average concentration levels for most chemicals examined in
Santa Clara Valley were below 5 micrograms per cubic meter (ug/m^). Monitoring
data for toxic metals indicate that they are present in Santa Clara Valley's
outdoor air at low levelsin some cases, the lower end of the range is below
the detection limit for the analytic equipment used. Dispersion modeling of
toxic organic substances indicates that pollutant concentrations are generally
highest in the northern part of the study area, which is more industrialized
and more heavily populated, and lower in the mountains to the south and west
of the Valley.
One reason that the toxic air pollutants we studied are present at
relatively low levels may DP that existing regulations designed to reduce
emissions of criteria pollutants reduce emissions of toxic air contaminants
as well. As cited in this chapter, other studies suggest that such indirect
control may be significant for toxic compounds. The ACMD has compiled infor-
mation indicating that implementation of its rules to reduce organic compounds
that are ozone precursors has reduced emissions of a number of toxic organic
gases. The AQMD is compiling similar data on the reduction of emissions of
toxic metals that has been achieved as a result of controls on total suspended
particulates.
9. Among the key research needs identified in the air analysis is the
need for monitoring data for toxic air contaminants in Santa Clara Valley.
In the course of conducting the risk assessment for exposures to airborne
toxic pollutants, the IEMP identified a number of data gaps, many of which have
been discussed elsewhere. One such data gap is the lack of monitoring informa-
tion for toxic air pollutants in Santa Clara Valley.
No monitoring data for the organic gases of interest were available when
the IEMP began in early 1984. While the project did collect a one week data
set fron both stationary monitors and a mobile sampling unit, these data were
collected over such a short period that we generally could not use them with
confidence as quantitative indicators of annual average ambient conditions.
For organic particulates, no site-specific ambient monitoring data were avail-
able. For metals, ambient monitoring data for the study area were available
fron only one monitoring station run by the AQMD.
-------
Table 3-41
Comparison of Ambient Levels of Selected Air Toxics
in Santa Clara Valley with Levels in Other Cities
All values are in micrograms/cubic meter
Except where noted, all values are annual averages based on ambient monitoring
Substance
Beryllium
Cadmium
Chromium
Nickel
Benzene
Carbon
Tetrachloride
Perchloro-
ethylene
1,1, 1-TCA
Santa Clara
County
0.00003
0.001
0.0126
0.0058
0.2
0.2
0.4
0.1
to 0.00014
to 0.003
to 0.0138
to 0.009
to 2.6 (1)(2)
to 1.2 (3)
to 4.0 (1)
to 2.8 (1)
Elizabeth Philadelphia Downtown Baltimore
New Jersey Naval Hospital Los Angeles (Guilford)
0.00003
0.007
0.013
0.02
22.3
0.2
2.9
0.4
0.0004
0.004
0.06
0.023
10.4
4.0
5.4
28.8
0.0002
0.004
0.017
0.019
15.2
0.2
9.0
6.8
0.00007
0.001
0.093
0.018
13.9
8.5
4.1
15.7
1 These ranges are based on dispersion modeling, not ambient monitoring; as such, one must exercise
caution in comparing these results with the monitoring results from the other cities since different
methodologies were used.
2 Dispersion model may underestimate ambient concentrations for benzene. See text.
3 This range is based on short term monitoring; as such, one must exercise caution in comparing these
results with the monitoring results from the other cities since different methodologies were used.
Source: Bill Hunt, Bob Faoro, Tom Cucran and Jena Muntz, "Estimated Cancer Incidence Rates for
Selected Toxic Air Pollutants Using Ambient Air Pollution Data," July 1984.
-------
3-115
Both the AQMD and the ARE have recently embarked on monitoring programs
for many of the organic gases included in this analysis. The ARB's program,
which began in mid-1985, includes 22 monitoring stations statewide, with five
in the Bay Area including one in San Jose. The substances for which the ARE
is monitoring include 1,1,1-trichloroethane, trichloroethylene, carbon tetra-
chloride, perchloroethylene, chloroform, and benzene. The AQMD is augmenting
the ARB project by monitoring for organic gases at stations within its juris-
diction. As part of its recently adopted air toxics program, the AQMD is
expanding its air toxics monitoring network to fifteen stations. At least one
station (in Mountain View) is in Santa Clara County. Another station (Fremont)
is being used by both the AQMD and the ARB for quality control and quality
assurance reasons.
As part of the Stage II analysis, the IEMP plans to sponsor monitoring and
analysis of organic particulates in Santa Clara Valley. Such information will
help fill a significant data gap and better define the potential risks posed by
organic particulates.
These monitoring activities will help verify some of the tentative findings
presented in this report and will be useful in ongoing efforts to assess and
manage risks from airborne toxics in Santa Clara Valley.
-------
3-116
Key Assumptions and Analytic Decisions in the Air Analysis
For the convenience of the reader, we have prepared this section high-
lighting some of the key assumptions and inputs in the air analysis. All of
these assumptions and other relevant aspects of the analysis are discussed in
greater detail in the chapter. For a brief statement of the limitations in
data and scope of the air analysis, see point no. 1 in the summary and
conclusions section.
Toxicology
(For general assumptions regarding toxicology, see chapter 2. Only assumptions
unique to the air analysis are outlined here.)
1. We assume that people are exposed continuously over their lifetimes (assumed
to be 70 years) to the levels of pollutants modeled or monitored in outdoor
air. However, people spend a large portion of their time indoors.
2. We assume that the typical individual weighs 70 kilograms (or roughly 154
pounds) and breathes 20 cubic meters of air each day.
3. For some compounds, there appear to be insufficient data at this time for
an adequate evaluation of their potential carcinogenicity. For others, no
data were located in the available literature regarding their carcinogenic
potential. For these reasons, a number of compounds (as noted in table 3-3)
are not analyzed as carcinogens in the report.
4. While there is limited evidence that some lead compounds induce tumors in
experimental animals, CAG has not yet made a determination as to its potential
carcinogenicity. As a result, we do not analyze lead as a carcinogen in the
analysis.
There is strong scientific evidence that airborne lead can have adverse
health effects, particularly for children. However, because of the issue's
complexity and the lack of an estimated no-effect threshold, we have not
estimated health effects from lead exposure in this report.
5. There is considerable debate over whether TCA is a carcinogen. EPA's
current policy is that TCA should not be considered a possible carcinogen, and
we follow that policy as our base case. EPA formerly considered TCA a possible
carcinogen, but has suspended that classification because a key implicating
study is currently under review.
Because of the uncertainty on this issue, and pending the outcome of the
review, we have performed a sensitivity analysis of the possible impact of TCA
if it were a carcinogen. The value of this analysis is that it can indicate
the importance of further research on this issue in terms of local risk analysis.
These estimates, which appear in footnotes to the text and tables, should be
regarded as extreme upper-bound estimates. They are not part of our base-case
analysis.
-------
3-117
6. The IEMP conducted sensitivity analysis on TCA for fetal effects. See
footnote to the text in this chapter's methodology section.
7. There is uncertainty over the carcinogenicity of gasoline vapors. Much of
the evidence of carcinogenicity comes from a two-year animal study sponsored by
the American Petroleum Institute (API). Although the study was well conducted,
its relevance to human risk assessment is uncertain. One issue is that the
animal models used in the study may be highly susceptible to the observed
effects, thereby diminishing the relevance of the findings to humans. Another
uncertainty centers on the question of vapor content. In the API study, the
gasoline was wholly vaporized for animal exposure. However, this mixture is
not representative of the evaporative mix found in ambient situations and may
overestimate the toxic potential of gasoline vapors by a factor of three to five.
For our analysis, we follow EPA policy and assume that gasoline vapors are
possible human carcinogens. However, based on the findings of other studies
and the professional judgement of EPA's Office of Air Quality Planning and
Standards, we divide our risk estimates by a factor of four to account for the
lesser volatility of some gasoline constituents suspected of toxic effects.
8. Gasoline vapors are composed of a mix of compounds of varying molecular
weight. In order to estimate cancer risk from exposures to gasoline vapors
measured in micrograms per cubic meter, it is necessary to convert the CAG
potency value from parts per million to micrograms per cubic meter. For purposes
of this conversion, in keeping with the professional judgement of EPA's Office
of Air Quality Planning and Standards, we assume that the average molecular
weight of gasoline vapors is 110.
9. In keeping with other EPA analyses, we assume that we can use CAG's potency
value for gasoline vapors in estimating cancer risks to individuals using self-
service gasoline pumps. However, human exposure to gasoline vapors when filling
fuel tanks is typically brief and sporadic whereas the exposure experiments
that form the basis of .CAG's potency estimate have generally been sub-chronic or
chronic.
Metals
10. For screening purposes, we assume that we can apply the metals monitoring
data collected from the Fourth Street monitoring station in San Jose to the
entire study area. Such an assumption appears reasonable given what is known
about sources and TSP levels in Santa Clara Valley. In doing so, however, we
may overestimate individual lifetime risks from exposures to metals for some
County residents, and underestimate such risk for other residents. It is not
clear whether this assumption results in overestimation or underestimation of
annual increased incidence due to exposures to metals.
11. When the metals concentrations were below the detection limit for the
analytical techniques used on samples of ambient air, and there was reason to
believe the metal is present in Santa Clara Valley's ambient air (for instance,
if it had been detected in other samples), EPA used the convention of assigning
a value of one-half the detection limit as the metal concentration.
12. Our characterization of the sources of metals and organic particulates in
-------
3-118
Santa Clara County's ambient air is based on rough calculations of emissions
from some, but not all, possible sources. For the area sources (cars and
residential heating), the emissions estimates are based on information in the
national literature scaled to reflect local conditions, but have not been
confirmed by source testing. For the point sourcethe coal burning cement
plantthe estimates are based on preliminary source testing.
These emissions estimates form the basis for our apportionment of risk
from metals and organic particulates by source. Under one scenario we assume
that the estimated risks can be apportioned to the sources examined in direct
proportion to their estimated emissions. In an alternative scenario, we assume
that estimated risks from these compounds can only be apportioned to area
sources; this second scenario excludes the coal burning cement plant on the
assumption that its emissions may have only a localized impact and not result
in measurable exposures to Valley residents.
These apportionment estimates have not been confirmed by modeling of the
emissions estimates and are preliminary and uncertain. Because we have not
examined all possible sources of metals and organic particulates, and instead
focused only on those we thought might be the most significant, we may be
overestimating the contribution of these sources to the estimated risk. This
preliminary analysis is designed to guide further research rather than provide
definitive statements regarding the sources of risk from metals and organic
particulates.
13. There is uncertainty over the valence state of the chromium monitored in
Santa Clara County's air (as elsewhere). The hexavalent form of chromium is
considered a carcinogen whereas non-hexavalent chromium is not considered a
carcinogen on the basis of human or animal studies. The magnitude of the
health risk from chromium clearly depends on the assumption one makes about the
proportion of chromium that is hexavalent. Based on information from other EPA
analyses, we assume for our base case that 10% of the chromium in Santa Clara
Valley's anbient air is hexavalent.
14. Nickel poses the same difficulties as does chromium. Only two relatively
race subspecies of nickel (nickel subsulfide and nickel carbonyl) are considered
carcinogenic. However, there is uncertainty over the form of nickel present in
Santa Clara County. As such, we used the potency value for nickel from refinery
dust, rather than the higher potency value for pure nickel subsulfide. The
assumption is that the mix found in the refinery dust may be more representative
of the mix of nickel found in ambient air. In addition, we assumed that the
lower end of our range for risk of cancer from nickel exposures was zero to
account for the fact that perhaps none of the nickel in Santa Clara Valley's
ambient air is carcinogenic.
Organic Gases
15. The results of the dispersion modeling for organic gases are dependent upon
the emissions inventories provided by the AQMD. The inventories for area
sources, were not, for the most part, developed for the purposes of this analysis,
For some organic gases, the model may underestimate ambient concentrations,
either because the inventories are not complete or because the model does not
account for background levels. For point sources, the estimates are from a
-------
3-119
Toxic Use Survey conducted by the AQMD in 1984 and are thus considered more
reliable for purposes of this analysis.
16. The emissions estimates we used are based on conditions at the time the
analysis began. As a result, they do not take into account recently enacted
AQMD regulations such as Regulation 8 Rule 30 regarding the emissions of organic
gases from semiconductor manufacturing facilities. The AQMD estimates that
the regulation may reduce emissions of chlorinated solvents from semiconductor
facilities by 20% to 40%. In addition, by encouraging positive photoresist
over negative photoresist in facilities emitting at least 15 pounds per day of
precursor organic compounds, it may reduce xylene emissions bv 90%. Negative
photoresist typically uses xylene formulated resin and developer solutions
whereas positive photoresist typically uses glycol ethers for primer and resin
carrier.
17. Similarly, the emissions estimates do not take into account possible changes
in benzene emissions. The ARE estimates that benzene emissions from evaporation
and combustion of gasoline may increase by 10% in the state by the year 2000
due to the phase-out of lead. On the other hand, emissions may be reduced
under a plan outlined by the ARB staff in a report released in May. The report
outlines specific control measures that, if adopted and fully implemented, may
result in a 50% reduction in benzene emissions statewide by the year 2000 and a
decrease in statewide cancer risk of 39% between 1984 and 2000.
18. The Philadelphia IEMD study suggests that the dispersion model we used
may be underestimating ambient levels of benzene by a factor of two or three.
The short-term monitoring conducted by the IEMP in Santa Clara Valley found
benzene levels two times higher than the model predicted. In addition, more
recent monitoring by the ARB between March and December of 1985 found average
levels as much as four times higher than the model predicted. Neither of
these monitored estimates are annual averages as the model estimates are;
however, they do suggest that the model may be underestimating ambient benzene
levels. Although we note the possibility of risks being up to four times as
high as modeled, for our base-case we use the modeling results.
19. Appendix 3-B discusses the assumptions regarding the air dispersion model
itself (source information, transport data, receptor data).
20. EPA supplemented the AQMD's emissions estimates of organic gases by estimating
emissions from three sources not included in the AQMD's inventories: publicly
owned treatment plants, groundwater aeration facilities, and sanitary landfills.
Because emissions estimates from these three sources were so low relative to
the emissions estimates from other sources in the AQMD's inventories, we did
not model them for average ambient exposures as we did the AQMD sources.
21. The AQMD did not provide emissions estimates for EDB, so we could not use
the model directly as we did for other organic gases. Instead, we estimated
the ratio of benzene to EDB in the exhaust of light duty vehicles, using national
figures on EDB emissions. We used this ratio to estimate a range of EDB levels
and aggregate exposures to EDB from exhaust. The assumption is that the ratio
of EDB to benzene in exhaust is about the same in Santa Clara County as it is
nationally.
22. For carbon tetrachloride, \*e used the results from the short-term monitoring
-------
3-120
to estimate exposures rather than the results from the modeling of inventoried
sources. Although in general we do not regard the short-term monitoring data
as a sufficient basis for exposure estimates, we believe that in the case of
carbon tetrachloride it is a better indicator than our model of the actual
levels present in Santa Clara Valley air. One reason is that studies suggest
that there may be a global background level of carbon tetrachloride of roughly
0.7 ug/m^ which would not be captured in modeling of inventoried sources.
Organic Particulates
23. For our analysis of organic particulates (which focused on PAHs), we have
neither site-specific ambient monitoring data nor site-specific estimates of
emissions. Although the data are extremely sparse, we included organic particu-
lates in the analysis because a nationwide study on air toxics conducted by EPA
had shown that they could be significant contributors to airborne risk and
we did not want to ignore this possible source altogether. However, there is
considerable .uncertainty in the estimated PAH concentrations; results should be
interpreted in this light.
24. For the first method of analyzing exposures to organic particulates, we
assume that emissions of polycyclic aromatic hydrocarbons (PAHs) are related to
emissions of benzo(a)pyrene (BaP is a chemical constituent of PAH); that national
emission rates for PAH and BaP can be scaled appropriately to Santa Clara
County; and that data from other cities on BaP concentrations can be applied to
Santa Clara County.
25. For the second method of analyzing exposures to organic particulates, we
assume that we can correlate ratios of total suspended particulates to PAHs £rom
other cities to Santa Clara County.
26. In analyzing the health risks associated with exposure to PAHs, we considered
the three major groups of chemical components individually. We focused the
analysis on the BaP group of compounds since there is no known evidence of
carcinogenicity for naphthalene and relatively weak evidence for the anthracene
group in comparison with the BaP group. Sufficient data on potency do not
exist to characterize the health risks associated with exposure to the BaP
group on a compound by compound basis. Thus, to facilitate our screening
analysis, we applied the potency value for BaP, the only chemical for which we
have sufficient toxicological data, to the entire BaP group of compounds. It
is not clear whether this approach overestimates or underestimates risks from
PAHs since the analysis excludes chemicals in the anthracene group (which may be
carcinogenic), but may apply a higher potency value than is appropriate to
chemicals other than BaP in the BaP group.
-------
3-121
References for the Air Analysis
* Appendix 3-A: Annual Summaries of Metals Monitoring Data.
* Appendix 3-B: Description of IEMP Air Modeling in Santa Clara Valley.
* Association of Bay Area Governments (ABAC). "Air Emissions Associated with
Publicly Owned Treatment Works in Santa Clara Valley." May, 1985. 16 pp.
* Association of Bay Area Governments (ABAC). "Air Emissions Associated with Pumping
of Contaminated Groundwater in Santa Clara Valley." February, 1985. 50 pp.
Bay Area Air Quality Management District (BAAQMD). Memorandum to Chairperson
Silver and Members of the Board of Directors of the BAAQMD from Milton Feldstein,
Air Pollution Control Officer. "Potential Toxic Air Contaminant Emissions
from Kaiser Permanente Cement Company." March 26, 1986.
California State Air Resources Board, Staff Report. "Proposed Benzene Control
Plan." May 1986.
* Clark, Mae and Gary Silverman. "Evaluation of Air Emissions, Runoff to Surface
Water, and Leachate to Groundwater from Sanitary Landfills." Association of Bay
Area Governments (ABAC). July, 1985. 250 pp.
Health Effects Institute. "Gasoline Vapor Exposure and Human Cancer: Evaluation
of Existing Scientific Information and Recommendations for Future Research."
September 1985.
Memorandum from Bill DeWees (ENTROPY Task Manager) to Dennis Holzschuh (EPA
Task Manager). "Analytical Results of Hexavalent Chromium and Total Chromium
for Samples Submitted by EMB." May 29, 1985.
Office of Policy Analysis, "The Air Toxics Problem in the United States: An
Analysis of Cancer Risk for Selected Pollutants," USEPA, May 1985.
Simmonds, P.G. et al. "The Atmospheric Lifetime Experiment: Results for Carbon
Tetrachloride Based on Three Years Data." Journal of Geophysical Research.
October 20, 1983.
* TRC Advanced Analytics, Inc. "Air Quality Study of Santa Clara Valley,
California: October 8, 1984 - October 12, 1984." TRC Project 7016-Z88. 26 pp.
Data Appendix, 330 pp.
United States Environmental Protection Agency. "Air Quality Criteria for Lead."
External Review Draft, EPA-600/8-83-028B. 1984.
United States Environmental Protection Agency- "Evaluation of Air Pollution
Regulatory Strategies for Gasoline Marketing Industry." July 1984.
* Denotes documents prepared in support of the Santa Clara Valley IEMP
and available through the IEMP office.
-------
3-122
References for the Air Analysis (cont.)
* Versar Inc. "Analysis of Ambient Air Samples." May 1, 1985.
* Versar Inc. "Emissions and Exposures to Gasoline Vapors in Santa Clara County,
California." May 1986.
* Versar Inc. "Follow-up Air Quality Analysis in Support of the Integrated
Environmental Management Division's Santa Clara Project." June 12, 1985. 160 pp.
* Versar Inc. Memorandum to Eileen Soffer (IEMP) from Dennis Hlinka (Versar Inc.).
"Updated Santa Clara Valley Study Report." May 12, 1986.
* Versar Inc. Memorandum to IEMP from Versar Inc. "Technical Background and
Estimation Methods for Assessing Air Releases from Sewage Treatment Plants."
October 11, 1984. 57 pp.
* Denotes documents prepared in support of the Santa Clara Valley IEMP and
available through the IEMP office.
-------
CHAPTER FOUR
ANALYSIS OF RISKS FROM TOXICS IN DRINKING WATER
-------
CHAPTER FOUR
ANALYSIS OF RISKS FROM TOXICS IN DRINKING WATER
I Introduction 4-1
A Drinking Water Sources and Distribution System 4-2
II Exposure and Risk: Surface Water Contamination 4-8
A Contaminants in Surface Drinking Water 4-8
1. Disinfection Agents and By-products 4-8
2. Metals and Minerals 4-9
3. Pesticides 4-9
B Toxicology Evidence 4-10
C Exposure Methodology and Risk Calculations 4-15
1. Disinfection 'Agents and By-products 4-15
2. Metals and Minerals 4-24
3. Pesticides 4-29
D Summary: Surface Water Contamination 4-31
III Exposure and Risk: Groundwater Contamination 4-32
A Groundwater Contamination Issues 4-33
B Health Effects Information 4-37
C Hydrogeology 4-45
D Risks from Current Exposure to Groundwater Contamination 4-48
F IEMP Methodology for Estimating Future Exposure to
Groundwater Contamination from Underground Tanks 4-63
G Future Risks from Groundwater Contamination: Results and
Conclusions 4-78
IV References 4-104
-------
CHAPTER FOUR
ANALYSIS OF RISKS FROM TOXICS IN DRINKING WATER
This chapter describes the methodology and results of the Santa Clara Valley
ICMP's Stage I drinking water risk analysis. Vfe estimated risk by combining
estimates of individual and aggregate exposure to various chemicals with estimates
of the toxicological potency of those chemicals. We designed our drinking water
analysis to allow us to compare risks in Santa Clara Valley trom exposure to
toxics in drinking water with those from exposure to toxics in ambient air, which
we discuss in Chapter 3, and with other environmental exposures.
Most of this chapter addresses our exposure analyses, which we based on
local monitoring data and projections of future contamination* and exposure. We
also briefly discuss our toxicological analyses; these analyses are based on
analysis of laboratory experiments with animals and human epidemiologic data.
It is important to remember the uncertainties in both types of analyses when
considering the results.
We examined exposure to and risks from disinfection agents and by-products
in surface water, including chloramines and trihalomethanes. We analyzed exposures
and risks for pesticides from runoff to surface water, and from leachate to
groundwater. We also estimated exposure and risk for metals and minerals in
drinking water. These are primarily naturally-occurring substances in both
surface and groundwater.
We analyzed risks for the following groundwater contamination problems:
1) organic chemicals from spills, leaking underground tanks and pipes,
and illegal disposal;
2) fuel constituents from spills and leaking underground tanks and pipes;
3) organic chemicals and from other sources, including sanitary
landfills, sewer line leakage, septic tanks, and runoff to dry wells; and
4) nitrates from wastes, fertilizers, and septic tanks.
The drinking water analysis is divided into two parts: risks from imported
surface water, and risks from local groundwater. Because surface water and
groundwater sources have different contamination problems and sources of data,
our approach to estimating risks is different for each. When limitations in the
available data were significant, particularly for parts of our groundwater exposure
analysis, we have made extensive use of assumptions, professional judgement, and
analytic models. In addition, we have only estimated the risks from current and
future exposure to toxics in drinking water, and have not estimated risks from
past exposure. Our methodology for estimating exposure to drinking water pollutants
is illustrated in Figure 4-1.
We estimated drinking water risks from groundwater contamination for six hydro-
geologic "zones" in Santa Clara Valley because they differ in their vulnerability
to contamination and the distribution of contamination sources. This approach
differs from the estimation of surface water risks, in which we break down risks
by water treatment plant and water service district.
* NOTE: The term "contamination" has a specific meaning under California law.
Vfe use a more general interpretation of the term throughout this report.
4-1
-------
4-2
Our analysis provides conservative estimates of the exposure and risk
toxics in drinking water (i.e., we are more likely to overestimate than under-
estimate risks). We include estimates of the average level of risk to exposed
individuals, of the risk to the most exposed individual, and the expected
increased incidence of disease. In general, we present ranges of exposure
estimates, reflecting alternative plausible assumptions which are not necessarily
upperbound assumptions. Our toxicological assumptions, however, are plausible
upper bound assumptions, with uncertainties noted and analysis of alternatives
in a few cases where evidence is most equivocal.
We have tried to be as comprehensive as possible in estimating drinking
water risks, but we can not be sure that we have identified all of the significant
drinking water problems. Our results should be applied only within the context
of the IEMP. They are intended only for screening purposes, and are not designed
to produce unequivocal statements on the absolute risks from various sources
or pollutants. We intend our risk estimates to be used for evaluation and
comparison of classes of problems ~ sources, exposure pathways, pollutant
types rather than for characterization of specific problems, such as particular
plumes of groundwater contamination.
This chapter first presents background information on Santa Clara's drinking
water sources and distribution patterns. It then discusses exposure to and risk
from surface water, including water imported from outside Santa Clara and .local
surface water supplies. We then discuss exposure to and risk from groundwater
contamination, including risks from currently contaminated wells, and risks from
current and future contamination from various sources.
Drinking Water Sources and Distribution System
Drinking water in the Santa Clara Valley comes from three sources: ground-
water; surface water imported through the South Bay and Hetch Hetchy Aqueducts;
and impounded local surface water. Overall, about half of the drinking water
in the Santa Clara Valley is groundwater, and half is surface water. Large
volumes of imported and local surface water are used to recharge the groundwater
basin artificially, both to prevent the depletion of the aquifer and to store
water supplies at a relatively low cost. Figure 4-2 shows the drinking water
supply patterns for the Santa Clara Valley.
Residents in different parts of the Valley are served by different water
sources, or combinations of sources. In most of the Valley groundwater provides
at least part of the supplies. Some areas rely entirely on treated surface
water, however; others use only groundwater. Nineteen water retailers
eleven municipal water agencies and eight large private water companies
deliver water to the Santa Clara Valley's consumers, under regulation by the
California Department of Health Services. Figure 4-3 shows the systems for
conveying, treating, and distributing water to these retailers, and Figure 4-4
shows the areas they serve. In addition, about 200 small public systems and
numerous private wells also provide groundwater for drinking.
The City and County of San Francisco operate the Hetch Hetchy Aqueduqt,
which brings water from the Sierra Nevada to parts of the Bay Area. Hetch Hetchy
water supplies about one-fifth of the Valley's drinking water. It is treated
at four facilities between Yosemite and the Santa Clara Valley.
-------
4-3
The Santa Clara Valley Water District (SCVWD) operates two treatment plants
for the water imported from the Sacramento River Delta through the South Bay
Aqueduct, and sells the treated water to the water retailers. About one-fourth of
the Valley's water supply comes directly from the South Bay Aqueduct. The SCVWD
has been increasing its ability to serve treated surface water to most parts of
the Valley. The SCVWD is also responsible for managing most of the Valley's
groundwater resources, including recharge of Santa Clara Valley's aquifers
with local and imported surface water.
Most of the water retailers draw groundwater from the aquifers; the nineteen
major retailers operate about 300 wells. The state Department of Health Services
(DOHS) regulates these retailers. More than 200 additional small water systems
(those serving 199 or fewer connections; most have about 10 connections) also
draw on the aquifers. They are regulated by the Santa Clara County Public
Health Department. Numerous private wells also exist; these receive little
formal regulatory scrutiny. In the northern part of the Valley, the small
systems and private wells rely on about 1,200 wells. There are also about
4,000 wells in South County, an unknown number of which are in operation.
Together, these 5,200 wells supply about 7 percent of the area's total drinking
water supplies. Altogether, groundwater supplies about half of the Valley's
drinking water.
Generally, the SCVWD assures that the Valley has an adequate water supply,
while DOHS and the County Health Department are responsible for assuring that
the drinking water is of good quality and meets all federal and state standards.
The water retailers cooperate closely with these agencies to provide most
Valley residents drinking water. Those people using private wells are not
subject to the same degree of regulation as those using small or large public
wells.
-------
F,GURE4-1 METHODOLOGY FOR ESTIMATING
EXPOSURE TO TOXIC DRINKING
WATER POLLUTANTS
4-4
mm
SURFACE
WATER
Monitoring of:
Metals
Minerals
Pesticides
GROUND
WATER
Monitoring
Trihalomethanes
and Chloramines
Industrial
Contaminants
Present: Monitoring
Future: Model
V
]
:::::::::.::::::..:
-------
4-5
FIGURE 4-2 DRINKING WATER SUPPLY PATTERNS
1982-1983 (VALUES IN THOUSAND ACRE FEET)
HETCH
HETCHY
RESERVOIRS
& STREAMS
SOUTH BAY
AQUEDUCT
WATER
TREATMENT
PLANTS
DRINKING
WATER
WATER
CONTROL
RESERVOIRS
140
RAINFALL TO
GROUNDWATER
GROUNDWATER BASIN
-------
FIGURE 4-3
WATER CONVEYANCE, TREATMENT
AND DISTRIBUTION SYSTEM
I
CTi
-------
4-7
t Oly Ol Palo A'tG iCily Owned!
Water supply Hetch Hetchy
2 Gi( oi Mountain View (Cily OAned)
Water supply Helch Helchy Groundwaler
3 Cily of Sunnyvale (City owned)
Waif supply Groundwater Soulh Bay Aqueduct
Helen Hetchy
4 Punss'^a Hills Wale' District (Special D'Stncti
Wale' supply Hetch Helchy
5 Ca'i'ornra Wale1 Service Company (Investor o/.ned)
Water supply Groundwater South Bay Aqueduct
Hetch Hetchy
6 City ol Cupertino (City owned)
Wafer supply South Bay Aqueduct Grouodwater
7 City ol Santa Clara (Cily owned)
Water supply Groundwaler Soulh Bay Aqueduct
Helch Helchy
8 City ol San Jose (City owned)
Water supply Hetch Hetchy. G'oundwdler
9 City ol Miipiias [City owned)
Water supply Helch Helchy
10 San Jose Water Company (Investor owned)
Water supply Groundwaler Soulh Bay Aqueduct
SJWC operated Irealment plant
11 Cily ol San Jose (City owned) \
Wale' supply Soulh Bay Aqueduct Gioundwaler
12 Great OaKs Water Company (Pnvalely owned)
Water supply Groundwaler
13 City ol Morgan Hill (City owned)
Wale' supply Groundwaler
14 Wesl San Martin Water Works iPnvalely owned)
Water supply Groundwater
15 City Ol Gilroy iCily Owned)
Wale' supply Groundwaler
16 Santa Clara Valley Water District (Special Oistncti
Wa/e' supply Groundwater Soulh Bay Aqueduct
(The District serves I he entire Sanla Clara
Couniy with water supply and Hood control pro
grams II is a wholesale water supplier to various
cily owned and investor owned utilities It
operates two Irealment plants, eight dams and
also conducts programs in groundwater
recharge wale' reclamation, importation
distribution and weather modification )
A Magic Sands Mobile Home Park (Pnvalely owned)
Wale/ supply Groundwaler
Rancho Santa Teresa Mobile Home Park
(Privately owned) Water supply Groundwaler
Carnbbe Mobile Home Park (Pnvalely owned)
Wale' supply Groundwater
Redwood Mutual Water Company (Pnvalely owned) i
Water supply Groundwater
FIGURE 4-4
AREAS SERVED BY
MAJOR WATER PURVEYORS
-------
4-8
EXPOSURE AND RISK: IMPORTED SURFACE WATER
In this section, we discuss the methods used to estimate exposures to toxic
contaminants in imported surface water supplies of drinking water. Vfe also discuss
the possible health effects of the contaminants studied. We combined estimates of
human exposure to chemicals with information on contaminant toxicity to estimate
individual risks and aggregate incidence of disease.
Generally, we estimate current exposure by assuming that people are exposed
throughout their lifetime to the contaminant levels indicated by recent monitoring
data, if available. In some cases, relatively good data are available over
time; in others, the data are quite sparse. Vfe believe that current contamination
levels are generally a good indicator of future contamination levels, except
where changes may occur in supply patterns. We have not accounted for the
impact that the San Felipe supplies will have since we do not yet know how
this water will be distributed in the Santa Clara Valley, nor what the levels
of contaminants will be.
We analyze exposure to disinfection agents and by-products in surface
water. We also present, in this section, our analysis of minerals and metals
and pesticides in drinking water; minerals and metals and pesticides are present
in both surface and groundwater, and are included here because the methodology
for estimating risks is similar to that used for other surface water contaminants.
Vfe decided to study these pollutants through a screening process which e'xamined
evidence of local exposure and evidence of chronic toxicity; only where we had
positive indications of both did we attempt to develop quantitative estimates
of exposure and risk (see Appendix).
We break down our estimates of exposure to surface water contaminants in
different ways. For disinfection agents and by-products (trihalomethanes and
chloramines), we generally estimate exposure by the source of the water, and by
the plant at which the water is treated. Vfe also indicate which areas are
served by each source of surface water. For pesticides, we examined recent
groundwater monitoring data and historical data on surface water imported
through the Hetch-Hetchy and South Bay Aqueducts. For minerals and metals, we
estimate exposure for each water service district. Vfe have not attempted to
describe surface water exposure for smaller geographic units because variations
in the mix of sources and the widespread blending of many sources make such
detailed analysis extremely difficult, if not impossible.
Contaminants in Surface Drinking Water
Disinfection Agents and By-products
Disinfection is a process in which water is treated with chemical agents,
most commonly chlorine, to kill microbial pathogens (disease-causing agents).
Without disinfection, people would be exposed to numerous infectious agents
through their drinking water and periodic outbreaks of diseases such as cholera
or typhoid could result. Disinfection techniques comnonly leave a residual
disinfecting agent (chlorine if the disinfection process is chlorination) in
the treated water, so as to continue disinfecting water until it is used.
Unfortunately, potentially harmful byproducts- trihalomethanes- result from the
reaction of chlorine with organic matter (such as plant and animal matter) in the
water.
-------
4-9
We have analyzed the potential toxic effects of chlorination and chloram-
ination, the two disinfection techniques used to treat Santa Clara Valley
drinking water. Organic chemicals of concern created through these processes
include the trihalomethanes (chloroform, brcmoform, chlorodibromomethane, and
dichlorobromomethane), and the chloramines (monochloramine, dichloramine,
and trichloramine).
Until recently, chlorination was the only disinfection method used to treat
drinking water consumed in the Santa Clara Valley. The Santa Clara Valley Water
District has implemented an alternative treatment method, chloramination, at
the two major local treatment plants, the Rinconada and Penitencia plants.
Chloramination involves the addition of ammonia to water in combination with
some form of chlorine. The result is a lower chlorine residual, because the
chlorine is mostly bound up with ammonia in compounds known as chloramines.
Because of the lower chlorine residual, fewer THMs are formed but the
treated water contains chloramines. There is usually no free chlorine in
chloraminated drinking water. We estimate exposure to chloramines in chloraminated
water, as well as to THMs in all disinfected water. There is no evidence that
chlorine itself is toxic to humans at the levels found in treated drinking
water.
Metals and Minerals
A number of inorganic substances, primarily metals and other minerals, are
found in drinking water. Most of these substances are probably from natural back-
ground sources (i.e., they are present in soils and are picked up by water traveling
through those soils). Some of these chemicals may be the result of man-made
contamination, but, for most substances, this is difficult to determine and
specific incidents that caused the contamination cannot be identified. Metals
and minerals may be present in both surface and groundwater; we present our
exposure and risk estimates in this section because sources (surface or ground)
are not readily distinguishable and analytic methods are similar to those used
to estimate other surface water risks.
Metals and minerals for which we have estimated exposures and risks include:
o arsenic;
o barium;
o cadmium;
o chromium;
o lead;
o mercury;
o selenium;
o silver;
o zinc; and
o fluoride.
Some experts believe that asbestos may be of concern in drinking water.
Since asbestos is not monitored in the Santa Clara Valley, we did not estimate
exposure.
Pesticides
Pesticides could contaminate imported surface water primarily through runoff
-------
4-10
from agricultural areas through which the water travels; such contamination would
not necessarily have to occur in or near Santa Clara Valley to affect local water
Pesticides from local sources could cause giroundwater contamination. The
Appendix to this Chapter contains a list of pesticides monitored in surface and
groundwater. Monitoring has not detected any pesticides in large public drinking
water systems in the Santa Clara Valley.
Toxicology Evidence
Cancer effects
Tables 4-1 through 4-3 summarize the available evidence on the chronic
toxicity of those substances for which we were able to estimate exposures. Table
4-1 presents the estimates of the carcinogenic potential and potency for disinfection
agents and by-products, and metals and minerals. Evidence of the carcinogenicity
of trihalomethanes is strongest for chloroform. EPA classifies it as a probable
carcinogen, and has calculated a quantitative potency score for it. Experiments
indicate that chlorodibromomethane may also be a carcinogen, and many toxicologist
suspect that bromoform and dichlorobrromenethane may also be carcinogenic.
Since EPA's Carcinogen Assessment Group (CAG) has not developed a potency
evaluation for these chemicals, we analyzed risks from THMs as if all THMs
are as potent as chloroform.
No evidence currently exists to suggest that chloramines are carcinogenic,
although the National Toxicololgy Program (NTP) is examining this issue.
Among the metals and minerals, only arsenic has been found by EPA's CAG to
be potentially carcinogenic. (EPA considers cadmium and hexavalent chromium
carcinogenic via inhalation of air, but not via ingestion from drinking wateir.)
The evidence for arsenic's carcinogenicity by ingestion is epidemiologic evidence
correlating high drinking water arsenic levels with "Blackfoot Disease," a
form of skin cancer. However, the carcinogenicity of arsenic ingested through
drinking water is the subject of some dispute. Some scientists believe that
arsenic may not be harmful, and may even be beneficial at low levels. EPA's
Office of Drinking Water has concluded that arsenic is not harmful at low
levels, and proposed a Maximum Contaminant Level (MCL) of 50 micrograms per
liter. Blackfoot Disease is uncommon in this country, even among people drinking
from systems with arsenic levels far higher than those in Santa Clara Valley,
and is generally not life-threatening if treated. Because of the uncertainty
regarding arsenic's carcinogenicity, we have estimated arsenic's risk under
two different assumptions, detailed in the next section.
Non-cancer Effects
The estimated toxicity of water disinfection agents and by-products for
effects other than cancer is summarized in Table 4-2. Presumed no-effect
thresholds, below which exposure is deemed to be safe, are also presented.
Briefly, animal studies indicate that all trihalomethanes pose some threat of
liver effects, while some may also increase the risk of neurological and fetal
effects, and of kidney, thyroid and immune system effects. EPA and NTP are
evaluating the health effects of chloramines. Preliminary examination of the
literature indicates several possible adverse health effects. However, data
are insufficient to estimate no-effect thresholds.
-------
4-11
Substance
DISINFECTION
BY-PRODUCTS
Arsenic
Table 4-1
CANCER HEALTH EFFECTS DATA FOR CONTAMINANTS IN
SURFACE SUPPLIES OF DRINKING WATER
[All potency values are in (ug/l)~l]
Estimated
Potency
Weight of
Evidence1
Source
Reference
Chloroform
Brcmoform
Chlorod i-
bromcme thane
Dichloro-
bromome thane
Chloramines
METALS and
MINERALS
2
2
2
2
.3
.3
.3
.3
X
X
X
X
0
icr6
10~6
icr6
10~6
B2 CAG2 EPA 84
IEMD3
IEMD3
IEMD3
4.58 x 10~4
CAG'
EPA 83
Barium
Cadmium
Chramium+3
Chromium* 6
Lead
Mercury
Selenium
Silver
Zinc
Fluoride
0
0
0
c
0
0
0
0
0
0
1 The weight of evidence of carcinogenicity for the compounds listed varies
greatly, fron very limited to very substantial. According to EPA's cate-
gorization of levels of evidence of carcinogenicity, A = proven human car-
cinogen; B = probable human carcinogen; C = possible human carcinogen;
D = not classifiable; and E = no evidence.
2 Potency estimate derived by EPA's Carcinogen Assessment Group (CAG).
^ IEMD indicates a potency score estimated by IEMD toxicologists and consultants
using existing literature and following EPA procedures.
* The Carcinogen Assessment Group has calculated a potency score for arsenic
based on epidemiological evidence of "Blackfoot Disease," a form of skin
cancer. However, some scientists believe that arsenic may not be carcinogenic
and may be a necessary element at low levels.
-------
4-12
TABLE 4-2
NON-CANCER HEALTH EFFECTS FOR SURFACE SUPPLIES OF DRINKING WATER
Substance
Health Effect
Presumed Human
Threshold (ug/1) Source^
Reference
Chloroform Liver 350
Kidney 225
Fetal 795
Neurological 117
Bromoform Liver 210
Thyroid 210
Immune System 210
IEMD
IEMD
IEMD
EPA/RFDV
IEMD
IEMD
Heywood et al. 79
Heywood et al. 79
Thompson et al. 74
Challen et al. 58
EPA 85
Chu et al. 82;
Tobe et al. 82
Chu et al. 82
Munson et al. 78
Dichloro-
bromomethane
Chlorodi-
bromomethane
Fetal 794
Liver 210
Thyroid 210
Immune System 210
Liver 210
Thyroid 210
Immune System 210
Kidney 1000
IEMD Ruddick et al. 83
EPA/RFD Tobe et al. 82
IEMD Chu et al. 82
IEMD Schuller et al. 78
Munson et al. 77
Saunders et al. 77
EPA/RFD NTP 85; Tobe et al. 82
IEMD NTP 85
IEMD Munson et al. 78
IEMD NTP 85
Chloramines
Bone marrow2
Reproductive2
Body weight2
Blood2
1 "IEMD" indicates a threshold estimated by IEMD toxicologists and consultants
using existing literature and following EPA procedures. "EPA/RFD" indicates
a threshold derived from an EPA "Reference Dose" (RFD) or "Acceptable
Daily intake" (ADI) level. ADIs and RFDs are estimated no-effect thresholds
that are intended to protect an individual from the most potent non-cancer
effect. EPA has reviewed and verified internally some of these thresholds,
although they have not necessarily undergone peer review. Those EPA has
verified internally are indicated by v; those which are unverified internally
are indicated by uv. ~*
2 Tentative
-------
4-13
TABLE 4-3
NON-CANCER HEALTH EFFECTS FOR SURFACE SUPPLIES OF
DRINKING WATER: METALS AND MINERALS1
[All values are in micrograms per liter (ug/1)]
Presumed Human
Substance
Arsenic
Barium
Cadmium
Chrcmium-t-3^
Chromium+6
Mercury
Selenium
Silver
Health Effect Threshold
Liver
Neurobehavioral
Periph. Vascular
Skin
Arsenic Poisoning
Reproductive
Teratogenicity
Multiple Organ
Hypertension and
Cardiotonic effects
Kidney
Liver
Mutagenicity^
Reproductive
Fetal
Fetal 35
Liver 35
Kidney 35
Reproductive
Fetal
Fetal
Neurobehaviora 1
Selenosis
Liver
Neurobehavioral
Kidney
Repro
Fetal
Skin Discoloration
133
133
133
133
1700
133
133
133
1750
17.5
17.5
4900
874
,000
,000
,000
170
170
11
11
105
105
105
105
105
105
105
Source-^ ' J
IEMD
IEMD
IEMD
IEMD
IEMD
IEMD
IEMD
IEMD/RFDV
EPA/RFDV
EPA/RFD1
IEMD
IEMD
IEMD
EPVRFDUV
IEMDUV
IEMD
IEMD
IEMD
EPA/RFD
EPA/RFDV
EPA/RFD
IEMD
IEMD
IEMD
IEMD
IEMD
EPA/RFDV
Reference
Clertent
Clement
Env Health Crit #18
WO 31
Perry et al. , NIOSH 75
Zaldivar 77
Hood et al. 77
Hood et al. 77
Heywood & Sortel 79
Perry et al. 83
Brenniman et al. 79
Kjellstrom et al . 77
Friberg 50
Scharpf et al. 72
Schroeder & Mitchener 71
Clement
Clement
Clement
Gale 78
Gale 78
EPA 80
EPA 80
Venugopal and Lucky 79
Natl Academy Sci 76
EPA 80
Smith et al. 36
Harr et al. 67
Venugopal & Lucky 79
Schroeder & Mitchener 71
NAS 76
Schroeder et al. 71
Gaul & Staud 35
Blumberg & Carey 34
East et al. 80
-------
4-14
TABLE 4-3 (cont.)
Presumed Human
Substance Health Effect Threshold Source Reference
Zinc^
Fluoride**
Reproductive
Gastric Problems
Skeletal effects
Dental fluorosis
(cosmetic)
2590
74,900
4000
1000-4000
IEMD
IEMD
EPA/ODW
EPA/RFD5
Schlicker and Cox,
EPA 80
68
For sake of brevity, negative evidence (i.e., where a laboratory test or
epidemiological study has been performed, but no evidence was found of a
health inpact) is not reported here. More complete toxicological information
is available from IEMP.
"IEMD" indicates a threshold estimated by IEMD toxicologists and consultants
using existing literature and following EPA procedures.
"EPA/RFD" indicates a threshold derived from an EPA "Reference Dose" (RFD)
or "Acceptable Daily Intake" (ADI) level. ADIs and RFDs are estimated
no-effect thresholds that are intended to protect an individual from
the most potent non-cancer chronic health effect. EPA has reviewed and
verified some of these thresholds internally, although they have not
necessarily been peer reviewed. Those which EPA has verified internally are
indicated by v; those which are unverified internally are indicated by
uv.
Mutagenic effects cannot be expressed meaningfully in quantitative terms.
Therefore, we note positive evidence of mutagenicity by listing it (on this
table) as an effect.
zinc and Chromium +3 are considered necessary elements at low doses.
Fluoride is considered beneficial (i.e., it reduces dental caries formation)
at levels of 1000 to 2000 micrograms per liter, with ah optimal level of
about 1,000. The threshold for cosmetic effects varies with the sensitivity of
individuals.
-------
4-15
Evidence on the non-cancer toxicity of metals and minerals is summarized
in Table 4-3, along with estiinated no-effect thresholds for the non-cancer
effects. Strong scientific evidence exists that airborne lead can have adverse
health effects, especially for children (see discussion in Chapter 3). Waterborne
lead may also have adverse health effects. Evaluating effects from chronic
lead exposure is a complex issue involving evaluation of the total body burden
fron air, water and dust exposure and is the subject of current EPA research.
No EPA reference dose identifying a no-effect threshold has yet been estimated.
Because of the issue's complexity, the ongoing Federal initiatives, and the
lack of an estimated no-effect threshold, we have not estimated health effects
frcm lead exposure.
Cadmium, mercury, and selenium may pose risks of non-cancer effects at low
levels. Other substances, such as barium and zinc, are quite a bit less toxic.
At least one substance, fluoride, is considered beneficial in drinking water at
some concentrations but harmful at higher concentrations.
Pesticides in either groundwater or imported surface water may also pose
non-cancer health risks.
Exposure Methodology and Risk Calculations
Disinfection Agents and By-products
The quality of the data we used to estimate levels of disinfection agents
and byproducts in surface water varies widely. Because trihalomethanes are
regulated by EPA and the state Department of Health Services (DOHS), all public
water systems serving over 10,000 people are required to monitor for them regularly.
Thus, we have relatively good data for estimating trihalomethane concentrations-
and hence exposure- for the water provided via the Hetch-Hetchy Aqueduct. While
the same monitoring requirements apply to the water delivered via the South Bay
Aqueduct, fairly recent changes in treatment practices by the Santa Clara Valley
Water District make much of the historic monitoring data for South Bay Aqueduct
water inappropriate for our analysis. We relied on monitoring data from September
to December of 1985 to estimate THM levels in water imported through the South Bay
Aqueduct. We assumed that the levels found in water distributed from the treatment
plants were representative of the levels reaching consumers; in fact, levels may
be slightly higher as THMs may continue to be formed throughout the distribution
system. The monitoring data for THMs are poor for small public systems.
Table 4-4 shows average THM and chloramine levels and populations served
for the sources of surface water. Table 4-5 presents estimates of the levels of the
individual THMs where these data were available. Table 4-6 presents information on
the service areas and population served by the sources of water and treatment
plants. To reduce THM levels, the Santa Clara Valley Water District (SCVWD) now
chlorinates and chloraminates the water at its Rinconada and Penitenicia treatment
plants, rather 'than only chlorinating it.
The SCVWD is also examining other disinfection methods. While this study
may eventually result in different exposure levels, such changes are too speculative
to incorporate into this baseline analysis. A new source of supply, the San
Felipe aqueduct, is scheduled to begin supplying additional surface water to the
Santa Clara Valley and Gavilan Water Districts in a few years. While this water
roay have different characteristics than water currently imported, we have too
little information to project the effects of this change.
-------
4-16
TABLE 4-4
SURFACE WATER: SOURCES, POPULATIONS, AND DISINFECTION AGENT AND
BY-PRODUCT LEVELS
SOURCE
POPULATION
SERVED
TOTAL
THMS (ug/1)2
TOTAL
CHLORAMINES (ug/1)
Hetch-Hetchy
212,000
59
South Bay Aqueduct
- Rinconada 206,600
- Penitencia 136,700
80
56
1,000
1,000
City of Santa Clara1 87,700
20
<1,000
Local Surface Water
- Redwood Mutual 1,300
- SJ Water Company
- Saratoga 2,300
- Montevina 34,500
- Small Systems3 1,000
Total 682,000
57
36
47
0
0
1 The City of Santa Clara blends water from Hetch-Hetchy, the South Bay
Aqueduct, and groundwater.
2 Estimates for Hetch-Hetchy, City of Santa Clara, and Redwood Mutual are
based on data for periods between 1980 and 1984.
Estimates for South Bay Aqueduct are based on data from September to December,
1985.
Estimates for San Jose Water Company are based on the 1985 running average
of distribution system samples.
Estimates for small systems are based on average of San Jose Water Conpany
levels for local surface water.
3 Five small public residental systems disinfect local surface water.
These systems have a total of 391 connections. We estimate that these
connections serve 1,000 people.
-------
4-17
TABLE 4-5
ESTIMATES OF AVERAGE LEVELS OF INDIVIDUAL TRIHALOMETHANES
IN SURFACE WATER (ug/1)
DICHLOROBRO- CHLORODIBRO-
SQURCE CHLOROFORM MOMETHANE MOMETHANE BROMOFORM TOTAL
Hetch-Hetchy 52 5 1 1 59
South Bay Aqueduct
- Rinconada 17 26 29 8 80
- Penitencia 12 17 18 9 56
City of Santa Clara1 20
Local Surface Water1
- Redwood Mutual 4
- SJ Water Company
- Saratoga 57
- Montevina 36
- Small Systems 47
1 Data on levels of individual THMs unavailable.
-------
4-18
TABLE 4-6
SERVICE AREAS AND POPULATIONS SERVED FOR SOURCES OF SURFACE WATER
SOURCE
SERVICE AREAS
POPULATION SERVED
Hetch-Hetchy
Milpitas
Mountain View
Sunnyvale
Palo Alto
Purissima Hills
San Jose/Alviso
Stanford University
42,000
52,000
43,000
56,000
7,000
1,000
11,000
212,000
South Bay Aqueduct
Rinconada
California Water
Cupertino
San Jose Water Company
Sunnyvale
33,600
14,000
127,000
32,000
206,600
Penitencia
City of San Jose
San Jose Water Cortpany
37,300
99,400
136,700
City of Santa Clara1
87,700
Local Surface Water
Redwood Mutual
San Jose Water Company
Saratoga^
Montevina
Small Systems
.1,300
2,300
34,500
1,000
39,100
Total:
682,000
1 The City of Santa Clara blends water from Hetch-Hetchy (8%), the South Bay
Aqueduct (via Penitencia; -9%), and groundwater (33%).
2 The Saratoga Water Treatment Plant is not used full-time.
-------
4-19
We have calculated individual risks and aggregate incidence for THM exposure
by assuming that all THMs have the same carcinogenic potency as chloroform.
Structural similarities suggest that all four may eventually be found carcinogenic.
EPA has used this approach in the past in developing drinking water regulations.
Table 4-7 presents our estimates of average individual risks from THMs, and
Table 4-8 presents our estimates of aggregate increased incidence for each source
of surface water supplies.
We estimated that the average increased lifetime risk of cancer for
individuals drinking surface water ranges from about ten to 200 chances in
1,000,000, depending on the levels of THMs. For the entire study area, the IEMP
projects a possible increased incidence of slightly over one additional case
of cancer per year.
Table 4-9 compares the highest average levels of disinfection-related
contaminants found in drinking water with the estimated thresholds for toxic
effects other than cancer. Unlike cancer, exposure to a contaminant at a dose
below the estimated threshold for a given health effect is presumed to entail no
risk of that toxic effect. All trihalomethane levels appear to be below estimated
thresholds for non-cancer effects.
Measured trihalomethane exposures are comparable with those in other
areas where chlorinated surface water is drunk. Table 4-10 presents information
on THM levels in some other cities.
The impact of chloramination on the SCVWD's drinking water supplies is
difficult to measure precisely. According to the SCVWD, chloramination in
combination with chlorination at the Rinconada and Pentincia treatment plants
reduces THM levels by about 30% from levels if the SCVWD only chlorinated the
water.
It is important to note that while treatment practices are an important
factor in determining THM levels, the quality of the original water supply also
affects THM levels. Water with less decayed organic material, which are the
precursors for THMs, will generally have lower THM levels than water with a
higher proportion of such material. The importance of this factor is indicated
in the Santa Clara Valley by the fact that the THM levels in the chlorinated
Hetch-Hetchy water are comparable to the levels in the chloraminated SCVWD
water. This is due to the lower level of THM precursors in the Hetch-Hetchy
water. The SCVWD is investigating ways to improve the quality of its incoming
water, including using alternative sources from the Sacramento-San Joaquin
Delta.
We have too little information to say whether the increased levels of
chloramines found in chloraminated water pose a health irisk. While there is
sane evidence that chloramines may cause adverse health effects, no thresholds
for those effects have been calculated.
-------
4-20
TABLE 4-7
ESTIMATED LIFETIME INDIVIDUAL CANCER RISKS
FROM DISINFECTED SURFACE WATER1
SOURCE
Hetch-Hetchy
ESTIMATED
RISK
1 X 1(T4
CHANCES IN A MILLION WEIGHT OF
OF CONTRACTING CANCER EVIDENCE
100 B22
South Bay Aqueduct
- Rinconada
- Penitencia
1 x 10~4
1 x 10~4
100
100
B2
B2
City of Santa Clara
Local Surface Water
- Redwood Mutual
- SJ .Water Company
- Saratoga-^
- Montevina
- Small Systems
,-5
4 x 10
9 x 10~6
1 x 10~4
8 x 10~5
40
1 x 10
-4
100
80
100
B2
B2
B2
B2
B2
1 Exposure estimates are discussed in the text and the Appendix. Toxicological
potency estimates are discussed in the text.
2 EPA considers chloroform a class B2, or probable, human carcinogen. EPA has
not classified the other THMs foir carcinogenicity. See text for details.
3 The Saratoga Water Treatment Plant is used part-time; estimates reflect
risks as if exposure is full time.
-------
4-21
SOURCE
TABLE 4-8
ESTIMATED ANNUAL CANCER INCIDENCE
FROM DISINFECTED DRINKING WATER1'2
ESTIMATED ANNUAL
INCIDENCE
WEIGHT OF
EVIDENCE
Hetch-Hetchy
B2
South Bay Aqueduct
- Rinconada
- Penitencia
.5
.3
B2
B2
City of Santa Clara3
.05
B2
Local Surface Water-3
- Redwood Mutual
- SJ Water Company
- Saratoga^
- Montevina
- Small Systems
.0002
.004
.04
.002
B2
B2
B2
B2
TOTAL
1.3
Exposure estimates are discussed in the text and the Appendix, lexicological
potency estimates are discussed in the text.
EPA considers chloroform a class B2, or probable, human carcinogen. EPA
has not classified the other THMs for carcinogenicity. See text for details.
3 Data on levels of individual THMs unavailable; risks computed using total
THM data, as if all THMs =jre chloroform.
The Saratoga Water Treatment Plant is used part-time; estimates reflect
risks as if exposure is full-time.
-------
4-22
TABLE 4-9
COMPARISON OF MONITORED SURFACE WATER LEVELS OF TRIHALOMETHANES
WITH PRESUMED THRESHOLDS FOR NON-CANCER HEALTH EFFECTS
[All values are in micrograms per liter (ug/1)]
Presumed Human
Highest Avg. Estimated
Substance
Chloroform
Bromoform
Dichloro-
bromome thane
Chlorod i-
bromome thane
Health Effect
Liver
Kidney
Fetal
Neurological
Liver
Thyroid
Immune System
Fetal
Liver
Thyroid
Immune System
Liver
Thyroid
Immune System
Kidney
Threshold 1
699
225
795
117
210
210
210
794
210
210
210
210
210
210
1000
Concentration
52
9
26
29
1. See Table 4-2 for further information on thresholds.
-------
4-23
TABLE 4-10
COMPARATIVE DATA
TOTAL TRIHALOMETHANE LEVELS IN DRINKING WATER
CITY TOTAL THM CONCENTRATION (ug/1!
Baltimore, MD* 50/54
Concord, CA** 111)
Los Angeles, CA 49
Oakland, CA 45
San Diego, CA 9"?
* From 1981-1983 data; concentrations from two treatment
plants
** All California data from National Organics Monitoring
Survey, 1976-1977
-------
4-24
Metals and Minerals
A nunber of inorganic substances are regulated by state and Federal
authorities, who require public systems to monitor for the chemicals regularly.
As with THMs, we have a relatively good database with which to estimate exposure
to and risk from these substances. We have used data for recent years, drawn
from Department of Health Services records, to estimate current levels of
exposure to these substances. These data are summarized in Table 4-11. We
used a slightly different method to estimate arsenic levels in drinking water;
these estimates are summarized in Table 4-12, and the methodology explained in
the Appendix.
Because of the nature of the contamination sources (natural soil constituents
and, in some cases, unidentified past contamination incidents), it is likely that
contaminant levels will be similar to current levels.
Arsenic is the only inorganic substance we examined in drinking water which
EPA's CAG considers carcinogenic if ingested. However, substantial disagreement
exists as to the toxicity of low levels of arsenic in drinking water. We have
therefore estimated a range of possible cancer effects, shown in Table 4-12.
Estimated individual cancer risks range from zero to seven thousand in a million,
aii.i estimated annual increased incidence ranges from zero to about seven cases.
If arsenic is carcinogenic, it would pose the greatest increased risk of cancer
ir- drinking water. However , the higher estimate is suspect, since increased
incidence of skin cancer has not been observed in areas with far higher levels
of arsenic than those in Santa Clara. Arsenic levels are well below the lowest
presumed threshold for non-cancer effects.
Of the other inorganic substances examined, lead is probably the most
significant cause for concern. Substantial evidence of lead's toxicity exists,
particularly for children exposed to lead in air and dust. Unfortunately, no
thresholds have been calculated that would allow an evaluation of the risks
posed by monitored levels of lead in Santa Clara Valley water ( it is generally
agreed that estimation of lead effects should take into account the total "body
burden" from air, water and dust, so that a water- or air-specific threshold
would be inappropriate). Lead is the subject of substantial research work at the
Federal level.
All other metals appear to be present at levels well below no-effect
thresholds for non-cancer effects, with the exception of mercury and cadmium,
for which the highest monitored value was close to the threshold for neurobehavioral
effects (see Table 4-13). Most of the inorganic substances for which monitoring
data are available (with the exceptions of arsenic and possibly lead) are not
considered carcinogenic by ingestion. Table 4-14 provides national information
on the levels of minerals and metals in drinking water.
-------
4-25
TABLE 4-11
SUBSTANCE
EXPOSURE TO METALS AND MINERALS
[system average/highest monitored ppb)l
WATER PURVEYOR
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Zinc2
Fluoride 3
Hetch Cuper- Gilroy Great Los Morgan San Jose
Hetchy tino Oaks Altos Hill Water Co.
[See Table 4-12]
<500 <1000 <250 <250 <173/220 <1000 <136/180
<1 <10 <10 <10 <1 <10/10 <1
<1 <5 <5 <30 <1 <6/8 <1
<1 <5 <5 <11/40 <5 <5 <5
<0.5 <1 <1 <5/6 <1 <1 <1
<5 <10 <10 <10 <2 <10 <1
<1 <20 <20 <10 <1 <20 <1
<1 <50 <55/80 <21 <10 <60/90 <10
Sunny- Snail
vale Systems
<100 <112/200
0 <2/4
<5/5 <5/9
<1 < 10/10
0 <1/1
<1 <5/7
<1 <2/3
CIO
depends
MCL/Action
Level
10,000
10
50
50
2
10
50
5,000
varies;
on temper ati
Single values are detection limits. Dual values are system averages (with results below detection limits
averaged at the detection limit), followed by highest monitored level.
zinc is added by the SCVWD for corrosion control at levels of 500 - 700 ug/1.
Monitored fluoride levels in systems that are not fluoridated were less than 1,000 in all tests.
-------
4-26
TABLE 4-12
ESTIMATED EXPOSURE TO AND CANCER RISKS FROM
ARSENIC IN DRINKING WATER1
Source and Purveyor
Concentration
(ug/l)
Individual
Risk
Population
Exposed
Incidence
(cases per year)
Rinconada
Cupertino
Cal Water
SJWC
Sunnyvale
3.1
3.1
3.1
3.1
- 1.4 x 1C-3
14,000
33,600
127,000
32,000
n
\^
0
0
0
,.3
,7
2.5
,6
Penitencia
SJWC
S.J. Evergreen
2.1
2.1
- 9.5 x 10-4
99,400
37,300
u -
1.3
.5
Other
City of Santa Clara2 1.2 0
SJWC 6.0 0
Redwood Mutual 15.0 0
Boys Ranch 11.0 0
Oakmont Mutual 6.0 0
San Ysidro School 5.0 0
- 5.4 x 10-4
- 2.7 x 10~3
- 6.8 x 10-3
- 5.0 x 10~3
- 2.7 x ID'3
- 2.3 x 10-3
87,700
2,300
1,300
175
175
175
0
0
0
0
0
.7
.09
.1
.01
.007
.006
Hetch-Hetchy
Milpitas
Mountain View
Palo Alto
Purissima Hills
Stanford
SJ/Alviso
Sunnyvale
Average
Total
.3
.3
.3
.3
.3
.3
.3
- 1.2 x 10~4
0 - 7.7 x 10-4
42,000
52,000
56,000
7,000
11,000
1,000
43,000
647,125
0
0
0
0
0
0
0
.07
.09
.1
.01
.02
.002
.07
0-7.2 cases/yr.
Some systems do not appear to contain arsenic; exposure estimates here reflect
levels in both surface and groundwater.
The City of Santa Clara blends water from Hetch-Hetchy, Penitencia, and
groundwater. Concentration estimate is average of Hetch-Hetchy and Penitencia
levels.
-------
4-27
TABLE 4-13
COMPARISON OF HIGHEST CONCENTRATIONS OF METALS AND MINERALS WITH
LOWEST PRESUMED HUMAN THRESHOLDS
Lowest Presumed Highest
Hunan Threshold Concentration
Substance (ug/1) (ug/D*
Arsenic
Barium
Cadmium
Chromium
Mercury
Selenium
Silver
Zinc
Fluoride
133
1,750
17,5
170 (Cr +6)
35,000 (Cr -t-3)
11
105
91
25,900
1,000-4,000
15
220 U,000
10
9 (30)**
6
7 (10)
3 (20)
90
***
* Numbers in parentheses indicate highest detection limit. Where there
is no such number, the highest concentration exceeded all detection
limits.
** Combined figure for chromium +6.
*** Monitored fluoride levels in systems that are not fluoridated were
less than 1,000 ug/1 in all tests.
-------
4-28
TABLE 4-14
COMPARATIVE DATA
FROM NATIONAL STUDIES ON
METAL AND MINERAL LEVELS IN DRINKING WATER*
[All values are in micrograms per liter (ug/1)]
frequency
range
average date of study
Arsenic
NA
<10-1000
NA
NA
Barium
94%
6%
0-100
100-172
NA
1970
Cadmium
1%
99%
NA
1971
Chromium
74.5%
24.5%
1-112
9.7
1967
Fluoride
NA
200-4400
<1,000
1969
Lead
NA
NA
13.1
1970
Mercury
Selenium
Silver
NA
NA
NA
0.1-1.8
0-10
NA
NA
NA
1970
1963
NA
Zinc
NA
NA
NA
NA
* All data from "Guidance for Issuance of Variance and Exemptions" by the
EPA Office of Drinking Water.
-------
4-29
Pesticides
Data on pesticides in drinking water are much less complete than data on
trihalomethanes and inorganic substances. The existing data indicate that
pesticides are typically not detected, or are detected only at trace levels in
imported surface water and local groundwater.
Pesticides and other agricultural chemicals could contaminate drinking
water in the Santa Clara Valley by several routes:
o Contaminants from other regions could be imported with surface water
supplies via the South Bay or Hetch-Hetchy Aqueducts;
o Chemicals used in local farms could seep downward and affect groundwater
supplies;
o Non-farm use or misuse or inappropriate disposal of chemicals could
result in contamination of local surface or groundwater;
o Cverspraying during chemical application could afreet local surface water.
The IEMP has examined potential pesticide contamination by reviewing
TOnitoring data for imported surface water and local surface water and groundwater
systems. This approach is applicable to all of the above exposure routes for
those sources that have been tested.
Imported Surface Water
In examining pesticide levels in imported surface water, the IEMP has
reviewed monitoring on contaminants at the Banks Pumping Plant, which feeds
into the South Bay Aqueduct. (Past monitoring has shown a good correlation
between contaminant measurenents at the Aqueduct intake and at the point of
delivery to the SCVWD.) Testing by the Department of Water Resources (DWR),
has shown only occasional traces of pesticidesless than 1 ug/1.
o The highest observed levels between 1971 and 1984 of a number of common
pesticides measured, for example, 2,4-D and 2,4,5-TP at 0.13 ug/1 and
0.16 ug/1. The Maximum Contaminant Levels for these chemicals are 100 ug/1
and 10 ug/1, respectively. Other chemicals tested are listed in the
Appendix.
o Recent testing (September 1984) at the Banks Pumping Plant in conjunction
with the Delta Health Aspects Monitoring Program showed no evidence of
25 pesticides.
o Periodic monitoring by DWR at the Banks Pumping Plant for groups of
organic compounds contained in pesticides have shown extremely low
levelsless than .1 ug/1 of chlorinated hydrocarbons, organic
phosphorous, and herbicides. The monitored levels are near the detection
limits and should not be viewed as exact measurements. Additional
monitoring at peak periods of pesticide use (early spring and summer)
is necessary before we can consider these data conclusive.
-------
4-30
Maximum Contaminant Level (MCL) Pollutants
Public water supply systems, which use both imported surface water and
local surface and groundwater, are required to test for several organic chemicals
regulated under Primary Drinking Water Standards established in the Safe Drinkirvg
Water Act. Pesticides regulated under this program are listed in the Appendix.
None of the targeted pesticides has been detected in major public water supplies
in the Valley.
Assembly Bill 1803
In addition to monitoring water systems under the Safe Drinking Water Act,
purveyors have undertaken further well monitoring under A.B. 1803, an extensive
monitoring program overseen by the Department of Health Services. This monitoring
is designed to detect local groundwater contamination, and tests for any suspected
contaminants from a group of 40 chemicals listed in the Appendix. The Department
of Health Services makes a case-by-case assessment of necessary pesticide
sampling, 'DOHS determined requirements for each water service district depending
on (1) use of chemicals in the last ten years, and (2) known agricultural
activity. No pesticides have been detected in the Santa Clara Valley under
AB 1803 monitoring.
South County
Small water systems and private wells in agricultural areas are more
vulnerable to pesticide contamination than deeper public wells, especially if
wells are relatively shallow. Past monitoring o£ these systems failed to
detect pesticides in groundwater:
o Tests by the DOHS for dibromcchloropropane (DBCP) in areas where this
chemical was suspected to have been used were negative.
o Between 1979 and 1983, four small system wells in Morgan Hill and
Gilroy were tested once each for the pesticides with MCLs (see Appendix).
The results were negative. At least one of these wells was near an
agricultural area that could have used a chemical that was included in
the monitoring, but the other wells were not highly vulnerable to
pesticide contamination.
o Since 1984, the State Water Resources Control Board has examined potential
agricultural chemical hot spots in parts of California, including the
Santa Clara Valley south of Morgan Hill. Twelve wells near Gilroy were
tested for halocarbon fumigants, and no contamination was detected,
although it was strongly suspected that these chemicals had been used.
This program is continuing in cooperation with the Gavilan WCD, and
will address other shallow wells in the South County.
Programs designed to detect pesticides in high risk South County wells and
water systems have been and are being conducted, but no contamination has been
found. See the Appendix for information on the pesticides for which monitoring
has been conducted.
-------
4-31
Sirmary: Surface Water Contamination
The IEMP analysis of risks from surface water is based largely on monitoring
data. We examined three classes of toxic substances: disinfection agents and
by-products (chloramines and trihaloroethanes), metals and minerals, and pesticides.
It is important to remember that this analysis is based on available monitoring
data and an initial assessment of the sources and pollutants that are likely to
cause toxic effects through chronic exposure. We believe that we have identified
the pollutants and sources that are of greatest concern; however, there is no
way to be certain that we have addressed all significant contaminants.
Trihalonethanes appear to account for a substantial portion of total health
risks from surface water sources of drinking water. We estimated individual
cancer risks of 10 to 200 chances in a million, and estimated aggregate incidence
of slightly over one additional cancer case per year. THMs appear to be below
estimated no-effect thresholds for non-cancer effects. Chloramines, which are
used as alternative disinfection agents in addition to chlorine, are not considered
carcinogenic but may increase the risk of certain other health effects.
THMs are one of the most significant risk issues in the IEMP Stage I analysis.
About half of the Santa Clara Valley's population is exposed to THMs, and THMs
appear to pose cancer and possibly other health effects. While disinfection is
necessary to prevent the spread of infectious disease, different disinfection
technologies exist that result in different THM levels; thus, THM exposure is
potentially controllable to a significant extent. Based on data from the SCVWD,
we estimate that chloramination reduces THM levels, and thus risks, by about
1/3 from THM levels generated by chlorination alone. The SCVWD is studying
alternative disinfection methods, which appears to be an appropriate first
step in addressing risks from THMs in the Santa Clara Valley.
Most metals and minerals appear to be present at levels sufficiently low
that they do not pose a health threat from chronic exposure. Two significant
exceptions are arsenic and lead. Substantial uncertainty exists as to the toxicity
of low levels of arsenic in water, and we estimated a range of possible effects.
Assuming that arsenic is carcinogenic implies local individual increased cancer
risk of slightly under 1,000 chances in a million and estimated incidence of about
seven cancer cases per year. However, this projection may be incorrect, since
there is no evidence of Blackfcot's Disease, the form of skin cancer attributed
to arsenic, in areas with far higher levels of arsenic than those in the Santa
Clara Valley. Local levels of arsenic are not particularly high by comparison
with other areas.
Substantial evidence exists that lead may cause adverse health effects,
particularly in children. Evaluation of the effects of chronic lead exposure
is a ccnplex area, and is currently the subject of ongoing research efforts. We
^re unable to estimate possible risks for this report.
Pesticide data are not as complete as data on the other classes of surface
water contaminants, because pesticide sampling is less frequent. The data that
do exist - good data on a few pesticides, more limited data on a broad range of
Pesticides - indicate that pesticide contamination of imported surface water is
present only at very low levels. Based on the available data, pesticide contamination
°f drinking water does not appear to pose a significant public health problem
in the Santa Clara Valley.
-------
4-32
EXPOSURE AND RISK: GROUNDWATER CONTAMINATION
Groundwater contamination differs from pollution in other media (e.g.,
air, surface water) in several ways that affect the evaluation of public health
risk and risk management options. Air emissions and surface water discharges
result in relatively rapid contamination of the resource in question, followed
quickly by exposure to people; elimination of the source of pollution generally
is followed fairly rapidly by a natural cleansing of the environment (through
contaminant degradation and dilution) and consequent reduction of exposure and
risk to people. In contrast, groundwater contamination is often a slow process:
releases generally take some time to reach the groundwater, and the groundwater
itself often moves quite slowly, so that contamination may be slow to spread
(this is less true of the Santa Clara Valley, where groundwater flow rates in
some areas are quite rapid). Moreover, groundwater contamination does not
automatically result in peoples' exposure to toxic contaminants; tainted groundwater
may or may not contaminate a drinking water well. Also, unlike exposure to
air pollution, a well that is known to be contaminated can be shut down, cutting
off exposure to toxics.
While contamination is a slow process with many links, it is often not
easily reversible. Once contaminated, aquifers are quite difficult to clean
up. Removing or containing sources of contamination helps, but often one
cannot rely on natural degradation of contaminants to restore the ecosystem;
indeed, the worst effects of contamination may not be felt until long after
the source has ceased releasing contaminants.
Our ability to understand groundwater contamination dynamics is hindered
by our poor grasp of the basic chemical, physical, and biological processes
involved. Local hydrogeologic idiosyncrasies can result in significant variations
in contaminant levels over small distances. These characteristics make it
difficult to model groundwatec contamination reliably. Thus, it is more difficult
to apply standard environmental modeling and monitoring techniques to groundwatetr
contamination, and our analysis of potential groundwater health risks is therefore
complex. We could not simply assess current conditions and extrapolate to the
future with few changes. Instead, beginning with the available information on
current conditions, we developed assumptions about the key processes in the
chain from substance release to final exposure to project future risks. Although
such prognostication is highly uncertain, there is no way to avoid it in evaluating
the future effects of current groundwater contamination, as well as the impact
of future leaks.
Groundwater's unique characteristics have implications for risk management.
Since groundwater contamination and human exposure are not necessarily linked,
a wide variety of actions can be taken to prevent either or both. For example,
reducing potential sources of contamination and acting to clean up aquifers
might be effective actions for preventing and repairing groundwater contamination,
while well-head monitoring might be the most effective way to control exposure.
In addition, the relative irreversibility that often occurs with groundwater
contamination has implications for risk management: mistaken actions (or
inaction) are more costly, since they may take lifetimes or longer to correct.
In this section, we first discuss the groundwater contamination issues and
pollutants we chose to examine. Then we briefly describe the hydrogeology of
the Santa Clara Valley area. Next, we present a 'snapshot1 of what we know
-------
4-33
about current gtroundwater contamination and exposure, and evaluate the health
risks of that exposure. Finally, we describe our methods for projecting the
future impacts of groundwater contamination, along with results. Our analysis
of future exposure levels and risks is necessarily speculative; it should not
be taken as a literal projection, but rather as an effort to. make plausible
bounding assumptions about key factors and thereby develop an understanding of
the probable range of future impacts. We have tried to be conservative in
projecting health effects (i.e., err on the side of over- rather than under-
estimating risks). We have made extensive use of sensitivity analyses in
examining key uncertainties in estimating future exposure.
Groundwater Contamination Issues
Numerous toxic contaminants ftrom various sources have contaminated groundwater
in the Santa Clara Valley. As a result of discussions with agencies, elected
officials, industry and citizen representatives, we decided to examine potential
health risks from the following sorts of contamination:
1) organic chemicals frcrn spills, leaking underground tanks and pipes,
above-ground tanks, and illegal disposal;
2) fuel constituents from spills and leaking underground tanks and pipes;
3) organic chemicals from other sources, including sanitary landfills,
sewer line leakage, septic tanks, and runoff to dry wells; and
4) nitrates from wastes, fertilizers, and septic tanks.
Since very few wells in the Valley are disinfected with chlorine, we did not
estimate risks from disinfection byproducts for groundwater.
Organic Chemicals frcm Tanks, Pipes, Spills, and Illegal Disposal
Background
The problem of toxic organics released from underground tanks and sumps at
industrial facilities in the Santa Clara Valley emerged in late 1981 when
Fairchild Camera and Instrument Company discovered a leak in an underground
storage tank at its facility in south San Jose. Subsequent study of inventory
and disposal records indicated that the tank had released about 60,000 gallons
of waste solvent and water mixture over at least 18 months. When Fairchild
discovered the leak, the Great Oaks Water Company found its well #13, serving
about 700 people, contaminated with 1,1,1-Trichloroethane (TCA) at about 5,800
ppb. (The state action level for TCA is 200 ppb.) Great Oaks immediately
removed this well from service, and has a policy of not serving water with any
toxic contamination.
Subsequent investigation by other firms, under the direction of the Regional
Water Quality Control Board, has revealed about 100 other soil or groundwater
contamination sites in the Valley. (These investigations examined tank sites
and found contamination in other parts of the Bay Area as well.) At most
sites, investigators eliminated tank failures as a cause of contamination,
suggesting that many of these contamination incidents are due to causes such as
spills.
-------
4-34
Levels of contamination found in the soil and groundwater vary widely.
The stages that sites have reached in assessing contamination or in planning
or undertaking clean up efforts also vary. A complete description of industrial
contamination sites is not attempted here, and the interested reader is referred
to "Assessment of Contamination from Leaks of Hazardous Materials in the
Santa Clara Groundwater Basin, 205 J Report," 1985.
Most industrial contamination has involved leaks or spills of solvents,
such as 1,1,1-trichloroethane or CFC-113, on the grounds of electronics firms.
In general, the organic chemicals considered most likely to be toxic are halogenated
organics, such as 1,1,1-TCA, trichloroethylene, and methylene chloride; and
aromatic compounds, such as benzene, xylene and toluene. In this report, we
examine exposures to and risks from organic contaminants that, have contaminated
public drinking water supply wells; many of the substances on the RWQCB's list
of the 20 most commonly found contaminants at industrial sites; and various other
substances for which we had adequate information to estimate risks.
Releases from above-ground tanks are a potentially important source of
groundwater contamination. About 275 above ground tanks exist in the Santa
Clara Valley, containing primarily waste water, caustic sodas, and acids.
Vfe have not estimated releases from above-ground tanks because we do not expect
these substances to pose significant chronic health risks in groundwater.
These substances generally are either very dilute (waste water), or are likely
to become largely neutralized in soils (sodas and acids). In addition, the
acids are probably not as potent in groundwater as other substances in our
analysis.
Contamination of Groundwater by Fuel
Although the bulk of local concern about groundwater contamination has
centered on industrial solvent sites, there are more than ten times as many
underground fuel tanks as industrial tanks (over 6,000 fuel tanks compared to
300 - 400 industrial tanks). These tanks are subject to the same problems -
corrosion leaks, piping failures, spills due to improper handling - as industrial
solvent tanks. Most older gasoline tanks are single-walled steel vessels,
with no corrosion protection. The risk that these tanks will fail due to
corrosion increases with tank age; the average age of steel tanks that have
failed is about 15 years. Many tanks in the Santa Clara Valley are approaching
this age. Some oil companies have been systematically replacing older steel
tanks in the Santa Clara Valley and elsewhere in the country. (We discuss our
estimates of the rate of tank failures for gasoline tanks later.)
About three years ago, releases of 300,000 and 100,000 gallons occurred at
tank farms in San Jose. Since then, authorities have discovered about 400
additional leaks in Santa Clara County, ranging from about 100 to 10,000 gallons
each. Many of these leaks are due to pipe rather than tank failures. Authorities
are discovering additional leaks as monitoring results are reported under
local Hazardous Materials Management Ordinances, or HMMOs.
Sources of data on gasoline leaks and groundwater contamination include
site monitoring data required under local HMMOs and voluntary tank removal and
station upgrade programs. Cities have not yet fully implemented site monitoring
under the HMMOs, and we may discover more gasoline leaks as investigations
continue.
-------
4-35
Sane common fuel constituents may be toxic at low levels. Lead may
be toxic, but it is probably not a major groundwater problem in Santa Clara
because it does not appear to move readily through soil and groundwater. We
have assumed that lead is relatively immobile in Santa Clara Valley soils. In
this report, we concentrate on estimating exposures to and risks from benzene
and ethylene dibroraide, the two most mobile and toxic gasoline constituents.
Toxic Organics from Sewer Lines, Sanitary Landfills, Septic Tanks, and Dry Wells
Sewer lines leading to municipal treatment plants carry a range of toxic
organics and metals, as shown by effluent sampling at industrial dischargers
and influent sampling at municipal plants. Sewer lines are not perfectly tight,
resulting in infiltration during rainy periods and leakage during dry periods.
Toxic organics are a more significant groundwater concern than metals, because
they are far more mobile. Local sewage treatment plant operators regulate the
amount of total toxic organics discharged into sanitary sewers.
While toxic wastes are formally barred from Class III sanitary landfills
(the only landfills sited in Santa Clara Valley), such wastes may be present
in household wastes (home pesticides, for example) and commercial waste.
Toxic wastes may also have been deposited in the past, before today's relatively
stringent waste disposal regulations. Finally, some toxics may be created
through reactions of other substances.
Under contract to EPA, the Association of Bay Area Governments (ABAC) analyzed
landfills in Santa Clara County to characterize potential contaminant releases
to air, surface water, and groundwater. This analysis indicates that, with
the possible exception of the Singleton Road landfill, the eleven permitted
landfills in the Valley have little potential to contaminate drinking water
with toxics because leachate would probably move toward San Francisco Bay and
away from drinking water. Groundwater monitoring at the Singleton Road landfill
near San Jose (now closed) has detected contamination by toxic organics, including
trichloroethylene. We characterized releases of TCE and methyl ethyl ketone
for a facility of approximately the same age and acreage as the Singleton Road
site. We then estimated potential exposure from those releases. Since our
estimated exposure levels were very low for both TCE and MEK, we did not estimate
risks from this source. For further information on this part of the analysis,
see the Appendix to this Chapter.
Septic tanks are used widely in parts of the Santa Clara Valley, particularly
in Los Altos Hills and in the South County area. The use of solvents for
cleaning out septic tanks can result in significant releases of toxic chemicals
to groundwater. The concentrations of contaminated groundwater reaching drinking
water wells depend on several factors, including the location of wells and
septic tanks, the construction of the wells, the amount the released material
is diluted in the soil by rainfall, and hydrogeologic characteristics. Under
some conditions, releases from septic tanks could pose siginificant risks.
Because of the difficulty in modeling exposure from releases to septic tanks,
and because we do expect this exposure to be substantially less than the exposure
to other sources in our analysis, we have not estimated risks from toxics
released from septic tanks.
Dry wells are shafts, up to three feet in diameter and up to 50 feet deep,
sunk to accept runoff from percolation back into the water tablo. In some
-------
4-36
parts of the Santa Clara Valley dry wells are used in lieu of storm drains.
Where runoff is contaminated or dry wells are used foir illegal disposal, the
wells can speed the introduction of contaminants to groundwater. Although
runoff in the Santa Clara Valley may contain fairly high levels of metals, risks
from runoff are probably substantially lower than risks from other sources we have
analyzed because runoff occurr only during and immediately after rain, and
concentrations of contaminants are diluted by the rain and by the groundwater.
Nitrates
Nitrates may contaminate groundwater from past and present fertilization
of crops and lawns, animal waste, septic and sewage treatment plant effluent,
feedlots and mushroom farms. Parts of South County have high nitrate levels,
partly because of past agricultural activity. The health risk on which drinking
water standards for nitrates are based is methemoglobinemia, or "blue baby
syndrome," which affects infants under six months old. However, other effects
may also occur in older children and adults at similar levels of exposure.
Formal and informal notification programs in South County, including bill
inserts from purveyors serving water over thresholds, and occasional articles
in local newspapers help inform users that their water may be unsuitable for
use by infants. These measures probably substantially reduce exposure to
infants.
In the Santa Clara Valley, exposure to high nitrate levels is most likely
to occur in households drinking groundwater from small public systems or private
wells. Small system wells are required to test for nitrates once every three
years, and private wells are usually not monitored. Although actual exposure
data for private and small public system wells are lacking, the IEMP has made
some rough estimates of potential exposure above thresholds.
We know of several systems in the Santa Clara Valley, particularly in the
South County area, with nitrate levels above standards:
0 The Morgan Hill system is under DOHS order to comply with nitrate
standards by 1988. The city is notifying customers of the problem.
0 One third of the wells tested in the San Martin area in 1978 and 1981
had nitrate contamination above the standard, and about one third of the wells
monitored by DOHS and the County Health Department in San Martin, Morgan Hill,
and Gilroy were consistently above standards. Gilroy has also encountered high
nitrate levels in tests to locate potential future water supplies.
Based on these monitoring results, we estimate that one third of private
and small system wells may also have nitrate levels above standards. Extrapolating
from county-wide census data, we estimate that at any given time there may be
about 50 to 100 infants less than six months old in the South County area, and
therefore potentially vulnerable to nitrate contamination. It is very difficult
to estimate how aware consumers are about the nitrate issue, but the notification
mechanisms have probably helped increase awareness and therefore lowered the nunber
of infants potentially exposed. We do not know of any cases of blue baby
syndrome in the Santa Clara Valley.
-------
4-37
Health Effects Information
In this section, we summarize the qualitative and quantitative data on the
adverse health effects that may be caused by chronic exposure to the chemicals
we are analyzing. We examine the evidence for a number of toxic effects,
including cancer, kidney effects, and others. Vfrierever possible, we rely on
peer-reviewed EPA estimates of toxicological potency. In some cases, we present
additional information that is not as fully reviewed. (See Chapter 2 and the
Appendices for a fuller discussion of methodology.)
Table 4-15 presents data on the carcinogenicity of the substances we have
analyzed. Two types of information are presented: (1) a qualitative estimate
of the strength of evidence that a substance is carcinogenic; and (2) where a
substance is a suspected carcinogen, a quantitative estimate of the substance's
potency. To summarize the strength of evidence, we present the classification
scheme of the International Agency for Research of Cancer (IARC) and a similar
classification scheme developed by EPA. We briefly discuss these schemes in
Chapter 2. Quantitative potency estimates are from EPA's Carcinogen Assessment
Group (CAG)..
Many of the substances of concern are halogenated organic chemicals,
which are commonly used as industrial solvents. Significant controversy exists as
to whether these substances, including 1,1,1, trichlorethane (TCA), trichlorethylene
(TCE), perchloroethylene (PCE), 1,1, dichloroethylene (1,1 DCE or vinylidene
chloride), and methylene chloride, are carcinogenic. Much of the positive
evidence for these substances' carcinogenicity is the occurrence of liver
tumors in exposed laboratory mice. However, the mice used in these experiments
have a high background rate of liver tumors, and the scientific community
disagrees on the significance of these findings. While definitive scientific
consensus on the carcinogenicity of most of these substance has not been reached,
we follow EPA policy in this report in designating a substance as a "possible,"
"probable," or "proven" carcinogen.*
To examine the potential non-cancer health effects of the chemicals we are
analyzing, we compare measured or modeled concentration levels with estimated
thresholds, below which adverse effects are assumed not to occur. These thresholds
include both EPA-developed Reference Doses and additional health effect thresholds
estimated by a toxicology consultant for this study. Possible adverse health
effects, estimated quantitative thresholds, and information sources are presented
in Table 4-16. Where evidence for an effect exists, but there are insufficient
data to calculate a quantitative threshold, the effect is listed and no threshold
estimate presented.
* EPA's current policy is that 1,1,1 trichloroethane should not be considered
a possible carcinogen, and we follow that policy as our base case. However, a
significant portion of the scientific community believes that TCA may eventually
be found to be carcinogenic. EPA formerly considered TCA a possible carcinogen,
but has suspended that classification because a key implicating study is under
review.
Because of the uncertainty on this issue, we have performed sensitivity
analysis of the possible impact of TCA if it were a carcinogen.
-------
4-38
TABLE 4-15
CARCINOGENIC POTENCY VALUES AND STRENGTH OF EVIDENCE OF CARCINOGENICITY
FOR SUBSTANCES IN IEMP GROUNDWATER RISK ANALYSIS
Level of Evidence
Substance
Human
Animal
EPA Weight
of Evidence
foe Caircinogenicityl
Potency Value3 ug/1
Benzene
Chloroform
1,1 Dichloroethylene
Ethylene Dibromide
Methylene Chloride
Perchloroethylene
Trichloroethylene
Vinyl Chloride
Ttrichloroe thane 3
S
I
I
I
I
I
I
S
I
S
S
L
S
S
L
S
S
A
B2
C
B2
B2
B2
B2
A
Not Classified
8.2 x 1CT7
2.3 x 1CT6
1.7 x 10~5
1.17 x 10~3
2.10 x 10~7
1.5 x 10~6
3.2 x 1C-7
6.60 x 10~5
0
S: Sufficient evidence
L: Limited evidence
I: Inadequate evidence
1 The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to very
substantial. According to EPA's categorization of levels of evidence of carcinogenicity, A = proven human
carcinogen; B = probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2);
C = possible human carcinogen; D = not classifiable; and E = no evidence.
2 Source: EPA's Carcinogenic Assessment Group (CAG)
Potency values are plausible upper-bound estimates of human effects.
3 Alternative case TCA Potency Score: 4.58 x 10~8 (ug/1)"1
-------
4-39
TABLC 4-16
NON-CANCER HEALTH EFFECTS AND PRESUMED HUMAN THRESHOLDS FOR
SUBSTANCES IN IEMP GROUNDWATER RISK ANALYSIS (1)
Health
Effect2
Blood
Fetal
Liver
Neurological
Kidney
Reproduct ive
Fetal
CFC-113
(l,l,2-Trichloro-l,2,2-trifluoroethane)
Psychomotor performance
Substance
Benzene
Carbon
Tetrachloridi;
Presumed
Human Threshold (ug/1)
Source
3,4
Reference
1
4
24.5
41
24
24
,080
,300
242
EPA/RFDUV
IEMD
EPA/RFDUV
IEMD
IEMD
IEMD
IEMD
Snyder, et. al. , 19
Kuna and Kapp 1981
EPA, 1980 & 84
Moller, 1973
EPA, 1980
Adams, et. al. , 1952
Schwetz, et. al. 1974
1,050,000
EPA/RFDV
Imbus & Adkins 72
Chloroform
Liver
Kidney
Fetal
Neurological
Mutagenicity
350
225
795
117
EPA/RFDUV HevwDOd, et. al. 79
IEMD Heywood, et. al., '79
IEMD Thompson, et. al.,'74
IEMD Challen, et. a.., '58
EPA, 1985
-------
4-40
TABLE 4-16 (Continued)
Substance
1,1
Dichloroethylene
Ethylene
Dibrcnide
Methyl Ethyl
Ketone
Methylene
Chloride
Nitrate
Health
Effect
Liver
Kidney
Mutagenicity
Reproductive(male)
Reproduct ive(female)
Liver
Kidney
Mutagenicity
Presumed
Human Threshold (ug/1)
Source
Liver
Blood
Fetal
Neurological
Kidney
Mutagenicity
Blood
Cardiovascular
Fetal
310
24.8
17.5
119
626
128
1,700
2,100
86,000
2,100
86,000
6,990
EPA/RFDV
IEMD
IEMD
IEMD
IEMD
IEMD
EPA/RFDV
EPV'RFDV
IEMD
IEMD
IEMD
IEMD
10,000 EPA/RFDV
(up to 6 months)
133,000 IEMD
(7 months to 13 years)
78,700 IEMD
399,000 IEMD
Reference
Nitsche, et.al., 79
NTP, 1982
NTP, 1982; Hurtt & Zenick, 1985
Short, et. al., 1977
NTP, 1982
NTP, 1982
EPA, 86
EPA 1984; Burek, et al, 80 & 84
Burek, '80; NIOSH '76
Schwetz, et. al. 1975
EPA, '85; NIOSH '76
NTP, 1985 (Draft)
Winton, et. al., 1971
Putokov, et. al., 1970
Shuval and Gruener, 1977
Sleight and Atallah, 1968
-------
4-41
TABLE 4-16 (continued)
Substance
Perchloroethylene
Toluene
Health
Effect
Liver
Kidney
Fetal
Blood
Liver
Neurological
Presumed
Human Threshold (ug/1)
699
699
9,090
10,100
10,100
5,800
Source
EPA/RFDV
IEMD
IEMD
EPA/RFDV
IEMD
IEMD
Reference
Rossmiller, 1953
NCI, 1977
Nelson, et. al. ,
CUT 80
EPA, 1983
Hanninen, et. al.
1979
, 1976
Seppalainen et. al. 1978
Trichloroethane
Trichloroethylene
Kidney
Reproductive
Fetal
Liver
Neurological
Fetal6
Liver
Neurological
Kidney
10,100
5,000
4,760
979
12,500
260
260
37,700
IEMD
IEMD
IEMD
EPVRFD^
IEMD
EPVRFDUV
IEMD
IEMD
EPA 1983
Matsumoto, et. al
Hudak & Ungvary,
McNutt, et. al.,
Bruckner, et. al.
EPA, 1984
Grandjean, et. al
Tucker, 1982
. , 1971
1978
1975
, 1985
. , 1955
-------
4-42
TABLE 4-16 (continued)
Substance
Vinyl Chloride
Xylene
7
Health
Effect
Liver
Cardiovascular-
Blood
Kidney
Reproductive
Fetal
Liver
Neurological
Cardiovascular
Blood
Kidney
Reproductive
Fetal
Presumed
Human Threshold
45.5
246
3,010
246
1,640
1,640
2,150
2,150
2,150
2,150
2,150
528
528
Source Reference
EPA/RFDUV Feron, et. al., 1979
IEMD Feron, et. al., 1979
I HMD Feron, et. al., 1979
IEMD Feron, et. al., 1979
IEMD Ungvary, et. al., 1978
IEMD John, et, al., 1977
EPA/RFDUV Bowers, et. al., 1982
Tatrai, et. al., 1981
EPA/RFDUV Savolainen, et. al., 1979
EPA/RFDUV Hipolito, 1980
Smyth & Smyth, 1928
EPA/RFDUV Browning, 1965
Hipolito, 1980
EPA/RFDUV EPA, 1984
IEMD Ungvary et. al. , 1980
IEMD Ungvary et. al., 1980
FOOTNOTES ON NEXT PAGE
-------
4-43
TABLE 4-16 (continued)
FOOTNOTES
(1) For sake of brevity, negative evidence (i.e., where a laboratory test or epidemiological study has been
performed but no evidence found of a health impact) is not reported here. More complete toxicoloqical
information is available from IEMP.
(2) Mutaqenic effects cannot be expressed, meaningfully in guantitative terms. Therefore, we note positive
evidence of mutagenicity by listing it as an effect.
(3) "IEMD" indicates a threshold estimated by TEMD toxicologists and consultants using existing literature and
following EPA procedures.
(4) "EPA/RfD" indicates a threshold derived from an EPA "Reference Dose" (RfD) or "Acceptable Daily Intake" (ADI)
level. RFDs are estimated no-effect thresholds that are intended to protect an individual from the most
potent non-cancer chronic health effect. EPA has reviewed and verified some of these thresholds internally,
although they have not necessarily been peer reviewed. Those which EPA has verified internally are
indicated by v; those which are unverified internally are indicated by uv.
(5) This threshold reflects the highest dose given at which no adverse effect was seen, with an additional safety
factor of 1,000. Therefore, this threshold has no particular adverse health effect as its basis.
(6) A limited number of studies have been con-lucted to evaluate the potential for adverse effects on the fetus
from exposure to TCA. These studies are summarized in U.S. EPA 1984, Health Assessment Document for
1,1,1-Trichloroethane (methyl chloroform), Final Report, EPA-600/8-82-003F, Washington D.C. and in IEMD,
"Health Score Evaluations for Pollutants in the Santa Clara Valley Integrated Environmental Management
Program: l,l,lTrich loroethane. "In its Health Assessment Document, U.S. EPA indicates that it is not
possible, on the basis of limited available data, to define the full potential of TCA to produce teratogenic
effects. Each of the available mammalian studies had methodological drawbacks that do not allow for
conclusive evaluation of the ability of TCA to produce a teratogenic response over a wide range of doses.
All of the studies performed in mammals at the time the Health Assessment Document was published had been
done in rats and mice. On the basis of these studies, it does not appear that short- or long-term exposure
to TCA results in teratogenic effects in rats ot mice. Therefore, the lEMP's base-case analysis assumes
that exposure to TCA poses no risk of fetal effects.
-------
4-44
TABLE 4-16 (continued)
An unpublished study, which has not undergone scientific peer review and which was completed subsequent
to the publication of EPA's final Health Assessment Document for TCA, reports fetotoxic effects (cardiac
malformations) in rat pups exposed in utero to TCA (Dapson et al., 1984). Given that this study was
extremely limited (using only one dose level) and that the resultsif reproduciblecould have significant
implications, the National Toxicology Program (NTP) has commissioned a project to repeat the Dapson study
using multiple dose levels. Results of the NTP study are expected in Fall 1986.
In order to assess the importance to Santa Clara Valley residents of further research on this issue, we
need to examine whether local exposures to TCA may be of concern if the Dapson findings are validated.
For this reason, IEMD toxicologists have calculated a threshold of 16 ug/m^ based on the Dapson study to
be used in this report as a sensitivity case pending the outcome of the NTP study. The sensitivity analysis
examines the possible impact of TCA under the assumption that exposures above the estimated threshold could
pose the risk of cardiac malformations. This analysis, which appears in footnotes to the text is not part
of our base-case analysis. THE SENSITIVITY RESULTS SHOULD NOT BE INTERPRETED AS INDICATING WHETHER A RISK
IN FACT EXISTS AND EPA RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK MANAGEMENT DECISION-MAKING OR
REGULATORY ACTION pending completion of the follow-up study by NTP.
(7) For all effects with a threshold of 2,150 ug/1 noted, we have also estimated an alternative threshold of
350 ug/1. This 350 ug/1 level is also an unverified RFD.
-------
4-45
Hydroqeology
It is important to have a basic undertandinq of local hydrogeology to
understand the water supply and contamination problems in the Santa Clara
Valley. Details on the Valley's hydrogeology, and the bases for many of the
assumptions used in the analysis, are presented in greater detail in the Appendix.
The Santa Clara Valley encompasses three groundwater subbasins: the
Santa Clara, extending southward from the county line to Coyote; the Coyote,
from Coyote to Morgan Hill; and the Llagas, from Morgan Hill to Gilroy. Of
the three, the Santa Clara subbasin serves the largest number of high-yielding
public water supply wells. This subbasin is composed of four zones (see Figure 4-5
for location of tour zones, and Figure 4-6 for a schematic cross-section diagram).
The Bay Sands/Bay Muds Zone borders parts of San Francisco Bay, and extends
about three to ten miles inland. Two types of areas, with different hydrogeologic
characteristics, are interspersed throughout this zone. The Bay Muds areas are
characterized by thick muds, and very few wells draw from these sections. We
estimated that groundwater flows less than one foot per year in the Bay Muds
zone. Bay Sands areas are characterized by permeable soils near the surface,
but groundwater moves relatively slowly; we estimated a groundwater flow rate of
60 feet per year in the Bay Sands. A key characteristic of the Bay Sands/Bay
Muds Zone is a clay confining layer 100 to 200 feet below the surface which
separates an upper aguifer serving shallow private wells -from the lower aquifer
from which public wells draw. This clay layer helps keep contaminants from reaching
the deep confined aquifer, but does not provide complete assurance against
contamination of deep wells.
The Inland Confined Zone borders the Bay Sands Zone and extends south
into San Jose. Its soils are somewhat less permeable near the surface than the
Bay Sands Zone, but groundwater gradients are steeper, and groundwater flows more
quickly. We estimated a groundwater flow rate of 500 feet per year in the Inland
Confined Zone. The clay layer in the Bay Sands Zone, described above, extends
into the Inland Confined Zone.
The Recharge Zone surrounds most of the Inland Confined Zone to the east
and west, and extends south to South San Jose. It is the section of the Santa
Clara subbasin furthest from the San Francisco Bay. Near the surface, Recharge zone
soils are similar to those in the Inland Confined zone. The clay layer protecting
the lower aquifer in the Bay Sands and Inland Confined Zone does not extend to
the Recharge Zone. Both shallow private and deeper public wells are therefore
more vulnerable to contamination. We estimated a flow rate of 3,500 feet per
year in the Recharge Zone.
Groundwater in the Santa Clara subbasin generally flows in a northerly
direction, from South San Jose toward San Francisco Bay. Water artificially
recharged in this area, therefore, tends to flow toward the Inland Confined and
Bay Sands and Muds zones.
The Llagas and Coyote subbasins in South County are characterized by permeable
soils and shallow groundwater gradients. Groundwater is essentially unconfined as it
neves in a northwesterly direction up the Valley in the Coyote subbasin. In the
Llagas subbasin, groundwater moves Southeast, and becomes more confined near
Gilroy. We modelled the Llagas and Coyote subbasins as a single South County
zone, with a groundwater flow rate of 180 feet per year and no confining layer.
-------
4-46
BAY SANDS/MUDS ZONE
SANTA CLARA
Recharge Zone
Inland Confined Zone
. Confined Zone Boundary
FIGURE 4-5
GROUNDWATER BASINS
AND THE CONFINED ZONE
-------
FIGURE 4-6 SCHEMATIC CROSS SECTION FOR SANTA CLARA SUB-BASIN
Southwest
Unconfmed
L Forebay Zone
Northeast
Unconilned Area lor Upper Zon
Confined Area for Lower Zone -
Well Symbols
Note: Arrows Indicate direction of groundwater movement without regard to quantity
-------
4-48
Risks from Current Exposure to Groundwater Contamination
Introduction
We analyzed both current and future risks fron various sources of groundwater
contamination. This section discusses risk from current exposure to known
groundwater contamination by organic chemicals. Leaks or spills from underground
storage tanks have caused at least some of this contamination, but we have been
unable to determine the cause of contamination in all cases. Figure 4-7 shows
the locations of known industrial groundwater contamination sites, and Figure
4-8 shows the locations of large public drinking water wells, including the
affected wells. We discuss groundwater contamination by nitrates and pesticides
separately.
It is important to remember the distinction between groundwater contamination
and human exposure. People may not be exposed to contaminated groundwater, either
because plumes have not affected drinking water wells, or because purveyors
have closed affected wells. If contaminated groundwater is not used as drinking
water, it does not pose any current risks. However, it may pose future risks
if the contamination subsequently reaches drinking water wells.
Analyzing risks fron current exposure at contaminated public wells is
reasonably straightforward. We have good data on concentrations in public
wells as a basis for estimating exposure. We also know how frequently these
wells are used, and the number of people they serve. We combined these exposure
data with our toxicological potency estimates to estimate risks.
Under Assembly Bill 1803, DOHS required all large public systems to monitor
for toxic substances in 1984-1985. This and other monitoring has tarqetted the
toxic chemicals of concern in Santa Clara County, including industrial solvents,
fuel constituents, pesticides (in some systems) and other organic chemicals,
and metals. (These substances are listed in the Appendix.) Most systems
tested all of their wells; some tested a representative sample. Future monitoring
will include annual testing of representative wells for toxic organics, and
more frequent monitoring of wells near known groundwater contamination, and of
contaminated wells.
Thirty-six large public system wells have been contaminated. Purveyors
have taken eight wells out of service, and have placed seven wells on a stand-by
basis. Twenty-one others with low levels of contamination remain in service.
Table 4-17 lists the contaminated large public wells, the sources of contamination
when known, the levels of contamination, the status of the wells, and the
number of people exposed at those wells which remain open.
Authorities "have not tested most small systems and private wells. However,
the County will monitor all small systems beginning in 1986 under AB 1803. In
addition, the County monitored 171 private wells in 1985 with assistance fron
the Regional Water Quality Control Board, and plans to monitor about 1,000 - 1,500
private wells in 1986-1987.
Thirteen of the 171 private wells tested in 1985, or about 8% of those tested,
were contaminated with organic chemicals. Five of these wells had concentrations
exceeding standards. Since the County monitored these wells because of their
-------
4-49
proximity to known contamination sites, the County had expected to find more
wells contaminated. In addition, bacteriological contamination was more widespread
than anticipated, with about 36% of the wells tested showing contamination
above standards. Table 4-18 summarizes the results of the 1985 monitoring for
organic chemicals, and of previous monitoring conducted in 1984. These results
reflect only known contamination at private wells; most private wells have not
been tested.
Since about 5,000 private wells exist in the Valley, recent monitoring has
not provided enough information for us to estimate current exposure or risks
for small public or private wells. In the following section, we use a modeling
approach to estimate future exposure and risk at both public and private wells.
Small public and private wells may be more vulnerable to groundwater contamination
because they are typically shallower and less protected by clay aquitards. In
the County's recent monitoring program, twice as many wells shallower than 150
feet were found contaminated than those deeper than 150 feet.
-------
WV^w "' X va"e
' -vV'*. '-' 1* Fl
-"''irVxVv. viV*
Ul
o
FIGURE 4-7 LOCATIONS OF CONTAMINATION SITES
-------
^ i
s V
V '
1-4 I
t.»
rVv^ "-
'x \ > «^ Santa Clara
* "V^V. Valley
,' Floor
;, =A/
}
\
FIGURE 4-8
LOCATIONS OF
WELLS SERVING
MAJOR WATER
PURVEYORS
A CONTAMINATED WELLS, STANDBY
A CONTAMINATED WELLS, ACTIVE
CONTAMINATED WELLS, CLOSED
O UNCONTAMINATED WELLS
.c-
I
-------
4-52
TABLE 4-17
CONTAMINATED LARGE PUBLIC DRINKING WATER WELLS
OPEN WELLS
Source Purveyor Well Contaminant
IBM SJWC Tully Well #1 TCA
CFC-113
#2 TCA
CFC-113
#3 TCA
CFC-113
#4 TCA
CFC-113
IBM City of San Jose
Evergreen #2 TCA
#3 TCA
#4 TCA
#5 TCA
IBM Rancho Santa Teresa
Mobile Home Park
Well #1 TCA
CFC-113
Well #2 TCA
Concentration
(ppb)a
<0.5 - 1.3
ND - 1.7
1.0 - 2.8
ND - 3.3
<1.0 - 1.8
<1.0 - 1.0
<1.0 - 2.0
<1.0 - 2.0
0.6 - 4.1
ND - 4.1
0.5 - 1.3
2.3 - 5.2
ND - 4
ND - 3
ND - 7
ND - 7
# of People
Use Exposed*5
in service
in service
35,000
in service
in service
in service during summer;
wells provide 10% of
annual supply
50,000
in service
700
in service
-------
4-53
TABLE 4-17 (continued)
OPEN WELLS (continued)
Source
Unknown
Unknown
Purveyor Wells
City of Santa Clara
Well #24
Well #20-02
SJWC Blossom Hill #4
#5
Contaminant
CFC-113
cis-1,2 DCE
TCA
TCA
Concentration
(ppb) Use
0.5 - 3.0 in service
0.5 in service
<0.5 - 2.4 in service
<0.5 - 1.1 in service
# of People
Exposed*3
4,900
200
9,300
(in 1985)
Unknown California Water Service Company
Hillview Well Carbon
Tetrachloride
ND - lc
0.9
in service
2,600
Unknown City of Gilroy
#1
#2
#3
TCA
PCE
TCA
PCE
TCA
PCE
ND -
.38 -
ND -
0.5 -
ND -
ND -
1.1
1.4
1.0
2.3
1.6
1.2
in service
in service
in service
26,200
-------
4-54
TABLE 4-17 (continued)
OPEN WELLS (continued)
Source Purveyor
Wells
Unknown City of Gilroy #4
(continued)
#6
#7
STAND-BY WELLS
Unknown SJWC
Ridgely #1
#2
#3
Contaminant
TCA
PCE
TCA
PCE
TCA
PCE
TCA
TCA
TCA
Concentration
(ppb)
ND - 0.5
ND - 0.61
ND - 0.5
ND - .79
ND - 0.5
ND - 1.2
2.3 - 2.5
2.2 - 3.8
1.3 - 1.8
Use
in service
in service
in service
stand-by
# of People
Exposed*3
4,300
(September, 1985)
Unknown SJWC
Springfield
PCE
ND - 0.5
last used 1981; stand-by
Unknown SJWC San Thomas #2
#5
PCE
PCE
<0.5 - 1.4
ND - 0.6
last used 1979; stand-by
IBM
SJWC
Senter
TCA
CFC-113
2.2 - 3.4
ND - 5.8
standby
4,500
(December 1985)
-------
4-55
TABLE 4-17 (continued)
Source Purveyor
Wells
Contaminant
Concentration
(ppb)
Use
CLOSED WELLS
IBM
IBM
Magic Sands
Mobile Home Park
Great Oaks
Water Company
Well #2
Well #8
TCA
CFC-113
TCA
CFC-113
TCA
ND - 2
ND - 1
ND - 2.6
ND - 12
ND - 17
out of service
March 1986
out of service
Sept. 1982
out of service
Feb. 1982
IBM or
Singleton
Road
Landfill
Carribee Mobile Home Park
Well #1
CFC-113
TCA
ND - 0.5
ND - 0.5
out of service
March 1986
Fairchild Great Oaks Well #13
Water Company
TCA
100
(varies)
out of service
Dec. 1981
-------
4-56
TABLE 4-17 (continued)
CLOSED WELLS (continued)
Source Purveyor Wells
Contaminant
Concentration
(ppb)
Use
Unknown SJWC First Street #1
#2
#3
TCA
2.8 - 10.6
TCA
CFC-113
TCA
1,1 DCE
PCE
CFC-113
2.
ND
14
ND
ND
ND
4 - 4.7
- 1.4
,6 - 37.
- 1.7
- 1.1
- 2.5
2
out of service, 1980
out of service, 1980
out of service, 1980
a Does not reflect blending of uncontaminated water fron other wells in the Tully, Evergreen, and Blossom fields;
nor at the City of Santa Clara wells.
b For the SJWC and the California Water Service Company, population estimates are based on extrapolating percentages
of well production for these wells to the population served by the water purveyor. For other purveyors, figures
are based on purveyors' estimates of population served.
c Reflects levels after treatment; levels before treatment range from 5.2 - 11.8 ug/1.
d Dibrcmo-2-chloro-2-fluorocyclopropane (DBCFCP) has been tentatively identified in this well. No standard exists to
confirm if this chemical exists, or to confirm that it is in the groundwater. DOHS is attempting to determine if any
health effects are associated with this chemical.
-------
4-57
TABLE 4-18
SUMMARY OF PRIVATE WELLS CONTAMINATED
WITH ORGANIC CHEMICALS
City
Mountain View
Number of
Contaminated Wells
40
Contaminants
TCA
CFC-113
TCE
PCE
1,2 DCE
1,1 DCA
Vinyl Chloride and
other compounds
Concentrations (ug/1
ND
ND
ND
ND
ND
ND
ND
24
520*
2,800
8.5
5.0*
1.1*
540
San Jose
11
Palo Alto
TCA
PCE
1,1, DCA
CFC-113
1,1 DCE
Chloroform
TCA
TCE
ND - 150
ND - <.05*
ND - .6*
ND - 430
ND - 3
ND - .7*
ND - .9*
ND - .9*
Sunnyvale
Los Gatos
56
TCE
PCE
ND -
ND - 8.9
1 Concentrations marked with * are from the Santa Clara County Health Department's
1985 monitoring program. All other concentrations are fron the County's 1984
monitoring.
-------
4-58
Exposure and Risks
We have calculated risks for the currently contaminated public wells which
remain in use. Table 4-17 includes exposure estimates average concentration
levels and populations exposed for those contaminated large public wells in
service, and for those wells available on a stand-by basis. Since the stand-by
wells are used infrequently, we have estimated risks only for those wells which
remain in regular service. The remaining contaminated public wells have been
removed from service, and therefore do not pose any risk. For comparative
purposes, we also present estimates of what aggregate risks might be if the closed
contaminated wells remained open.
The six contaminants in public wells which remain open are 1,1,1 trichloro-
ethane (TCA), l,l,2-trichloro-l,2,2-trifluoroethane (CFC - 113), 1,2 dichloroethylene,
DBCFCP, perchloroethylene (PCE), and carbon tetrachloride. EPA considers PCE
and carbon tetrachloride possible human carcinogens. Combining the exposure
information provided in Table 4-17 with our potency estimates, we estimated
risks for the PCE in the Gilroy wells and for carbon tetrachloride in the
California Water Service Company Hillview well. We have estimated risks as if
the concentrations in the Gilroy wells are consistently at 2.3 ppb, the highest
level in any of the wells. We estimated risks at the Hillview well as if the
concentration of carbon tetrachloride is consistently 1 ppb.
Cancer risks are summarized in Table 4-19. We estimate the lifetime
individual risk of cancer from current exposure at public wells at about two to
four chances in a million, and we estimate the increased annual incidence of
cancer at about one case in 800 years. The risk projections implicitly assume
70 years of exposure at current levels. We cannot confidently project future
exposure levels nor the likely directions of changes in concentrations at these
wells. We present our analysis of future drinking water risks from contaminated
groundwater in the next section.
EPA's Office of Research and Development has stated that there is insufficient
evidence to classify 1,1,1 trichloroethane (TCA) as a possible carcinogen;
therefore, our best guess is that cancer risk from exposure to TCA is zero.
However, because TCA is similar in chemical structure to otheir chlorinated
hydrocarbons suspected of being human carcinogens, and because scientists disagree
about the interpretation of laboratory studies of TCA, we have included a
sensitivity case to estimate risks from TCA as if it were a carcinogen.*
* If TCA were a carcinogen, we would conservatively estimate the average risk
among exposed individuals at about 7 x 10~8, or one chance in 10,000,000 of
contracting cancer. The total population exposed is about 121,000, and our
conservative estimate of the projected annual incidence about 1.2 x 10~^, or
one case every 10,000 years. This estimate assumes an average level of TCA equal
to the the highest levels in the Tully, Evergreen, Rancho Santa Teresa, Blossom
Hill, and Gilroy wells.
The highest level of individual risk from currently contaminated wells
could occur at the Magic Sands and Rancho Santa Teresa Mobile Home Parks, where
TCA concentrations range as high as seven ppb. To be conservative, we have
estimated risks as if the concentrations were consistently seven ppb. At this
concentration, the individual chance of contracting cancer fpcm TCA is 3 x
10~7, or about three in ten million, if TCA is carcinogenic.
-------
4-59
TABLE 4-19
AVERAGE CANCER RISKS TO EXPOSED INDIVIDUALS AND PROJECTED ANNUAL INCIDENCE
FROM EXPOSURE TO CURRENTLY CONTAMINATED WELLS (1)(2)
Average
Purveyor
Gilroy
Cal. Water
Service Co.
Total
Contaminant
PCE
Risk Among
Exposed Individuals
( Chances
in a million)
2
Carbon 4
Tetrachloride
Population
Exposed
26,200
2,600
28,800
Aggregate
Annual
Increase in
Incidence
.001 (one case every 800 years)
.0001 (one case every 10,000 years)
.001 (one case every 800 years)
Weight
of
Evidence 3
B2
B2
B2
(1) Sources: Exposure estimates are discussed in vext and decuranted in the Appendix.
Toxicological evidence is summarized in text.
(2) These estimates of individual risk and incidence are rough approximations of actual risk.
They are based on conservative potency estimates, and are more likely to overestimate risks than
to underestimate them. See text.
(3) The weight of evidence of carcinogenicity varies greatly among compounds. EPA has classified
PCE and carbon tetraeh]oride as B2, or probable human carcinogens.
-------
4-60
All levels of toxic contamination in open public wells in Santa Clara County
appear to be below the estimated thresholds for non-cancer effects. Therefore, we
estimate that there is no risk of non-cancer effects from currently contaminated
public wells, assuming that the concentrations remain similar to current levels.
Estimated Risk Reduction from Closing Contaminated Public Wells
Tb estimate the effectiveness of removing wells from service, we have
estimated cancer risk from public wells in the Santa Clara Valley which are
either out of service, oir which may be used on a seasonal stand-by basis, and
which include substances EPA considers possible carcinogens. These include
sane of the Springfield (PCE), First Street (PCE and 1,1 DCE), and San Thomas
(PCE) wells of the San Jose Water Company.
We estimated the populations potentially served by these wells by extrapolating
the percentage of water produced by these wells when they were last used regularly
to the current population served by the Company. These estimates were 0.35%
for Springfield, 0.25% for First Street, and 0.73% for San Thomas.
For simplicity, we assumed that the highest levels of 1,1, DCE and PCE found
in each system apply to the entire system. This tends to overestimate the
potential risks. We did not estimate risks for TCA, CFC-113, 1,2 DCE, or DBCFCP,
which are also in closed public wells, since EPA does not consider these substances
carcinogenic.
We estimated that closing contaminated public wells in the Santa Clara
Valley has reduced the increased incidence by about one case every thousand
years. Compared to risks from open wells, closing these wells has reduced risk
by about 50%.
Organic Chemical Contamination Elsewhere
Table 4-20 presents national data on contaminated public wells for PCE, TCA,
and carbon tetrachloride (data on the other substances in public wells were
unavailable). Contaminant levels in public wells in the Santa Clara Valley are
similar to those in contaminated wells elsewhere (i.e., most are in the low
parts per billion).
-------
4-61
TABLE 4-20
ORGANIC CHEMICAL CONTAMINATION IN DRINKING WATER SYSTEMS NATIONWIDE
1984
Percentage of Systems Above
Indicated Levels (ppb)
# of Systems
Contaminant Affected >0.5 5 20 50
PCE 1552 3.2% .7% .1% .01%
TCA 1390 2.9 .8 .2 .01
Carbon Tet. 336 .7 .2 0 0
1 or more VOCs* 5359 (11%)
These results include 48,458 public drinking water systems in the U.S. The
levels presented are influent levels they are the levels of organics measured
in the water entering the systems and do not necessarily reflect the levels of
organics after treatment.
The data presented in the table above are considered representative data and are
from five studies: National Organics Monitoring Survey (1977 data), National
Screening Program for Organics in Drinking Water, Ground Water Supply Survey
(State data - 1982), 1978 Community Water Supply Survey, and the Rural Water
Survey. For Carbon Tetrachloride, the data are also from the National Organics
Reconnaisance Survey.
* Number of systems affected by one or more volatile organic compounds at a
level greater than or equal to 0.5 micrograms per liter (ppb).
-------
4-62
Summary; Current Groundwater Contamination
Our estimates of the risks from current groundwater contamination of large
public wells in the Santa Clara Valley indicate that the risks are low, with a
conservatively estimated total annual increased incidence of about one case
every 800 years. All contaminants appear to be well below estimated thresholds
for non-cancer effects. The risks from current exposure in private wells,
which we have not included, may be higher since private wells are generally
shallower than public wells and are therefore vulnerable to higher levels of
contamination. In addition, private wells generally are not monitored, so
users could be exposed to drinking water contamination for long periods without
becoming aware of it.
Several factors contribute to the low level of estimated risks. Closure
of some public wells has reduced risks by about 50%. Clean-up of contaminated
groundwater has reduced the levels of contaminants in the groundwater, and thus
the contamination to which people are potentially exposed. In addition, some
contaminants may degrade into harmless substances in groundwater. For example,
there is no evidence of drinking water contamination by fuel constituents, such
as benzene, in Santa Clara Valley wells despite many known fuel leaks. Chemical
or biological degradation may be converting benzene and other constituents to
harmless forms before reaching drinking water wells.
-------
4-63
IEMP Methodology for Estimating Future Exposure to Groundwater' Contamination
from Underground Tanks'
The characteristics of groundwater make our estimates of future exposure
and risk from groundwater contamination the most complex and uncertain part of
the Stage I analysis. Future contamination levels are difficult to predict
since groundwater contamination is understood poorly and can take years to
appear. To estimate future exposure, we have developed a model linking the key
physical processes and regulatory activities affecting exposure. This section
briefly summarizes that methodology, and is intended to provide only a basic
understanding of the methodology. We also present exposure and risk estimates
for groundwater contamination from industrial and fuel sites. The Appendix to
this Chapter provides a more complete explanation of our groundwater exposure
methodology, and of the results.
Our analysis of future groundwater exposure is a generic analysis. It
does not model the probable impact of specific plumes throughout the Valley.
Instead, we have analyzed various factors affecting groundwater contamination,
including different sources, sizes of releases, and hydrogeologic characteristics.
We have attempted to determine the classes of drinking water wells and hydrogeologic
zones at greatest risk, and the classes of sources posing the most significant
problems. This generic approach is intended to assist us in setting priorities
among problems, and not to design solutions for particular contamination incidents.
Simulation Approach; From Sources to Human Exposure
Our framework for estimating exposure from underground tanks tracks contamination
from initial sources, through the fate and transport of contaminants in the
groundwater, to drinking water contamination and human exposure. We characterized
the important steps in this process and the links among them. We also accounted
for remedial and protective activities, such as public well monitoring, implementation
of Hazardous Materials Management Ordinances (HMMOs), and current clean-up.
Where possible, we based this analysis on existing laboratory or field data; in
many instances, however, we relied on engineering assumptions and expert professional
and scientific judgment.
Our methodology for estimating future exposure from groundwater contamination
is shown schematically in Figure 4-9. This diagram shows the sequence of steps
which must occur before a leak or spill results in exposure through drinking
water. We can group the major factors affecting the movement of contaminants
fron sources to human recipients in the following way:
1. Contamination sources: tank numbers, locations, failure rates;
contaminants;
2. Releases fron sources: frequency, duration, and volume;
3. Plume Characteristics: sizes, concentrations, and distribution;
4. Transport of contaminants in plumes: groundwater flow; contaminant
retardation, dispersion, and transformation; aquitards affecting vertical
flow;
5. Well Characteristics: locations with respect to sources of contamination;
role of conduit wells; and
6. Monitoring and Intervention Practices at wells.
-------
4-64
FIGURE 4-9 OVERVIEW OF IEMP ANALYSIS OF FUTURE
EXPOSURE TO GROUNDWATER CONTAMINATION
MONITORING
INTERVENTION
Fuel Tanks
Industrial Tanks
Landfills
Frequency
Duration
Volume
HMMOs
Sizes
Concentrations
GW Flow
Dispersion
Retardation
Transformation
Distance
Distribution of Sources
Distribution of Wells
Conduit Wells
Dilution
Public Well Monitoring
Public Well Closure if
Violates Standards
FUTURE
EXPOSURE
ASSESSMENT
-------
4-65
Vfe first identified potential sources of contamination. We then estimated
the frequencies of releases from these sources. In the third step, we linked
releases to contaminant plumes. Wa then modeled the transport of contaminants
from sources to vulnerable wells. We then estimated the number of wells the
contamination was likely to affect. Finally, we accounted for monitoring and
intervention practices, such as closing wells when contamination reaches certain
levels, before estimating the extent of exposure.
Analysis Organized Around Important Variables
Many factors could affect the processes just outlined. Wa have organized
our analysis around what appear to be the most important variables. The most
important ways that we disaggregate the analysis include:
o Pre-HMMO vs. Post-HMMO; We assumed that full implementation of the
hazardous materials ordinances* reduces the frequency, duration, and
volumes of leaks and spills. We estimated risks from releases before
implementation, and also estimated risks from releases after implementation.
In addition to implementation of the ordinances, this distinction
represents other important factors, such as replacement of many tanks,
changes in the substances contained in tanks, and a substantially
longer time frame for considering potential post-HMMO releases.
o Hydrogeologic Zones; Hydrogeologic characteristics vary within the
Valley. We accounted for differences in groundwater flow rates, well
locations, and other factors.
o Source Type; We accounted for different numbers of fuel and solvent
tanks, and the different behavior of the contaminants. We also
analyzed other sources, including above-ground tanks, illegal
disposal, sewer lines, septic tanks, and sanitary landfills.
o Public vs. Private Wells: We accounted for different monitoring
requirements, depths, and locations of private and public wells.
Vfe projected the potential for a source to result in drinking water
contamination for cases representing each of the above categories (e.g.. We
analyzed pre-HMMO industrial releases in the Recharge Zone separately from similar
releases in the other hydrogeologic areas and from fuel releases in the Recharge
Zone). Our estimates of key variables, such as contaminants of concern and rates
of contaminant travel in groundwater, differed for these cases. Wa combined
estimates of risks in each of these separate cases to arrive at an estimate of
aggregate risk for the Santa Clara Valley.
* The implementation of the Hazardous Materials Management Ordinances (HMMOs)
is important in understanding the analysis. Adopted by the County and most
cities, the HMMOs require industry to include a secondary containment system (a
second tank wall, or a lined vault for the tank to catch leaks) in all new storage
facilities, and to monitor between containment systems. The HMMOs require existing
facilities to monitor groundwater near their tanks to detect contaminant releases.
All facilities must submit plans for managing their hazardous materials, report
suspected releases, and clean up contaminants if a release occurs. We have
assumed that the HMMOs will be fully implemented by the end of 1987 on average.
As discussed below, we do not assume that full implementation will result in
complete compliance or effectiveness.
-------
4-66
Alternative Assumptions: Sensitivity Analysis
We have also examined a number of alternative assumptions to assess their
importance in analyzing important but uncertain variables. For example, we
used various assumptions about the sizes of leaks and the effectiveness of the
HMMOs.
We briefly describe some of the key assumptions below. Where significant
uncertainty exists, and where we examine alternative assumptions, we generally
designate one set of assumptions as our base case. Table 4-22 at the end of
this section lists many of the factors in the analysis, and roughly indicates
both the importance of these factors to the analysis, and the degree of confidence
we have in them. For more detailed discussion and documentation please refer
to the Appendix.
Overview of Methodology
Time Considerations
Our analysis of future exposure and risk frcm groundwater contamination
accounts for exposures over the next 70 years. This 70 year factor is standard
in EPA risk assessments, and is consistent with our analysis of risks frcm air,
surface water, and current exposure to groundwater contamination. We are not
projecting population growth and industrial development, but rather are estimating
risks from current conditions over 70 years. It is important to note that in
some parts of the Santa Clara Valley groundwater contamination may not result
in exposure for years and that exposure can continue for decades. Therefore,
exposure to future contamination may continue beyond the 70 year limit of this
analysis.
1. Sources
As outlined in Figure 4-9, the first step in our analysis is identifying
potential sources of contamination, their locations, and the contaminants of
concern. We identified underground tanks in the following ways:
Pre - HMMO Industrial: We applied data from the Regional Water Quality
Control Board (RWQCB) to identify 220 underground industrial tanks that contain
substances of concern that we modeled as potential sources before implementation
of the HMMO. Some of these tanks have since been removed.
Post - HMMO Industrial: We based post-HMMO tank populations on the State
Water Resources Control Board's (SWRCB) underground tank data base. These data
reflect tanks in place as of about 1983. Industry has removed many industrial
tanks which were used in 1982 or 1983. We accounted for the removal of tanks
described in RWQCB records, and for some additional removals reported by industry.
We also assumed that tanks projected to fail in 1985 and 1986 would be removed.
In a sensitivity analysis, we assumed that removal patterns for a sample of ten
electronics firms were typical of practices for large electronics and semiconductor
companies.
Fuel tanks: We used data from the Bay Area Air Quality Management District
to estimate that 6,453 fuel tanks exist in Santa Clara Valley.
-------
4-67
Location: Vfe based our estimates of the distribution of industrial tanks
across hydrogeologic zones on the RWQCB's maps. We did not model industrial
sites that may be located in South County because almost all underground industrial
tanks are located in the industrialized northern part of the County. We assumed
that fuel tanks, and therefore fuel leaks, are distributed in proportion to the
population in each hydrogeologic zone. Our estimate of the proportion of tanks
in each hydrogeologic setting is shown below:
Zone
% of sites- Industrial
% of sites- Fuel
Inland Confined
Recharge
Bay Sands
Bay Muds
South County
17%
15%
43%
25%
0%
14%
37%
18%
18%
13%
Contaminants Analyzed
Pre - HMMO: To characterize groundwater contamination from industrial tanks
before full implementation of the HMMO, we identified the twenty most ccmon
groundwater contaminants. We have analyzed nine contaminants, based on their
characteristics in the environment and their toxicity:
1,1,1 Trichloroethane (TCA)
Trichloroethene (TCE)
Chloroform
Methylene Chloride
Benzene
- Vinyl Chloride
- Perchloroethylene (PCE)
- Toluene
-1,1 Dichloroethylene (1,1 DCE)
Post-HMMO: We selected substances for analysis after the implementation of
the HMMO from data in the State Water Resources Control Board (SWRCB) tank data
base. We excluded substances considered relatively immobile or non-toxic. We
modeled six substances in two ways. First, in the base case we assumed that
they were present only to the extent indicated by the data (with adjustments for
tank removal). Since we did not have enough information to exclude many additional
substances, we conducted a sensitivity case in which we adjusted our estimates
of the presence of the six substances upward by a factor of 4.26. Based on the
SWRCB data, there are 4.26 times more tanks containing substances which may be
of concern than tanks containing the substances we modeled. The substances we
modeled are:
- 1,1,1 Trichloroethane (TCA)
- Methylene Chloride
- Xylene
- Toluene
- Methyl Ethyl Ketone (MEK)
-1,1 Dichloroethylene (1,1 DCE)
We analyzed potential releases of benzene and ethylene dibrccnide (EDB)
from fuel tanks. We focused on benzene and EDB as being the most toxic, soluble,
and mobile gasoline constituents. We also examined toluene, xylene, lead, and
undisolved gasoline, but our preliminary analysis indicated that these substances
are probably not as signficant as benzene and EDB. In addition, we modeled releases
of toluene and xylene from industrial sources.
-------
4-68
2. Releases
The second step in our analysis was estimating releases from the sources, as
shown in Figure 4-9. We analyzed three aspects of tank releases: location,
frequency, and volume. We assumed that future releases from industrial tanks
would follow the geographic pattern of past releases, and that releases from fuel
tanks would follow the population distribution. We assumed the frequency and
volume of releases varied across different analytic subcategories (i.e., pre
and post-HMMO releases, fuels and industrial). We conducted sensitivity analyses
using both low and high estimates of release sizes.
Release Frequency: We estimated the frequency of releases in two ways.
For pre-HMMO industrial releases, we relied primarily on the RWQCB's field data.
These data presumably reflect the combined effects of tank and pipe leaks, and
spills. For other releases (post-HMMO industrial and pre and post-HMMO fuel)
we relied on a simple tank release model based on field data and engineering
estimates. We discuss this model in the Appendix. We conducted sensitivity
analysis on the number of tanks to account for a lack of data on many pollutants.
We assumed that the rate of spills and other similar small releases would
continue after implementation of the HMMO at the same rate as before implementation.
While we expect the frequency of spills at industrial facilities to decline as
handling practices improve, we have attempted to account for other potential
sources such as illegal disposal by assuming that the rate of these relatively
small releases would continue at the same rate in the future.
Ws also analyzed releases associated with above ground tanks, urban runoff
to dry wells, sewer lines, septic tanks, and sanitary landfills. We did not
develop quantitative estimates of risks from these sources because preliminary
analyses of releases and exposure indicated that risks would be be minor compared
to risks from the other sources we analyzed. Details are presented in the
Appendix.
Release Volumes: The volume of contaminants released from a leaking tank
depends on two factors: tank release rates (how much of a contaminant leaks
over a period of time), and the duration of the leak. For most of the analysis,
we estimated the number of sites caused by leaks from tanks, and the number
from spills. This distinction is important because leaks generally release
larger volumes than spills. The volume released through spills depends on the
amount spilled at one time.
Because the assumptions on the volumes released are important and very
uncertain, we developed "high" and "low" release estimates for each category
of releases (pre-HMMO fuels and solvents, post-HMMO fuels and solvents) by
applying different assumptions on leaks. Wa combined judgments about release
rates with estimates of the delay before detection, and hence duration, of the
releases. Table 4-21 presents details on this part of our analysis.
We have assumed that HMMO implementation will reduce the duration of leaks
(and hence volumes) substantially. We expect that handling requirements, and
greater operator awareness, will reduce the frequency of spills after HMMO
implementation. It is important to note that we did not assume that implementation
of the HMMOs results in complete compliance with the Ordinance, nor that it
will be effective completely even for facilities that do comply. In some
cases, we assumed that it would be 95% effective, while in others we assumed
-------
4-69
TABLE 4-21
ASSUMED TANK RELEASE VOLUMES
(Base Case)
Reason
Release Release foir Leak Release % of
Rate (gph)-*- Duration Discovery Volume (gallons)^ All Releases
Pre-HMMO
Industrial - Low Estimate Based Primarily on Local Field Data See Text
100%
Industrial - High Est.
Spills - Hose Failure
Spills - Tank Overfill
Fuels - Low Estimate
Fuels - High Estimate
Post-HMMO
Industrial - Low Est.
& Fuel
Industrial - High Est.
& Fuel
0-0.1
0.1
-
-
0-0.1
0-0.1
0-0.1
0.1
0-0.1
0-0.1
0.1
0-0.1
3 years
4 years
-
-
3 years
4 years
3 years
4 years
6 months
3 years
70 days
4 years
Various
Various
-
-
Various
Various
Various
Various
On-site
Monitoring
Back-up3
On-site
Monitoring
Back-upb
788
2,900
50
150
788
1,400
788
2,900
50
788
140
1,402
42
2
28
28
100%
95
5
100%
95
5
100%
88
12
100%
88
12
100%
* Vfe applied two release rates: A gradual leak, starting at zero and increasing
to 0.1 gallons per hour over five years (designated as 0 - 0.1 gph); and
a steady leak of 0.1 gallons per hour, or roughly two gallons per day (0.1
gph).
^ Volumes do not reflect immediate clean-up activities, which reduces the amount
of contaminants reaching groundwater. See the Appendix for details.
3 "Various" and "Back-up" reasons for discovering leaks include tank testing,
inventory control, and on-site construction and maintenance activities.
-------
4-70
that it would be only 75% effective. In addition, even when we consider the
ordinance effective we still expect to have some small releases. We believe
that the estimate of 75% effectiveness is quite pessimistic.
The estimates of the volume of contaminants released firom tanks are quite
uncertain. A cursory review of recent field data indicates that most leaks
leaks (perhaps 95%) are smaller than those we have estimated. However, in at
least one recent case (at the IBM site in south San Jose) a leak released about
8,000 gallons of kerosene. These recent field data suggest that, overall, our
estimates of the sizes of releases are reasonable.
3. Plume Characteristics
In this section, we summarize our methodology for estimating the "initial"
size and concentration of plumes. Vfe subsequently modeled the movement of these
plumes. To characterize initial plumes, we combine the source and release data
just discussed with hydrogeologic data. For pre-HMMO solvent plumes, we also
analyzed a case in which we applied field data on actual plumes.
Plume Sizes: Wa had little available data on existing plume sizes, so we
relied extensively on local data on several industrial sites in each hydrogeologic
zone, additional local data on hydrogeology, and professional judgment to estimate
initial plume sizes.
Since plume sizes depend largely on groundwater velocity, and since ground-
water velocity varies among hydrogeologic settings, plume sizes vary among
hydrogeologic settings. We assumed that larger plumes would occur in the hydro-
geologic zones with faster groundwater flow rates. Initial plume lengths varied
from less than 400 meters in the Bay Sands zone to almost 5,000 meters in the
Recharge zone.
In each hydrogeologic zone, we specified a small and a large plume. We
assumed that each plume consists of an inner, higher-concentration area surrounded
by a larger, lower-concentration area.
Initial Plume Concentrations: Our groundwater transport model required the
specification of initial plume sizes and concentrations. Wa estimated plume
concentrations in most cases by applying our model on releases and engineering
analysis on hydrogeologic factors, and in other cases by applying limited field
data (for the pre-HMMO industrial low release case).
Because groundwater flow rates and the size of plumes vary across hydrogeologic
zones, the same size release results in plumes with substantially different
concentrations among zones. For example, concentrations in the Recharge Zone
(where we estimated groundwater flow at 3,500 ft/yr) are substantially lower
than those in the Bay Sands Zone (where we estimated groundwater flow at 60
ft/yr), if the same volume of contaminants is released.
Wa accounted for adsorption in estimating plume concentrations. Adsorption
is the prrr-'sss of chemicals bonding to organic materials in saturated soil. It
prevents sorse contaminant itsass from dissolving in groundwater, and therefore
slows the movement of contaminants within groundwater and reduces our initial
concentration estimates. However, the duration of exposure is greater since
some of the contaminants eventually separate from the soil and remain in the
-------
4-71
groundwater. We account for adsorption by extending the period of exposure
while reducing average concentrations.
Sane portion of releases will volatilize, transform, or became immobile in
the unsaturated soils between the surface and groundwater, known as the vadose
zone. The vadose zone can also delay the release of contaminants to the groundwater.-
so that even sudden releases may reach groundwater over several years. Based on
limited local data, we estimate that a 25 foot vadose zone of soils with some
organic material exists in the Inland Confined, Recharge, and South County zones.
We estimate that this zone could eliminate 50 to 90% of the mass of a contaminant
released just below the soil surface. We conservatively estimate a 50% removal
of contaminants in the vadose zone in these zones. We assume no loss in the Bay
Sands zone, due to different soil characteristics and a higher water table.
In our second approach to estimating plume concentrations (used for risk
estimates only in the pre-HMMO industrial case), we used the limited available
field data from a sample of industrial sites to estimate plume concentrations.
Since we based these concentrations on plumes from pre-HMMO industrial leaks,
we used this case as our pre-HMMO industrial low release case. Groundwater
clean-up activities have substantially affected these plume concentrations,
although we have not accounted explicitly for the impact of plume clean-up.
However, comparing this case with our modeled estimates (which do not account
for plume clean-up) allows us to make rough comparisons of the potential exposure
and risk reductions from groundwater clean-up activities. This case is best
interpreted as representing a small leak or spill, or a larger release that has
been partially cleaned up.
4. Contaminant Transport and Transformation
Thus far, we have discussed our assumptions about how a set of sources and
releases result in plumes. In this section, we discuss how those plumes may
move and change.
The movement (transport) and change (transformation) of contaminants depend
on four factors:
1) Groundwater flow rates;
2) Dispersion;
3) Retardation (the speed a substance travels relative to the speed of the
water); and
4) Transformation (changes in a substance due to chemical or microbiological
processes).
We used a simple computer model to analyze plume movement. The transport
nodel predicted plume movement and change, accounting for groundwater flow,
dispersion, and retardation. (We analyzed transformation separately.) The
model estimated concentrations over time at various distances from the plume.
We modeled transport in all hydrogeologic zones except the Bay Muds, where
the slow flow rates make it less likely that releases will have significant
effects than releases in other zones. In addition, this area is less likely to
be used as a source of drinking water. Since we have more data on the Southeast
Recharge zone than the other zones, we modeled transport in this zone somewhat
differently. We also analyzed conduit wells and their impact on the lower
-------
4-72
aquifer in the Bay Sands and Inland Confined zones. We discuss this analysis
later.
Dispersion: Contaminants tend to spread over a wider area and plume concen-
trations tend to decline over time. This spreading tendency is know as dispersion.
Dispersion affects both the length and width of plumes, and hence the speed at
which plumes travel, the area they cover, the number of wells they affect, and
the concentrations of contaminants at affected wells. Since we have little
local evidence on dispersion, we have relied heavily on data from other areas.
We assume that dispersion in Santa Clara is similar to dispersion in similar
aquifers elsewhere.
In the Southeast Recharge Zone, we assumed ten times less lateral dispersion
than in other zones because of groundwater "channeling". This effect occurs
when groundwater flows in narrow channels, much like a stream. Groundwater
channeling may be caused by underground stream beds or other geologic conditions.
Retardation: Contaminants may move more slowly than groundwater due to
interaction with soils in the saturated zone. This effect is called retardation.
A chemical with a retardation factor of two would move half as fast as groundwater,
and a chemical with a retardation factor of ten would move one-tenth as fast.
Retardation depends on both soil and contaminant characteristics. Wa have
very few data on key soil characteristics in Santa Clara Valley, so we applied
data for the Palo Alto Baylands area to all hydrogeologic zones, except the
Southeast Recharge zone, where we accounted for observed contaminant transport.
We modeled contaminants with estimated retardations between one and nine
with a retardation factor of two. We modeled contaminants with estimated
retardations between ten and eighteen with a retardation factor of ten. We did
not model contaminants with higher retardation factors.
In the Southeast Recharge zone, we adjusted our estimates of retardation
downward to account for different soil characteristics. This increases the
rate at which contaminants travel in this zone and affect wells, and increases
the concentration of the plumes;. We also conducted sensitivity analysis by
analyzing a case in the Southeast Recharge zone in which we assume no retardation
(this assumption implies that there is very little organic carbon content in
this area, and that all contaminants would move at the same speed as the
groundwater).
Transformation: Chemical and microbiological processes can transform
contaminants in the subsurface, but predicting these transformations is extremely
difficult. Many substances might transform into other substances in a few
years, or in hundreds of years. They could become harmless daughter products,
or more hazardous substances. We assumed that most substances do not transform.
While they could degrade into harmless daughter products, we did not wish to
underestimate risks by assuming that they would.
For EDB and benzene, however, stronger evidence exists that transformation
to harmless daughter products was likely and that our base case analysis should
reflect this. For TCA, transformation to a more harmful substance (1,1 DCE) is
sufficiently likely that we have analyzed that as our base case. Vfe examined
both transformtion and non-transformation cases for these three substances.
-------
4-73
Chlorinated solvents, such as methylene chloride, perchloroethylene, and
trichlorethylene, can be tranformed chemically or biologically into other
chlorinated compounds. Field evidence from Santa Clara Valley indicates the
presence of 1,1 DCE and vinyl chloride where they were never used, but where
other chlorinated compounds were used. Since predicting the likely transformations
is highly uncertain, we have modeled most chlorinated compounds as if they do
not degrade. This accounts, albeit crudely, for the possible transformation
frcm one hazardous substance to another, as well as for the possibility of slow
transformation. This assumption makes it unlikely that we significantly underestimate
risks from these substances. Overall, we account for the conservative, plausible
case for each substance.
5. How Plumes Affect Wells
Thus far we have discussed how sources and releases result in contaminant
plumes, and how those contaminants move and change over time. This section
discusses the next link in the chain: how plumes affect drinking water wells.
The extent of exposure from groundwater contamination depends on the distribution
of sources and wells, natural hydrogeologic barriers protecting some wells,
dilution at the well, and monitoring and intervention practices.
Transport Distances and Directions: We considered all wells vulnerable to
fuel leaks, but identified particular downgradient public wells as vulnerable
to industrial leaks. In the Bay Sands, Inland Confined, and South County zones
we assumed that plumes could significantly affect wells up to a mile from a
source. We modeled plumes in the Northwest Recharge zone for two miles, and
those in the Southeast Recharge zone for six miles because of higher flow rates
and groundwater channeling in these zones. We assumed that wells near the edge
of the Inland Confined Zone could be affected by plumes from the Recharge zone.
Our transport model predicted plume concentrations at 200, 1,000 and 3,000
feet frcm the leading edge of the initial plume. In the Southeast Recharge zone
we also modeled plumes in a well two miles downgradient. We assumed that wells
falling between two of the modeled distances were affected as if they were at
the closer distance from the source. Thus, in all zones but the Southeast
Recharge zone, we modeled public wells within 1,000 feet of a source as 200
feet away, wells within 1,000 feet and 3,000 feet away as 1,000 feet away, and
wells over 3,000 feet away as 3,000 feet away.
Overlaps of Plumes and Wells: We conservatively assumed that a plume would
contaminate a well if it reached any spot under the well, regardless of plume
depth or well depth or construction. The only exception was in the Bay Sands
and Inland Confined zones, where we assumed that plumes from sources within
these zones affect large public wells only if the well was poorly constructed
or a conduit well intervened.
Assuming that wells are affected by any plume underneath them tends to
overstate exposure since some public wells draw exclusively frcm zones deeper
than nearby plumes, and some private wells tap zones shallower than contaminated
zones.
Dilution at the Well: Contaminants can be diluted by clean water during
punping if the well also pumps uncontaminated water. Dilution will typically
be greater in higher yielding wells. In our base case, we assumed no dilution
-------
4-74
of plumes in private wells, slight dilution in small system wells, and higher
dilution in large public wells. Dilution in public wells in the Inland Confined
and Bay Sands zones is affected by our conduit well assumptions, discussed
be low.
Conduit Wells; large public wells in the Bay Sands and Inland Confined
zones are perforated only in the lower aquifer, so we assumed that contaminants
can reach these wells only if they are poorly constructed or are near conduit wells
piercing the clay layer separating the upper and lower aquifers (except for
wells near the Itecharge zone). See Figure 4-10 for an example of how conduit
wells may affect drinking water.
We accounted for two sources of conduit wells:
1. Past irrigation practice in parts of the Valley included flood
irrigation of relatively small areas, with deep wells perforated in
all aquifers. With twenty to forty acre parcels, approximately 20
wells per square mile would have been necessary. In our base case, we
assume conduit wells exist throughout the Inland Confined and Bay
Sands zones at this density.
2. About 18% of the large public wells in the Bay Sands and Inland
Confined Zones were constructed with gravel packs along their sides
before construction ordinances came into effect in 1975. Gravel
packs on these wellsf and on some wells constructed since 1975, may
extend through the confining layer into the lower aquifer. Some older
public wells may also have deteriorated, creating potential conduits.
Combining these wells, we assume that 20% of existing wells may act as
their own conduits. Other conduit wells (i.e., deep private wells)
may also exist, but probably are less significant than these two kinds
of conduits.
Contaminants reaching the lower aquifer will be diluted substantially by
the uncontaminated water in the lower aquifer, thereby reducing exposure from
contaminants transported via conduit wells. Because of this dilution, we
assume that conduit wells and drinking water wells must be within 300 feet of
each other for contamination :;;av; a conduit well to affect a drinking water
well. To account for dilution in conduit wells, the lower aquifer, and drinking
water wells,, we examined cases with dilution measured at factors of 50 in our
base case. This factor is the ratio of the amount of a contaminant to the amount
of water likely tc cascade down a conduit well, and the amount of water pumped
by a typical large, public well* We also examined an alternative, lower dilution
factor of ten, to simulate dilution at an intermittent or lower-flow public
well.
The SCVWD i& identifying and sealing some potential conduit wells near
contamination sites. While these efforts may reduce the potential for contaminants
to reach the lower aquifer, we have not accounted for them in our analysis.
6, Well Monitoring And Intervention
We have explained how we characterized sources, releases, plume character-
istics, contaminant transport, and plume-well interactions to determine the
levels of contamination reaching wells. The final step in our groundwater
-------
UPPER
AQUIFER
CLAY
LOWER
AQUIFER
PUBLIC
WELL
PRIVATE
WELL
CONDUIT
WELL
PUBLIC
WELL
FIGURE 4-10 HOW PLUMES AFFECT WELLS IN THE BAY
SANDS AND INLAND CONFINED ZONES
Public drinking water wells drawing from the lower aquifer are generally protected from contamination
by the clay layer.
When located near a conduit well, which may transport contaminants to the lower aquifer, public wells
are vulnerable to contamination.
Private drinking water wells drawing from the upper aquifer are vulnerable to contamination.
-------
4-76
exposure analysis is determining the duration and levels of exposure from
contaminated wells. Once we know how much of a contaminant individuals and
populations are exposed to, and for how long, we can combine this information
with our potency estimates to calculate risks.
Well monitoring and intervention practices are important in determining the
extent of exposure. If a well is closed immediately after it is contaminated,
the exposure from that contamination is minimal. If, however, a contaminated
well remains in use, then the exposure from that contamination may be significant.
Our assumptions on well monitoring and intervention practices are key factors
in estimating the duration and levels of exposure.
Santa Clara Valley Practices: The California Department of Health Services
recently required large public water purveyors to monitor their wells for a
wider range of organic pollutants than those for which purveyors historically
test. Purveyors have completed two rounds of extensive monitoring for these
organic priority pollutants, and DOHS will require at least annual testing.
Future testing will include most of the pollutants we examined in Stage I.
We assume annual testing of large public systems. Santa Clara Valley purveyors
have all stated their intention to close any well exceeding state action levels
(Great Oaks Water Company has a policy of closing any well with detectible levels
of contamination). We therefore assume that purveyors will, on average, close
wells as soon as they detect contamination over action levels. In the Recharge
zone, detection may occur after contamination exceeds standards because plumes
move quickly. For other zones, purveyors should detect plumes before contamination
reaches standards. DOHS requires certification of all laboratories analyzing
monitoring samples. While laboratory errors may occur occasionally, we assume
that all monitoring results will be accurate.
Small public wells are monitored infrequently, and for fewer substances than
large public wells. DOHS and the County will probably monitor all small systems
in Santa Clara once under A.B. 1803 in 1986 or 1987. Closure requirements
for these wells are the same as for large public wells.
Private wells are not regularly monitored. However, the HWQCB and Santa Clara
County tested 171 private wells in late 1985. The County, under a pilot program
funded by DOHS, will be testing about 1,000 - 1,500 private wells for a range of
organic pollutants, including industrial contaminants, in 1986 and 1987. The
Gavilan Water Conservation District is also instituting a program to sample selected
wells in areas of potential concern. Our analysis does not account for these programs,
and may therefore overstate the degree of likely exposure and risk at private wells.
Contamination can also be detected by taste or smell when levels are high
enough. Our estimates of taste and odor thresholds are as follows:
Benzene 24 ug/1 (ppb) PCE 300
Toluene 24 1,1 DCE 800
Xylene 50 Methylene
TCA 50,000 Chloride 3,200
TCE 500 Chloroform 100
EDB 50,000 MEK 1,000
We modeled small systems and private wells as if they were not monitored,
and assume contamination is detected and wells closed only when concentrations
reach taste and odor thresholds.
-------
4-77
TABLE 4-22
FACTORS IN ESTIMATING FUTURE EXPOSURE
TO GROUNDWATER CONTAMINATION FROM UNDERGROUND TANKS
Factor
Importance
Confidence
Tank Population (# Tanks)
Tank Locations
Missing Sources
Contaminants Addressed*
Releases
- Frequency
- Volume*
- Vadose Zone Effectiveness
- Soil Clean up
High
Medium
Low
High
High
Medium
Medium
Low
Medium
Medium
Medium
Medium
Medium
Low
Low
Low
Plumes: Initial Sizes
Transport
- No exposure in Bay Muds
- Effectiveness of Confining
Layer
- Dispersion
- Retardation
- Transformation*
0 Fuels
0 Industrial
- Plume Directions
Distances Between Sources/Wells
Plume/Well Overlap
Dilution in Lower Aquifer*
Conduit Wells
Dilution at Well
Monitoring
Well Closure
- Public
- Private
Medium
Medium
Low
High
Medium
Medium
Medium
High
High
Low
Medium
Medium
Low
Medium
Medium
High
High
High
Medium
Medium
Medium
Medium
Low
Medium
Medium
Low
High
Low
Medium
High
High
Medium
* We have conducted sensitivity analysis for these factors.
-------
4-78
Future Risks fron Groundwater Contamination; Results and Conclusions
This section discusses the IEMP estimates of risks from future exposure to
giroundwater contamination. We discuss our methodology for making these future
exposure estimates above. In this section, we explain our presentation of the
results, discuss the results and conclusions of QUIT analysis, discuss the
importance of various key factors in our analysis, identify important areas for
further research, and discuss some of the uncertainties in this analysis.
We present conservative estimates of the risks for cancer and other health
effects from groundwater contamination for the entire Santa Clara Valley. In
addition to presenting overall groundwater risk estimates, we present estimates
of risk disaggregated in the following ways:
0 by source type (fuels or industrial);
0 by pollutant (benzene, tsichloroethane, etc.);
0 by hydrogeologic zone (Recharge, Bay Sands, etc.);
0 by well type (private or public).
The complexities in this analysis make it infeasible to present in this
report detailed results for all combinations of assumptions. Instead, we present
disaggregated results only for our base case, which we describe in the previous
section. The one assumption varied within the base case is the estimate of
release sizes and plume concentrations: the base case includes both "high" and
"low" release estimates. This is one of the more uncertain, and important,
assumptions in our analysis, and is presented in most of the tables.
We have also summarized the results of the sensitivity analyses for several
other assumptions: 1) transformation and degradation, 2) dilution via conduit
wells, 3) the carcinogenicity of TCA, and 4) substances included in the analysis.
We do not present sensitivity analysis of carcinogenic potency estimates explicitly
but these can be inferred by noting the weight of evidence attached to the
potency estimates, and by recalling that, in most cases, the lower-bound estimate
of risk would be zero.
We present several types of risk estimates for cancer: risk to the maximally
exposed individual (MEI), the average individual risk among those exposed, and
the aggregate increased incidence of disease. For effects other than cancer,
we have estimated the number of people potentially exposed and the levels of
exposure (but not likely incidence or probability of disease) where contamination
may exceed thresholds. The combined figures on the tables represent the highest
risks for the maximally exposed individuals, the average risks to all exposed
individuals, the totals for the estimated populations affected and increased
incidence of disease.
Future Risks; Base Case Results
1. Our analysis of future cancer risks from groundwater contamination
indicates that most of the individual risks and the annual incidence will be
low by comparison with the estimated risks from other drinking water sources
and from air toxics. We also estimate that few people are likely to face
exposure to levels of groundwater contamination posing risks of effects other
than cancer. While we recognize the uncertainty of our estimates, the low
projected risks under a range of pessimistic assumptions and sensitivity analyses,
detailed below, suggest that this conclusion is fairly solid.
-------
4-79
Estimated risks reflect exposure patterns in air and drinking water.
About 10% of Santa Clara Valley's population is currently exposed to groundwater
contamination in public wells. Most of this exposure is to substances which
EPA does not consider carcinogenic. We estimate that an additional 15% of the
population could be exposed to groundwater contamination in the future. (This
estimate does not account for possible overlap between current and future
exposure.) By conparison, about 50% of the population is exposed to contaminants
in imported surface water, and 100% is exposed to air pollution. We expect
that most exposure to toxics in groundwater will be to low levels of contamination.
Under our base case assumptions, we estimated a possible future increased
cancer incidence of about one case every 15 to 30 years from groundwater con-
tamination in the Santa Clara Valley. This estimate is significantly lower
than the estimates of risk from toxic air pollution and from iitported surface
water. This estimate is higher than the estimates of risk from current exposure
to groundwater contamination in public drinking water wells (about one case
every 800 years). Current risk estimates are based on monitoring of large
public wells rather than on conservative modeling (which was necessary to
estimate future risks), and do not account for potential exposure in the more
vulnerable private wells.
Average individual risks from future groundwater contamination appear
lower than the comparable risks to typical exposed individuals from air toxics
and from imported surface water. However, our estimates of possible future
risks for maximally exposed individuals some people drinking from some
private wells are the highest in our Stage I analysis.
While we have accounted for many factors in our groundwater risk analysis,
several appear particularly important. We can divide these factors into two
categories natural physical factors; and regulatory programs and voluntary
actions. Important natural physical factors include the degradation of some
contaminants in groundwater and the Santa Clara Valley's hydrogeology. Key
regulatory programs and voluntary actions include preventive, clean-up, and
monitoring efforts. We discuss the importance of these factors later.
Future Risks by Type of Source
2. Underground industrial tanks appear to be the most important source of
risk from future groundwater contamination. Table 4-23 presents our estimates
of the individual risks and aggregate incidence from future exposure to fuel
and industrial contaminants. Although many more fuel tanks (about 6,400) than
industrial tanks (about 500) exist in the Santa Clara Valley, we estimated that
industrial contaminants will pose about 86% of future aggregate risk from
groundwater contamination. This is due largely to the assumed rapid degradation
of fuel constituents in the groundwater. If benzene and EDB do not degrade,
they could pose risks about equal to those from industrial contaminants. We
discuss degradation in greater detail below.
Several important similarities exist between fuel and industrial groundwater
contamination in Santa Clara: the sources of primary concern (underground
storage tanks); hydrogeology and transport considerations; preventive regulations;
and vulnerable wells and populations. However, there are also important differences
between fuels and industrial contaminants primarily in their behavior in
-------
4-80
groundwater and the number of sources so it is useful to distinguish between
the risks from fuel and industrial sources.
We incorporated estimates of risks from spills and illegal disposal of
contaminants in our analysis of releases from underground tanks. While we have
not quantitatively compared risks from these releases to risks from leaks from
underground tanks, we assumed that spills and illegal disposal occur frequently.
Under this assumption, they appear to contribute significantly to the number of
wells and people exposed to groundwater contamination. However, these smaller
releases appear less significant than leaks from tanks.
We also analyzed various other sources of toxic groundwater contamination,
including sanitary landfills, sewer lines, above ground tanks, and urban runoff.
As discussed in the methodology section and in the Appendix, our preliminary
analysis indicated that leaks and spills from underground tanks are substantially
more important than releases from these sources. Ihese sources appear less
significant because they generally release small volumes of contaminants, release
substances that are less toxic than those we have analyzed, or (as is the case with
most operating landfills) are located where they are not likely to affect many
wells.
We also analyzed potential exposure to nitrate levels above standards. As
discussed on page 4-36, we estimated that at any given time there are about 50 to
100 infants who are potentially vulnerable to nitrate contamination above estimated
no-effects thresholds. Important sources of nitrates include fertilizers and
septic tank leachate.
Future Risks by Pollutant
3. The chlorinated organic compounds appear to be the pollutants posing the
most significant risks. Table 4-24 presents our estimates of the individual
risks and aggregate incidence from future exposure to individual pollutants.
Many pollutants we considered do not appear on this Table because EPA does not
consider them carcinogenic, they are less potent than the substances we analyzed,
or because they are probably significantly less mobile in groundwater than the
chlorinated organic compounds.
1,1 DCE appears to be the substance posing the greatest risk of increased
cancer incidence. Other pollutants that appear to pose higher levels of risk
include vinyl chloride, methylene chloride, perchloroethylene, ethylene dibromide,
and chloroform.
It is important to note that predicting risks from particular contaminants
is highly uncertain, since quantities released are rough estimates and since
the transformation of chemicals once in groundwater is also quite uncertain.
These uncertainties are particularly relevant in the case of 1,1 DCE, since we
estimate that a significant part of the exposure may be from transformation from
1,1,1 trichloroethane. However, 1,1 DCE would remain among the more significant
industrial contaminants even without our assumption on transformation. In addition,
the evidence of carcinogenicity for 1,1 DCE is less substantial than that for
many of the other substances in our analysis.
Since we did not have information on many substances in the Santa Clara
Valley, and since we did not want to underestimate potential risks, we conducted
-------
4-81
sensitivity analysis on the pollutants included in this analysis. The results of
this analysis are discussed on page 4-85.
Future Risks by Hydrogeologic Zone
4. We estimate that future risks frcm groundwater contamination may be
substantially greater in the Bay Sands zone than in the other hydrogeologic
zones, as shown in Table 4-25. This is due to the greater concentration of
industrial sources, and the smaller and more concentrated plumes in the Bay Sands
zone. People drinking from wells affected by more concentrated plumes face
higher levels of exposure, and greater risks.
Our estimate that more people in the confined zone will be affected may seem
inconsistent with current conditions in the Santa Clara Valley. Since groundwater
flows rapidly in the Recharge zone, contamination has already affected more
public wells in this zone than in the confined zones. However, more public wells
exist in the confined zones, and in the 70 year span of our analysis we account
for the potential for contamination to reach more public wells eventually and
therefore more people in the confined zones. We expect most of this exposure
will occur through conduit wells. In addition, our base-case assumptions on
groundwater clean-up are rather pessimistic. In fact, slow moving plumes in the
Bay Sands zone may be somewhat easier to contain and clean up than faster-moving
plumes elsewhere. If such clean-up occurs, the impact in the Bay Sands zone
compared to the other zones may be less than we projected.
5. The rapid groundwater flow in the Recharge zone places public wells in
this zone at some additional risk. Plume containment and clean-up may be more
difficult in the fast-moving Recharge zone. Thus, more wells in this zone are
likely to be affected before plumes are controlled. While monitoring is likely
to detect contamination in public wells before concentrations exceed standards,
contamination could exceed standards before regular annual monitoring is conducted.
More freguent monitoring, as is conmonly practiced, should minimize this concern.
Future Risks by Type of Well
6. Some private well users may face substantial individual risks. Table
4-26 presents our estimates of future risks from groundwater contamination for
the different types of wells. As indicated, some of the individual risks in
private wells are among the most significant in the Stage I analysis. Although
many more people in the Santa Clara Valley receive their drinking water from
public wells than private wells, our estimates indicate that the incidence of
cancer may be slightly higher from exposure through private wells.
One of the key differences among large public, small public, and private
wells is the depth frcm which wells typically draw water. Small public and
private wells generally draw from the shallower aguifers, where contaminant
concentrations are higher. Large public wells generally draw frcm the deeper
aguifers. In the Bay Sands and Inland Confined zones, where about half of these
public wells are located, they are largely protected frcm high concentrations by
a clay aguitard. (We discuss the importance of this clay layer, and potential
transport of contaminants through it, later.)
Monitoring practices are a second key difference among different types of
wells. While purveyors regularly monitor large public wells, small public and
-------
4-82
private wells are generally not regularly monitored for toxic substances. If
contamination reaches a private well we do not assume, in this analysis, that it
is closed unless the concentration exceeds the taste and odor thresholds, which
are generally much higher than the standards which determine when a public well
is removed from service. Our assumption that private wells are closed when
contamination exceeds taste and odor thresholds is very important in reducing
our projections of potential risks at private wells from even higher levels.
Although we project that risks will be higher from private wells, we also
project that many more people will be exposed to contamination through public
wells because public wells serve many more people than private wells. We estimate
that about 78 public wells could become contaminated, and that about 30 to 40 of
these wells could close because contamination exceeds standards. (Thus far, we
know of 36 contaminated public wells. Two of these wells have exceeded standards.
There is sore overlap between currently contaminated wells and projected future
contamination. It would be inappropriate to add the two figures.) We estimate
that up to about 1,900 private wells may become contaminated, and that about
40% of these wells would close because contamination exceeds taste and odor
thresholds. It is important to note that these estimates of the numbers of
wells affected do not fully account for possible double-counting as some
wells may be affected by more than one plume.
Future Cancer Risk; Sensitivity Analyses
Transformation and Degradation
7. The transformation and degradation of contaminants in groundwater is
an important uncertainty affecting our estimates of the severity of groundwater
contamination from industrial and fuel contaminants. We have conducted sensitivity
analysis on this issue to determine if transformation may deserve further
attention. Table 4-27 presents the results of that analysis.
Our base case estimates of future risks from groundwater contamination
include several assunptions on the behavior of contaminants in groundwater. In
particular, we assumed that benzene degrades with a half life of ten weeks,
that EDB degrades with a half life of four months, and that TCA may be transformed
to 1,1 DCE with a half life of two years. In this section, we discuss the
importance of these assumptions first for benzene and EDB, and then for TCA.
Fuels: The IEMP base case transformation assumptions imply that most benzene
and EDB reaching groundwater degrade before reaching drinking water wells, and
that most of the contamination reaching wells does so at extremely low levels.
To test the importance of these assumptions, we analyzed the risks from fuel
contamination under the alternative assumption that benzene and EDB did not
degrade at all. Laboratory and field evidence in the Santa Clara Valley and
elsewhere suggests that the "no transformation" assumption is overly pessimistic,
particularly for benzene. Monitoring in the Santa Clara Valley has not detected
any benzene or EDB contamination in public drinking water wells, despite known
large fuel leaks. However, fuel contamination has affected some shallow wells
elsewhere in the Bay Area.
Field evidence in the Santa Clara Valley suggests that benzene degrades
into harmless daughter products, which is our base case assumption. However,
under certain conditions EDB may transform into vinyl bromide, which may be
-------
carcinogenic, and possibly more potent than EDB. Since EDB is more likely to
transform to harmless daughter products, we did not estimate risks from its
transformation to vinyl bromide. Our base case estimate of a four month half-life
for EDB, and our sensitivity case of no EDB transformation, should be adequately
conservative bounds for plausible estimates of the risks from EDB and possible
daughter products. In addition, potential exposure to EDB will decrease as lead
becomes phased out of gasoline.
Assumptions as to the degradation of benzene and EDB are important in
estimating risk from groundwater contamination. We estimate that about six times
as many public wells, and therefore people, would be exposed to contamination
from fuels if benzene and EDB did not degrade. We estimated aggregate risks
would also increase. However, these aggregate risk estimates are still fairly
low because the drinking water action levels of 0.7 ug/1 for benzene and 0.05
ug/1 for EDB would require public wells to close if they are affected at or
above these low levels. These results suggest that even if the ten week half
life estimate for benzene and the four month half life estimate for EDB are
optimistic, any other reasonable estimate of degradation is not likely to
affect our risk estimates substantially. It is important to note that, while
estimated risks remain comparatively low under a wide range of degradation
assumptions, the assumption of less rapid degradation rates significantly
increases our estimate of the number of wells which could be contaminated and
closed.
Industrial: For industrial contaminants, we assumed in our base case that
benzene degrades rapidly, as discussed above. More importantly, we also assumed
that TCA transforms to 1,1 DCE with a half life of two years. This implies
that most of the exposure to the substance posing the greatest aggregate risk
(1,1 DCE) is a result of transformation from another substance (TCA). Although
there is substantial controversy over the transformation of TCA to 1,1 DCE,
this assumption is appropriate for Stage I, given both the screening nature of
our analysis and the field evidence of 1,1, DCE's presence in the Santa Clara
Valley.
The transformation of TCA to 1,1 DCE is more important than the transformation
of benzene to harmless substances. Even if TCA is not transformed, 1,1 DCE
still appears to pose relatively high individual risks. Our estimates of
annual incidence from 1,1 DCE decline substantially (although the risks under
either assumption are fairly low), but we still estimate that incidence from
exposure to 1,1 DCE will be significantly higher than for most other substances.
Since EPA considers TCA non-carcinogenic, changing our transformation assumptions
does not affect our estimates of individual risk or incidence of cancer from
exposure to TCA. More generally, some chlorinated organic compounds may be
transformed to other chlorinated organics (such as vinyl chloride). While the
mechanisms and circumstances of particular reactions make these transformations
difficult to predict, it would be unwise to assume that these substances degrade
quickly to harmless daughter products.
The degradation of benzene to harmless products is an important issue, as
is the transformation of TCA to 1,1 DCE. For the chlorinated organics, the
potential exists for transformation to other equally or perhaps more potent
substances. It would therefore be inappropriate to assume that these substances
'o to harmless products.
-------
4-84
Dilution
One of the more uncertain assumptions in this analysis is the dilution of
contaminants as they migrate through conduit wells into the lower aquifer in
the Bay Sands and Inland Confined zones. In our base case, we have assumed
that contaminants are diluted by a factor of 50 from upper-aquifer plume con-
centrations, reflecting conduct well flow, dilution in the lower aquifer, and
dilution in pumping public wells. We also conducted sensitivity analysis in
which we assumed that contaminants are diluted by a factor of ten, which might
be more typical of an intermittent or low-volume public well.
As shown in Table 4-28, changing the estimate of dilution leads to a
slightly higher estimate of risk in our low-release case. This implies that
changing our assumptions about dilution in conduit wells and the lower aquifer
does not affect risks significantly. This is partially because public wells in
the lower aquifer are assumed to close if concentrations exceed standards.
TCA Carcinogenicity
Considerable debate exists about whether TCA is a carcinogen. EPA's
current policy is that it should not be considered a carcinogen, and we follow
that policy as a base case. However, a significant portion of the scientific
community believes that TCA may eventually be found to be a carcinogen. EPA
formerly considered it a possible carcinogen, but has suspended that classification
because a key implicating study is under review.
Because of the uncertainty on this issue, we have performed sensitivity
analysis of the possible impact of TCA if it were a carcinogen. These estimates,
which appear in footnotes to the text and tables, should be regarded as extreme
upper bounds of possible risks.* They are not part of the base case analysis.
Size of Releases
Estimating the size of past leaks and spills is very difficult, and estimating
the size of future leaks and spills is even more uncertain. Because of the
uncertainty, we have analyzed risks under two sets of assumptions about the
size of future releases: a "high release" set, and a "low release" set. We
discuss these assumptions briefly in the methodology section, and in detail in
the Appendix.
The effectiveness of the Hazardous Materials Management Ordinances is a
key factor in estimating the size of future releases from underground tanks.
We have conducted some additional analysis on this issue, which we discuss
*If TCA is a carcinogen, we estimate, as extreme upper bound values, the
range of individual lifetime cancer risks to be between .001 to .01 chances in
a million, and the risks to the maximally exposed individuals range from 20 to
70 chances in a million. Our estimates of aggregate incidence are somewhat
lower than the estimated risks from the other halogenated organics at about one
case every 11,000 to 25,000 years. Our assumption about the transformation of
TCA to 1,1 DCE is much more important to the analysis than our assumption about
TCA's carcinogenicity.
-------
4-85
later. It is important to note that effective clean-up and well monitoring
programs can significantly reduce the impact of large releases.
Substances Included
As discussed in the methodology section, our base case analysis accounts
only for those substances for which we had adequate information to estimate
risks. Since some other substances could pose risks, we conducted a sensitivity
case in which we assumed that sore other substances are as toxic and mobile as the
substances we analyzed in the base case. The total number of tanks with these
additional substances is about four times greater than the number of tanks for
the substances we did analyze. If these additional substances are as toxic
and mobile as the substances we modeled, cancer risks would be about four times
greater than under our base case assumptions (i.e., estimated annual increased
cancer incidence would be about one case every four to eight years).
Future Risks; Exposure Above Thresholds for Non-Cancer Effects
8. We estimated that exposure above non-cancer health effects thresholds
from industrial and fuel contamination may occur only in private wells, and
that up to about 500 people may be exposed above these thresholds. As shown in
Table 4-29, we estimated that this exposure may occur for five substances in
our analysis: TCA, TCE, methylene chloride, vinyl chloride,, and 1,1 DCE. Peak
exposure levels for these substances can be high, but average exposure levels
over 70 years (the duration of this analysis) are generally low.
For TCA, our estimates indicate that exposure may occur substantially
above a threshold for liver effects, and to a lesser degree above a threshold
for neurological effects.* For TCE, we project that exposure may pose some
risks of liver and neurological effects. For methylene chloride, exposure
appears to exceed thresholds for liver and fetal effects. For vinyl chloride,
exposure may pose risks of liver, cardiovascular, and kidney effects. For 1,1
DCE, we estimated that concentrations nv-,y exceed thresholds for liver and
kidney effects.
* TCA has not demonstrated any teratogenic potential in published studies
conducted using rodent species, Therefore, the IEMP base case analysis assumes
that exposure to TCA poses no risk of fetal effects. An unpublished study,
which has not undergone scientific peer review, reports fetotoxic effects
(cardiac malformations) in rate pups exposed ir\ utero to TCA (Dapson et al.,
1984). In order to assess the importance to Santa Clara Valley residents of
further research on this issue, the IEMP uses the Dapson study to examine the
possible impact of TCA under the alternative assumption that exposures
above an estimated threshold of 49 ug/1 could pose the risk of fetal effects.
THE SENSITIVITY RESULTS SHOULD NOT BE INTERPRETED AS INDICATING WHETHER OR NOT
A RISK IN FACT EXISTS: EPA RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK
MANAGEMENT DECISION-MAKING OR REGULATORY ACTION. Under this alternative assumption,
we estimated that numerous private wells and one public well might be affected
above this level. Vfe estimated that from less than ten to about 3,500 people
could be exposed above the estimated threshold. This finding suggests that more
research is appropriate on TCA's potential adverse effects. The National
Toxicology Program has commissioned a project to repeat the limited Dapson
study; results are expected in the Fall of 1986.
-------
4-86
Several factors contribute to the finding that relatively few people may
be exposed above thresholds. One of the most important factors, particularly
for private wells, is the assumption that these wells will close whenever
concentrations exceed taste and odor thresholds. Regulatory programs, including
the HMMOs, clean-up, and well monitoring also help reduce potential exposure.
In addition, many public wells are partially protected by the clay layer overlying
much of the Valley.
For public wells, the lowest estimated thresholds for non-cancer health
effects are all above the standards established for closing wells. Since we
expect that purveyors will likely detect contamination and close wells before
concentrations exceed drinking water standards, we do not expect significant
exposure to concentrations exceeding thresholds for non-cancer health effects.
Since groundwater flows much faster in the Recharge zone, the potential threat
is scmewhat higher than in the other zones, but monitoring is probably adequate
to prevent contamination fron exceeding thresholds.
For private wells, our estimates of potential exposure are significantly
reduced by the assumption that these wells will close when concentrations
exceed taste and odor thresholds. For several substances, the estimated taste
and odor threshold is below the estimated health effects threshold. This implies
that wells affected by these substances will be closed before concentrations exceed
the health effects thresholds. For some other substances, where concentrations
may exceed health effects thresholds, exposure above health effects thresholds
is limited by the taste and odor threshold.
It is important to note that the estimated taste and odor threshold for
benzene is very close to the estimated threshold for blood effects. Because
actual taste and odor thresholds vary for individuals, it is quite possible
that benzene concentrations could exceed health effects thresholds for some
people.
For many substances, exposure may occur from sources or pathways (indoor
or outdoor air, for example) in addition to drinking water. Therefore, even
where exposure does not appear to exceed thresholds in drinking water, the
exposure from other sources or pathways combined with exposure in drinking
water may result in exposure above thresholds. This is particularly important
for substances such as benzene (blood effects) and chloroform (neurological
effects) where our estimated exposures are close to but still below the health
effect thresholds.
Appendix B (attached) includes a description of our methodology for
estimating non-cancer thresholds, preliminary information on the exposures
above thresholds, and unreviewed potency estimates for those substances approaching
or exceeding thresholds. A more complete discussion of the health effects
information is included in health evaluation documents available from the IEMP.
Vfe also estimate potential exposure to nitrate levels above standards. We
estimated that at any given time about 50 to 100 infants are potentially vulnerable
to nitrate contamination above thresholds. Motification programs have probably
helped to reduce the number of infants exposed. We have not estimated risks from
exposure to nitrates, but we do not know of any cases of blue baby syndrome, the
primary health effect of concern, in the Santa Clara Valley.
-------
4-87
Key Factors in Reducing Risk
The relatively low aggregate risk estimates for future groundwater con-
tamination are in large part shaped by assumptions about natural factors, and
regulatory programs and related activities. Key natural factors include the
degradation of some contaminants and the protection provided by the local
hydrogeology. Key regulatory programs include the Hazardous Materials Management
Ordinances, groundwater clean-up, and the monitoring of large public wells with
the closure of those wells if concentrations exceed standards. Important
voluntary activities include the replacement of numerous underground tanks by
local industry. We have conducted some additional analysis of these regulatory
programs and voluntary efforts to assess their importance in reducing risks.
In this section, we first discuss the importance of the natural factors in
reducing risk, and then discuss the key regulatory programs and their implications
for risk management.
Natural Factors
9. The clay layer overlying the deep aquifer in the northern part of the
Santa Clara Valley appears to be the most important natural physical factor in
reducing risks from groundwater contamination. About half of the large public
wells in the Santa Clara Valley draw water from under the major confining
layer; the rest draw from the Recharge zone or are located in the southern
part of the Valley. Contamination could reach wells under the clay layer in
three ways: by traveling through conduit wells piercing the clay layer; by
traveling through the clay itself; or by moving under the clay layer from the
Recharge zone.
Thus far, public wells in the confined zones have been affected only by
contaminants moving under the clay layer from the Recharge zone. These wells
have been affected at low levels. We have estimated risks from this exposure
in the earlier section on risks from current exposure to groundwater contamination,
and we have also accounted for possible future exposure from this route.
We assumed that the major confining layer was impermeable that contaminants
could not move through it except through conduit wells. This assumption is
probably overly optimistic, in that contaminants may eventually reach the
lower aquifer by moving through the clay layer. However, we lack the necessary
data to analyze how quickly contaminants would move through the clay layer,
and the field data indicate that no contaminants have reached the lower aquifer
through the clay layer (except through conduit wells).
We have conducted some rough preliminary analysis on this issue by estimating
the flow of groundwater from the upper to lower aquifers. This analysis suggests
that the clay layer may offer little protection to drinking water wells near
thin areas, if any such areas exist around the edge of the clay layer. There
also appears to be some potential for the eventual contamination of the deep
aquifer through the clay layer, if contamination of the upper aquifer is extensive
and is not cleaned up. It also seems very likely that conduit wells are a
greater concern than flow through the clay layer.
10. Conduit wells linking the upper and lower aquifers are probably the
primary mechanism for transporting contaminants to the lower aquifer. We
estimated that about 21 public wells in the confined zones could be affected
-------
4-88
by transport through conduit wells. So far, no public wells in the Valley
have been affected by contaminants moving through conduit wells. There are
two important types of potential conduit wells in the Santa Clara Valley:
abandoned agricultural wells, and large public wells constructed before more
stringent ordinances were adopted in 1975.
Authorities recently discovered contamination in the deep aquifer in
Mountain View which is likely to have reached the lower aquifer through conduit
wells. This is the first time contamination has been discovered in the deep
aquifer that may have been transported through conduit wells. This finding is
consistent with the IEMP analysis, which accounted for the likelihood of groundwater
and possibly drinking water contamination via conduit wells. As the
first strongly suggestive evidence of conduit well transport, this recent
discovery may increase the priority of finding and sealing conduit wells near
contamination sites.
As discussed previously, the degradation of some substances in the groundwater
also appears important in reducing risks, particularly fuel constituents. In
addition, the transformation of contaminants in the groundwater may increase
the risks for some substances. Our analysis indicates that the transformation
of TCA to 1,1 DCE may be significant for risks. Other substances may transform
i-.o more potent substances in the groundwater and pose significant risks. The
transformation of substances in the groundwater is one of the more uncertain
parts of our analysis.
Regulatory Programs
In addition to the base case and sensitivity analyses discussed previously,
we conducted additional analysis of the effectiveness of the HMMOs, clean-up,
,a">d well monitoring and closure. There are several reasons for analyzing the
importance of these regulatory programs: to estimate their combined importance
in reducing risks; to estimate the importance of each program in reducing
risk; and to help us determine where further risk management actions might be
itsost appropriate.
We have analyzed the importance of the regulatory programs in two ways.
First, we compare our estimates of risk with regulatory programs with estimates
of what risks might be without any regulatory controls. This allows us to
analyze the combined importance of these programs in reducing risk. The second
way in which we analyze these programs is by comparing the importance of each
program in reducing risk, assuming that the other programs exist. This allows
us to discuss the incremental (or marginal) importance of each program in
reducing risk.
In conducting the additional analysis, we made some changes from our base
case analysis. Vfe made two assumptions for each program: one with the program
in place, and one without the program. For the HMMOs, we first assumed (as
in the base case described in the methodology section) that the Ordinance is
in place, that it reduces leaks, and that many tanks have been replaced. For
comparison, we then analyzed a case in which the Ordinance is not in place,
leaks continue to be large, and only some tanks have been replaced. For clean-
up, we first assumed that all future clean-up will be similar to the actual
clean-up reflected by local field data (as described in the methodology section
on page 4-71). Vfe then analyzed a case in which no future clean-up occurs.
-------
4-89
For well monitoring and closure, we first assumed that wells are monitored
annually, and closed if contaminants reach or exceed standards. We then examined
an alternative case in which wells are not monitored, and are closed only when
concentrations exceed taste and odor thresholds.
we conducted this additional analysis only for industrial contaminants,
rather than for both industrial and fuel contaminants. Since industrial con-
taminants appear to pose close to 90% of future risks, the conclusions should
still be valid. Details of this analysis are presented in the Appendix.
Impact of Combined Programs
11. We estimated that the combined effect of the HMMOs, voluntary tank
replacement, clean-up, and monitoring and closure programs would be to reduce
future risks from groundwater contamination roughly a hundred times below what
the risks might be without these efforts. As shown in Table 4-30, we estimated
that about one case of cancer might occur every five years from groundwater
contamination without these programs, and that about one case every 500 years
might occur with these programs. (These figures are not comparable to results
presented earlier because the "no programs" case is more pessimistic than any
plausible case, and the "with programs" case is more optimistic than the base
case. In the "with programs" case we assumed that more clean-up would occur
than in the base case.)
In addition to reducing aggregate risks, we estimated that these programs
would reduce the number of people potentially exposed by about 45%, or from
about 184,000 without the regulatory programs to 98,000 with the programs.
(These estimates of populations exposed reflect some double-counting of people
who might be exposed more than once.)
These findings indicate that the regulatory programs and voluntary actions
we analyzed have a substantial impact on reducing risks from groundwater
contamination. It is also important to note that even without these programs
the risks from groundwater contamination appear significantly lower than the
risks from surface water supplies of drinking water, and from air toxics.
This is largely because we expect fewer people to be exposed to groundwater
contamination because of natural factors such as the clay layer (discussed above)
and the dilution of contaminants in the groundwater. Next we present an analysis
of the marginal impact of each program on reducing risk.
HMMOs
As shown on Table 4-30, we estimate that one case of cancer may occur
about every twenty years without the HMMOs, assuming that limited groundwater
clean-up and complete well monitoring and closure are in place. By comparison,
we estimate that one case may occur about every thirty years with the HMMOs in
place.
In interpreting these results, cecall that the assumptions we used on the
effectiveness of the HMMOs were pessimistic, so as to avoid underestimating
health risks. If the HMMOs are more effective than we have conservatively
estimated, then their impact would appear to be more significant.
-------
4-90
In addition, we have analyzed groundwater clean-up separately from the
HMMOs; in effect, we assumed that the HMMOs will stop leaks, but that clean-up
does not begin unless and until drinking water wells are affected. In practice,
most clean-up is likely to occur when leaks are detected as part of the HMMO,
and could thus be viewed at least partially as an effect of the HMMOs. Ihis
suggests that the HMMOs may be more important in reducing future exposure and
risk than our quantitative analysis indicates.
Clean-up
As indicated on Table 4-30, groundwater clean-up appears to have the most
significant marginal impact on reducing risk of the three programs we analyzed.
We estimate that one case of cancer may occur about every 20 years without
groundwater clean-up, assuming that the HMMOs and monitoring and closure are
in place. If future groundwater clean-up reduces concentrations to levels
similar to current levels, and if the other programs are in place, we estimate
that one case of cancer would occur about every 500 years. Groundwater clean-up
appears to reduce future risks about thirty times below levels without clean-up.
Our assumptions about the levels of future groundwater clean-up are very
important to our estimates of the importance of clean-up. We assumed that
future clean-up of all contamination will meet the levels of current clean-up,
which has reduced concentrations substantially. Thus, clean-up greatly reduces
the potential for longterm exposure at high levels.
Since clean-up is continuing, actual final concentrations of contaminants
in wells are likely to be lower than those we have estimated, and the importance
of clean-up may be greater than we have indicated. On the other hand, if the
HMMOs are more effective in reducing both the frequency and size of releases
than we have estimated, which is quite possible, or if we credit some clean-up
to the HMMOs, then clean-up triggered by detecting contamination at drinking
water wells will be less important.
In our analysis clean-up does not begin until it is too late to keep drinking
water wells from becoming contaminated. However, clean-up did affect the number
of wells we estimated would be closed. We estimated that about 22 public wells
would close from industrial contamination without clean-up, but that only about
eight would close with clean-up. (Thus far, two wells in the Santa Clara
Valley have exceeded the standards used in this analysis.) This illustrates
the interdependence of. the drinking water protection programs: more effective
clean-up reduces the need to close wells or to treat water at the well. In
reality, if clean-up begins when leaks are detected (as is likely under the
HMMOs), then some drinking water wells are likely to be protected from becoming
contaminated.
Monitoring and Closure
As shown on Table 4-30, the marginal impact of the monitoring and closure
of public wells appeals similar to the impact of the HMMOs. We estimate that
one case of cancer may occur about every thirty years with monitoring and
closure, and about every twenty years without monitoring and closure. This
assumes that the HMMOs are in place, and that fairly extensive clean-up occurs.
-------
4-91
One important reason for the fairly snail impact of well monitoring and
closure is our set of assumptions on groundwater clean-up. Since we expect
clean-up to reduce concentrations near, or in sane cases below, drinking water
standards, and since we do not expect purveyors to close wells unless they
exceed standards, we would not expect concentrations to exceed standards sub-
stantially. In addition, monitoring and closure of public wells do not reduce
risks in private wells, whereas clean-up does. The importance of monitoring
and closure would be greater if groundwater clean-up did not occur, or is less
effective or comprehensive than assumed.
In addition, the monitoring and closure of public wells may be an important
"last line" of defense against groundwater risk. If the HMMOs and groundwater
clean-up do not prevent high levels of contamination from reaching wells,
monitoring and closure can still protect public health.*
Our analysis of the importance of the regulatory programs is quite uncertain.
The results should not be considered predictions of the actual importance of
various programs. These results are best interpreted as providing a rough
indication of the combined impact of these programs, and the relative importance
of each one. Although the results are not as solid as our estimates of risk
under the base case assumptions, the conclusion that these programs should
reduce future risks significantly is appropriate.
Resource Implications
12. Despite the comparatively low projected aggregate risks, groundwater
contamination could pose significant natural resource costs. Under our base
case assumptions we projected that contamination could affect a large number of
public wells (we estimated about 78), and that some (about 30 - 40) of these
wells could be closed. These projections have important implications for risk
management, and for managing groundwater supplies of drinking water, because
regulatory programs reduce public health risks in different ways. While well
monitoring and closure protect health, they do not protect the groundwater
itself. The HMMOs and groundwater clean-up reduce risks by protecting the
groundwater before it becomes drinking water.
Our analysis indicates that clean-up is important in reducing risks from
groundwater contamination, particularly in conjunction with preventive and well
monitoring and closure programs. While clean-up begun only when contamination is
detected at drinking water wells (as assumed in our base case analysis) does
not reduce the number of public wells affected by contamination, prompt clean-up
at sites would probably have such an effect. Even delayed clean-up substantially
reduces the concentrations and risks at drinking water wells. We estimated
that clean-up will allow a significant number of large public wells (about 18)
to remain in service with concentrations below standards that would otherwise
have to close.
* If public wells are closed some alternative supplies may pose risks. Surface
water contains significant levels of THMs. If the groundwater is treated at
the well and then disinfected, THMs could form at low levels which could still
pose greater risks than th'- contaminated groundwater.
-------
4-92
While groundwater clean-up maintains the usefulness of much of the ground-
water supply, it does so at substantial costs. Clean-up has both direct costs
(for drilling and operating monitoring and clean-up wells), and indirect costs.
These indirect costs include replacing the groundwater used in clean-up. In
the Santa Clara Valley, groundwater clean-up generally involves pumping a
mixture of clean and contaminated groundwater, treating it, and discharging it
to San Francisco Bay. All of this water which is discharged to the Bay (currently
about 15 million gallons a day) must be replaced by importing more surface
water to recharge the aquifers. Other approaches to clean-up could reduce the
need to replace these large volumes of groundwater. Alternatives that could
be examined include using this water for drinking, agricultural, or industrial
purposes either with or without treatment, or possibly recharging the water.
It may be possible to reduce future clean-up costs while protecting the
public health. For example, our analysis indicates that the effective implementation
of the HMMOs should reduce substantially the number and size of future releases.
This will reduce the amount of groundwater which becomes contaminated, the
amount of water which needs to be cleaned up, and the amount of water which
needs to be imported to replace the water which is discharged as part of the
clean-up.
The HMMOs should also reduce the need to close public wells because con-
tamination exceeds standards. Closing these wells can pose significant costs
as alternative supplies must be provided. One alternative to closing wells is
to treat water at the well to ranove the contaminants, or to reduce concentrations
to levels below standards.
The IEMP analysis suggests that it may be important to consider the inter-
dependence of groundwater-related programs and problems in devising risk management
solutions. In addition, this analysis indicates that control strategies should
account for the economic and natural resource value of groundwater as well as
possible health risks. In cooperation with local agencies, the IEMP plans to
research and address many of these groundwater issues in Stage II, including
the relationship between groundwater quality and quantity in the Santa Clara
Valley.
Research Needs
13. Further research and data on groundwater contamination is needed,
both on local and Federal levels.More information on many of the technical
aspects of this analysis would improve the understanding of groundwater contami-
nation in the Santa Clara Valley, and would also be useful for understanding
environmental health risks elsewhere.
Further local research which seems particularly important would include
obtaining more information on the effectiveness of the clay confining layer in
protecting the lower aquifer. Local authorities are attempting to identify and
seal potential conduit wells, which should protect some public wells from
contamination. However, we have very little local information on how readily
water, and contamination, might flow through the clay layer itself. In addition,
further local data on groundwater and contaminant flow would be helpful in
designing groundwater clean-up and monitoring plans.
-------
4-93
National research on contaminant transport and transformation would also
be helpful. The transformation and degradation of contaminants in the groundwater,
particularly benzene, seems important. Additional national work, sane of
which is underway, on the importance of various sources of contamination and
the possible releases (both frequency and size) would be useful for both site-
specific and national analyses. Finally, additional toxicological information
on both the cancer and non-cancer health effects of groundwater contaminants
would be helpful in future risk assessment and risk management.
Uncertainties
14. Despite many uncertainties, the IEMP Stage I groundwater analysis is
useful for comparing the various risks from groundwater contamination to each
other, and to our estimates of risks from surface water and air toxics. Although
they are not predictions of the absolute risks from groundwater contamination,
they roughly indicate the potential magnitude of the actual risks. As discussed
in our methodology section, parts of the analysis have significant uncertainties.
This section briefly notes some of the more important ones.
We are relatively confident in the estimates of exposure to toxics from
currently contaminated public wells. It is unlikely that the risk estimates
from these levels are significant underestimates of actual risks. We have not
estimated current risks from private wells since we do not have enough monitoring
data. We have projected private well risks in the future, and current monitoring
will aid in assessing the risks at private wells.
The analysis of future risks from groundwater is the most uncertain part
of the Stage I analysis. While we have attempted to use conservative assumptions,
risks could be higher than we have estimated. It is not possible to quantify
the uncertainties in this analysis, but in the methodology section we have
qualitatively indicated which assumptions are the most uncertain and Mportant.
We may have over-estimated the actual risks by following standard EPA practice
and applying conservative potency estimates. We have also considered most halogen-
ated organic chemicals carcinogenic, although there is some controversy on this
point. Many of our assumptions for estimating exposure are also conservative,
although we have used available data as a basis for assumptions whenever possible.
We may have under-estimated the actual risks in several ways. We have not
included all potential sources of contamination, nor all substances. In addition,
the confining layer separating the upper and lower aquifers in the confined zones
is not entirely impermeable. Contaminants may move either through the clay
layer, or may travel through natural cracks and fissures. Finally, the toxicology
information is limited by the experiments and epidemiology work conducted to
date, and may not include health effects for which testing has not been conducted.
Even where we do estimate risks quantitatively, extrapolation from laboratory
studies may not provide a reasonable indication of effects on humans.
We have conducted sensitivity analysis on the most important and uncertain
factors in our exposure analysis, including release estimates, transformation,
and dilution in the lower aquifer. While we are confident that we have made
reasonable assumptions on these factors for screening purposes, we have not
included the full range of possible risks. We have not included either a most
optimistic or pessimistic case, although we have included what we consider a
very pessimistic but plausible case.
-------
TABLE 4-23
FUTURE CANCER RISKS FROM DRINKING WATER CONTAMINATION
ESTIMATED INCREASED INDIVIDUAL RISKS AND ANNUAL AGGREGATE INCIDENCE (1)
(Base Case)(2)
TYPE OF SOURCE
FUEL
INDUSTRIAL
COMBINED
Lifetime Risk to Lifetime Risk to
Maximally Exposed an Average Exposed
Individuals (chances Individual (chances
in a million) in a million)
LOW
7,000
10,000
10,000
HIGH
4,000
20,000
20,000
LOW
4
19
11
HIGH
5
32
21
Exposed
Population
LOW HIGH
87,000 92,000
111,000 111,000
198,000 203,000
Annual
Aggregate
Incidence
LOW
0.005
0.03
0.03
HIGH
0.007
0.05
0.06
Weight
of
Evidence(3)
A/B2
A-C
A-C
(1) Because of significant uncertainties in the underlying data and assumptions, these estimates of Individual
risk and disease incidence are only rough approximations of actual risk. They are based on conservative
estimates of future exposure and potency, and are more likely to overestimate risks than underestimate them.
See text.
(2) These re-suits reflect base case assumptions as discussed in the text. "Low" and "high" correspond to low and
high estimates of release size, explained in text.
(3) The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to very
substantial. According to EPA's categorization of levels of evidence of carcinogenicity,, A « proven human
carcinogen: B - probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2);
C = possible human carcinogen; D - not classifiable; and E = no evidence.
-------
FUELS AND INDUSTRIAL
TABLE 4-24
FUTURE CANCER RISKS FROM DRINKING WATER CONTAMINATION:
ESTIMATED INCREASED INDIVIDUAL RISKS AND ANNUAL AGGREGATE INCIDENCE
(Base Case)(2)
(1)
Lifetime Risk to
Maximally Exposed
Individuals (chances
in a million)
Lifetime Risk to
an Average Exposed
Individual (chances
in a million)(3)
Annual
Aggregate
Incidence
POLLUTANT
TCE
CHLOROFORM
METH CL
BENZENE
VINYL CL
PCE
1,1 DCE
EDB
COMBINED
LOW
HIGH
10
100
2,000
100
500
10
10,000
7,000
60
200
800
100
20,000
9000
20,000
4,000
10,000 20,000
LOW
2
11
19
19
140
14
3,500
580
11
HIGH
5
470
25
13
700
44
4,700
700
21
LOW
HIGH
Exposed
Population(3)
LOW HIGH
300
200
11,000
300
100
100
400
600
198,000 203,000 0.03 0.06
1,200
600
11,000
500
300
800
600
700
0.00001
0.00003
0.003
0.00008
0.0002
0.00002
0.02
0.005
0.00008
0.004
0.004
0.00009
0.003
0.0005
0.04
0.007
Weight
of
Evidence
B2
B2
B2
A
A
B2
C
B2
A-C
I
VD
Ln
(1) Because of significant uncertainties in the underlying data and assumptions, these estimates of individual
risk and disease incidence are only rough approximations of actual risk. They are based on conservative
estimates of future exposure and potency, and are more likely to overestimate risks than underestimate them.
See text.
(2) These results reflect base case assumptions as discussed in the text.
high estimates of release size, explained in text.
"Low" and "high" correspond to low and
(3) Average risks for the total exposed population are lower than risks for any chemical because of our use of
different definitions of "exposure". We include people exposed to any chemical at 0.1 ppb (0.01 ppb for EDB)
greater for any period of time in the combined average. Individual chemical figures only Include individuals
exposed to a 70 year average greater than 0.1 ppb.
(4) The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to very
substantial. According to EPA's categorization of levels of evidence of carcinogenicity,, A » proven human
carcinogen: B - probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2);
C = possible human carcinogen; D - not classifiable; and E » no evidence.
-------
TABLE 4-25
FUTURE CANCER RISKS FROM DRINKING WATER CONTAMINATION:
ESTIMATED INCREASED INDIVIDUAL RISKS AND ANNUAL AGGREGATE INCIDENCE (1)
(Base Case)(2)
FUELS AND INDUSTRIAL
Lifetime Risk to
Maximally Exposed
Individuals (chances
in a million)
Lifetime Risk to
an Average Exposed
Individual (Chances
in a Million)
Exposed
Population
Annual
Aggregate
Incidence
HYDROGEOLOGIC
ZONE
Weight
of
Evidence(3)
BAY SANDS
INLAND CONFINED
RECHARGE
SOUTH COUNTY
COMBINED
LOW
10,000
200
30
10
10,000
HIGH
20,000
10,000
300
100
20,000
LOW
59
3
0.8
0.003
11
HIGH
71
6
8
0.7
21
LOW
36
106
52
5
198
,000
,000
,000
,100
,000
HIGH
39,000
106,000
52,000
5,800
203,000
LOW HIGH
0
0
0
0
0
.03
.004
.0006
.0000002
.03
0.04
0.009
0.006
0.00006
0.06
A-C
A-C
A-C
A-C
A-C
(1) Because of significant uncertainties in the underlying data and assumptions, these estimates of individual
risk and disease incidence are only rough approximations of actual risk. They are based on conservative
estimates of future exposure and potency, and are more likely to overestimate risks than underestimate them.
See text.
(2) These results reflect base case assumptions as discussed in the text. "Low" and "high" correspond to low ar
high estimates of release size, explained in text.
(3) The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to verj
substantial. According to EPA's categorization of levels of evidence of carcinogenicity,, A ~ proven human
carcinogen: B - probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2);
C = possible human carcinogen; D = not classifiable; and E = no evidence.
-------
TABLE 4-26
FUTURE CANCER RISKS FROM DRINKING WATER CONTAMINATION:
ESTIMATED INCREASED INDIVIDUAL RISKS AND ANNUAL AGGREGATE INCIDENCE (1)
(Base Case)(2)
FUELS AND INDUSTRIAL
Lifetime Risk to Lifetime Risk to Exposed
Maximally Exposed an Average Exposed Population
Individuals (chances Individual (chances
in a million) in a million)
PRIVATE
LARGE PUBLIC
SMALL PUBLIC
COMBINED
LOW
10,000
50
0
10,000
HIGH
20,000
300
40
20,000
LOW
250
4
0
11
HIGH
430
7
2
21
LOW
5,700
191,000
1,700
198,000
HIGH
6,600
194,000
1,700
203,000
Annual
Aggregate
Incidence
LOW
0.02
0.01
0
0.03
HIGH
0.04
0.02
0.00006
0.06
Weight
of
Evidence(3)
A-C ±
A-C
A-C
A-C
(1) Because'of significant uncertainties in the underlying data and assumptions, these estimates of individual
risk and disease incidence are only rough approximations of actual risk. They are based on conservative
estimates of future exposure and potency, and are more likely to overestimate risks than underestimate them.
See text.
(2) These results reflect base case assumptions as discussed in the text. "Low" and "high" correspond to low and
high estimates of release size, explained in text.
(3) The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to very
substantial. According to EPA's categorization of levels of evidence of carcinogenicity,, A - proven human
carcinogen: B « probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2) ;
C = possible human carcinogen; D - not classifiable; and E « no evidence.
-------
TABLE 4-27
FUTURE CANCER RISKS FROM DRINKING WATER CONTAMINATION:
ESTIMATED INCREASED INDIVIDUAL RISKS AND ANNUAL AGGREGATE INCIDENCE (1)
(Sensitivity Case)(2)(3)
FUELS AND INDUSTRIAL
Lifetime Risk to
Maximally Exposed
Individual (Chances
in a million)
Lifetime Risk to
an Average Exposed
Individual (chances
in a million)
Exposed
Population
Annual
Aggregate
Incidence
Weight
of
Evidence(4)
DEGRADATION
SENSITIVITY CASE
DEGRADATION
(Fuels)
NON-DEGRADATION
(Fuels)
DEGRADATION
(Industrial)
NON-DEGRADATION
( Industrial)
LOW
7,000
7,000
10,000
10,000
HIGH
4,000
4,000
20,000
20,000
LOW
4
4
19
3
HIGH
5
5
32
13
LOW
87
532
111
111
,000
,000
,000
,000
HIGH
92,000
532,000
111,000
111,000
LOW
0.005
0.03
0.03
0.004
HIGH
0.007
0.04
0.05
0.02
A/B2
A/B2
A-C
A-C
.fc.
I
<&
CO
(1) Because of significant uncertainties in the underlying data and assumptions, these estimates of individual
risk and disease incidence are only rough approximations of actual risk. They are based on conservative
estimates of future exposure and potency, and are more likely to overestimate risks than underestimate them.
See text.
"Low" and "high'
(2) Results reflect base case assumptions as discussed in the text, except for degradation.
correspond to low and high estimates of release size, explained in text.
(3) Assumes transformation of Benzene, EDB, and TCA as discussed in text. Other substances assumed not to be
transformed.
(4) The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to very
substantial. According to EPA's categorization of levels of evidence of carcinogenicity,, A = proven human
carcinogen: B = probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2);
C = possible human carcinogen; D = not classifiable; and E = no evidence.
-------
TABLE 4-28
FUTURE CANCER RISKS FROM DRINKING WATER CONTAMINATION:
ESTIMATED INCREASED INDIVIDUAL RISKS AND ANNUAL AGGREGATE INCIDENCE (1)
(Sensitivity Case)(2)(3)
FUELS AND INDUSTRIAL
Lifetime Risk to
Maximally Exposed
Individuals (chances
in a million)
Lifetime Risk to
an Average Exposed
Individual (chances
in a million)
Exposed
Population
Annual
Aggregate
Incidence
Weight
of
Evidence(4)
DILUTION
SENSITIVITY CASE
LOW HIGH LOW HIGH LOW HIGH LOW HIGH
LOW DILUTION 10,000 20,000 35(5) 20 200,000 207,000 0.1(5) 0.06
HIGH DILUTION 10,000 20,000 11 21 198,000 203,000 0.03 0.06
A-C
A-C
(1) Because of significant uncertainties in the underlying data and assumptions, these estimates of individual
risk and disease incidence are only rough approximations of actual risk. They are based on conservative
estimates of future exposure and potency, and are more likely to overestimate risks than underestimate them.
See text.
t»»
(2) These results reflect base case assumptions as discussed in the text, except for dilution. "Low" and "high"
correspond to low and high estimates of release size, explained in text.
(3) High dilution is the base case assumption as discussed in the text. Low dilution is one fifth of the base
case dilution in Bay Sands and Inland Confined Zone public wells.
(4) The weight of evidence of carcinogenicity for the compounds listed varies greatly, from very limited to very
substantial. According to EPA's categorization of levels of evidence of carcinogenicity,, A - proven human
carcinogen: B - probable human carcinogen (Bl indicates more evidence of human carcinogenicity than B2);
C > possible human carcinogen; D - not classifiable; and E - no evidence.
(5) Average risks are higher with additional dilution because minimally exposed individuals drop below the threshold
of notice with additional dilution.
-------
4-100
TABLE 4-29
Exposures Above Thresholds (1)
Estimated
Chemical Effect Threshold (ug/1)
Population
above Threshold
Peak
Exposure for Each
Chemical(ug/1)(2)(3)
70 Year Average
Concentration(ug/1)(4)
Benzene
(fuels)
Benzene
(industrial)
Chloroform
1,1 DCE
EDB
Blood
Fetal
Blood
Fetal
Li vet-
Kidney
Fetal
Neuro
Li vet-
Kidney
Repro (male)
24.5(a)
41(a)
24.5(a)
41(a)
699(a)
225(a)
795(a)
117(a)
310
25
17.5
Repro (female) 119
Methylene
Chloride
MEK
PCE
Li vet-
Li vet-
Fetal
Blood
Kidney
Neuro
NOEL(5)
Liver
Kidney
Fetal
626
2,100
2,100
86,000(a)
6,990(a)
86,000(a)
l,700(a)
699(a)
699(a)
9,090(a)
Low
0
0
0
0
0
0
0
0
20
120
0
0
0
0
0
0
0
0
0
0
0
0
High
0
0
0
0
0
0
0
0
240
340
0
0
0
<50
<50
0
0
0
0
Public Private Public Private
0.1 - 0.7 0.5 - 24 0.1 - 0.4 0.6 - 19
0.7 24 0.2 0.1
0.9 - 100 0.9 - 100 2 - 9.5 9-40
6 28 - 800 1 0.4 - 143
0.1 0.2-2 0 0.1
27 - 40 820 - 3,200 1-15 2 - 1,591
27 - 710 820 - 1,000 0.4 - 62 2 - 270
4 92 - 300 0.2-2 1-123
-------
4-101
TABLE 4-29 (Continued)
Chemical
Toluene
TCA
TCE
Vinyl
Chloride
Xylene
Estimated
Effect Threshold (ug/1)
Blood
Li vet-
Kidney
Neuro
Fetal
Reproductive
Liver
Neuro
Fetal (6)
Liver
Neuro
Kidney
Liver
Cardiovas.
Kidney
Reproductive
Fetal
Blood
Liver
Kidney
Neuro
Cardiovasc.
Blood
Reproductive
Fetal
10,100(a)
10,100(a)
10,100(a)
9,800(a)
4,760(a)
5,000(a)
979
12,500
260
260
37,700
45.5
246
246
1,640
1,640
3,010
2,150(a)
2,150(a)
2,150(a)
2,150(a)
2,150(a)
528(a)
528(a)
Population
above Threshold
Low
0
0
0
0
0
0
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
High
0
0
0
0
0
0
100
<30
<10
<10
0
<10
<10
<10
0
0
0
0
0
0
0
0
0
0
Peak
Exposure for Each
Chemical (ug/1)
Public Private
70 yr Average
Concentration (ug/1)
Public Private
24
24
0.4 - 13
1 - 23
18 - 200 67 - 32,000 0.1-12
0.2 - 110
92 - 500
0.2 - 0.3 1 - 130
70 -1,200 0.2
6 - 130
5-50
50
0.2 - 4 0.9 - 12
-------
4-102
TABLE 4-29 (continued)
(1) NOTEs Because of significant limitations in the underlying data and the use of conservative assumptions,
these estimates of exposure are rough. Because they are based on conservative assumptions, they are more
likely to over- than underestimate exposure.
(2) Peak exposures represent highest exposure level at any one time.
(3) The low values for peak exposures represent the highest estimated peak under the low release case asssumptions,
described in the text.
(4) These are estimates of the average concentrations for 70 years.
(5) The threshold for methyl ethyle ketone represents the highest dose given at which no adverse effect was seen
(No Observed Effect Level), with an additional safety factor of 1,000. Therefore, this threshold has no
health effect as its basis.
(6) The IEMP conducted sensitivity analysis for possible fetal effects. See footnote to text.
-------
4-103
TABLE 4-30
FUTURE DRINKING WATER RISKS FROM GROUNDWATER CONTAMINATION:
IMPACT OF REGULATORY PROGRAMS (1)
Sensitivity Case(2)
Status of Programs (3) Annual Aggregate Incidence Weight of Evidence (4)
Assuming HMMOs, more extensive clean-up, and monitoring programs:
- ALL PROGRAMS IN PLACE .002 (one case every 500 years) A-C
- NO PROGRAMS IN PLACE .21 (one case every 5 years) A-C
Assuming limited clean-up, complete monitoring programs in place:
- WITH HMMOS .03 (one case every 30 years) A-C
- WITHOUT HMMOS .05 (one case every 20 years) A-C
Assuming HMMOs, monitoring programs in place:
- WITH MORE EXTENSIVE CLEAN-UP .002 (one case every 500 years) A-C
- WITHOUT CLEAN-UP .05 (one case every 20 years) A-C
Assuming HMMOs, limited clean-up programs in place:
- WITH MONITORING/CLOSURE .03 (one case every 30 years) A-C
- WITHOUT MONITORING/CLOSURE .05 (one case every 20 years) A-C
(1) Because of significant uncertainties in the underlying data and assumptions,
these estimates of disease incidence are only rough approximations of actual
risk. They are based on conservative estimates of future exposure and potency,
and are more likely to overestimate risks than underestimate them. See text.
(2) These results reflect various assumptions as discussed in the text and Appendix.
(3) Our results for cases with the Hazardous Materials Management Ordinances reflect
known tank replacement at industrial facilities as of mid-1985. Cases without
the HMMOs reflect tank replacement at only a few industrial facilities.
(4) The weight of evidence of carcinogenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's
categorization of levels of evidence of carcinogenicity, A = proven human
carcinogen; B = probable human carcinogen (Bl indicates more evidence of
human carcinogenicity than B2); C = possible human carcinogen; D = not
classifiable; and E = no evidence.
-------
4-104
REFERENCES
1. Association of Bay Area Governments (ABAC): Evaluation of Air Emissions, Runoff
to Surface Water, and Leachate to Groundwater from Sanitary Landfills^
July 1985.* ~ ' ~~~
2. Environmental Protection Agency: National Organics Monitoring Survey,
1977.
3. Environmental Protection Agency, Office of Drinking Water: Coninunity Water
Supply Survey; Sampling and Analysis for Purgeable Organics and Total
Organic Carbon (Draft), June 9, 1981.
4. Environmental Protection Agency, Office of Drinking Water: The Ground
Water Supply Survey: Sumnary of Volatile Organic Contaminant Occurence
Data, June 1982.
5.- Environmental Protection Agency, Office of Drinking Water: Ground Water
Supply Survey, 1982.
6. ICF, Incorporated: Stage I Screening Analysis of Possible Exposures to
Toxic Contaminants in Drinking Water in the Santa Clara Valley, June
1986 (Appendix to Chapter 4).*
7. Mackay, Douglas M.: Derivation of Quantitative Estimates of Retardation and
Transformation for Organic Contaminants in the Groundwaters of Santa Clara
Valley, California; A Report to the IEMP Project Staff, EPA Region IX,
November 11, 1985 (Included in Appendix).*
* These documents are available from the Santa Clara Valley IEMP.
-------
CHAPTER FIVE
ANALYSIS OF RISKS FROM TOXICS IN SOUTH SAN FRANCISCO BAY
-------
Chapter Five
Analysis of Risks from Toxics in South San Francisco Bay
I. Introduction 1
II. Characterization of South San Francisco Bay. 2
III. Pollutant Sources and Emissions 2
A. Types of Pollutants
B. Types of Pollution Sources
C. Pollution Source Loading Rates
D. Regulations of Point Source Qnissions
IV. Ambient Environmental Conditions 19
V. Potential Human Health Risks 21
VI. Conclusions 36
-------
CHAPTER FIVE
ANALYSIS OF RISKS FROM TOXICS IN SOUTH SAN FRANCISCO BAY
This chapter describes the methodology and results of the Santa Clara
Valley lEMP's Stage I analysis of health risks from toxics in South San
Francisco Bay (the South Bay). We present data on toxic pollution sources and
regulations, ambient environmental conditions and potential health risks.
This chapter does not provide a complete analysis of the health risk associated
with consumption of fish from the South Bay; rather, it is an attempt to
gather existing data and make preliminary estimates of potential health threats
posed by regular consumption of fish and/or shellfish from the South Bay
(the term "fish" is used to refer to both fish and shellfish in this chapter).
First we examine types and sources of pollution in the South Bay. We
separate the sources into two categories - point and nonpoint sources.
Point sources are specific identifiable sites that discharge directly into
surface water, such as a sewage treatment plant. Nonpoint sources cannot be
traced to a specific site; an example of nonpoint source pollution is runoff
fron a parking lot during a rainstorm. Estimates of the loading rates of
these sources, i.e. the amount of pollution emitted from different sources,
are presented. Point source pollution estimates are available from required
monitoring data. In contrast, nonpoint source pollution estimates are
extremely difficult to make, and should be considered less reliable. We then
briefly discuss the regulations controlling current sources of pollution and
the status of South Bay point sources in meeting these regulations.
Measures of ambient environmental conditions are presented to illustrate
the amount of toxic contamination of the South Bay. Data on toxics in the
South Bay water column, sediments and fish tissue are presented and then
compared with both historical data and other geographic areas. We do not
attempt to link directly the pollutant loading rates with the ambient levels
of toxics in the South Bay environment due to the complexity of the interaction
of the pollutants with the environment.
The IEMP analysis then discusses the methodology used to estimate exposure
to toxics in the South Bay, and the limitations of this exposure analysis.
Wnile source loading estimates and toxic contamination data are reasonably
reliable because the data are from monitoring, our estimates of human exposure
are extremely uncertain. We calculate potential health risks to a hypothetical
highly exposed individual who consume contaminated fish from the South Bay
regularly. While fish consumption appears to be the most likely route by
which toxics in the South Bay would affect people, we believe it is unlikely
that many people have a regular diet of fish caught from the South Bay.
Because of this study's limited resources and our focus on human health
risks, we did not attempt to duplicate or summarize the numerous studies that
have been done on the ecological health of the Bay, although we present some
data from these studies. In April 1986, the EPA initiated a major program to
preserve the Bay's envircmental health by adding the San Francisco Bay to
its National Estuary Program. This program will help to protect and restore
water quality and aquatic resources. It seeks to create a master environmental
plan to control point and non-point pollution. The plan is intended to
-------
5-2
protect living resources, control freshwater input and removal, foster
sound land use practices, increase public understanding of the Bay's environ-
mental problems and establish anti-degradation policies for pristine areas.
These regulatory programs for protection of the ecological health of the Bay
may reduce the amount of toxics entering the South Bay benefitting both the
environment and, indirectly, human health.
Characterization of Surface Waters in South San Francisco Bay
For the purpose of this study, the surface waters of the South San
Francisco Bay include all San Francisco Bay waters and inflowing streams
south of the Dumbarton Bridge (Figure 5-1). They comprise a complex system of
open water, tidal mudflats, salt ponds and associated dikes and levees, fresh
and brackish inflowing streams and sloughs, and fresh and saltwater marshes.
There is little water circulation in the South Bay. Tidal action is
the predominant force of water exchange. Streams are of local drainage and
are intermittent. Annual freshwater flow from wastewater treatment plant
discharges exceed annual stream inflow.
The waters of the South Bay support a variety of animal life. Resident
species include: benthic invertebrates (sea or river bottom organisms without
backbones), such as worms, oysters, clams, shrimp, crabs, and barnacles;
fish, including northern anchovy, shiner perch, staghorn sculpin, tonguefish,
speckled sanddabs, and starry flounder; and migratory and resident bird
species. Most fish species are believed to use the area for nursery grounds.
Birds use the South Bay as a feeding and resting area. The rare and
endangered bird species inhabiting the area include the California least
tern, California brown pelican, California black rail, and the salt marsh
yellowthroat. The salt marsh harvest mouse inhabits the area and is considered
an endangered species. In addition, the area supports a rich variety of
plant life in the tidal marsh fringes of the Bay and adjoining sloughs and
creeks.
There are few recreational and commercial activities associated with the
South Bay. The San Francisco Bay Wildlife Pefuge, extending along the shore-
line from Fremont to Palo Alto, greatly restricts Bay access. What little
sport-fishing there is occurs near the marinas at Palo Alto and Alviso Slough,
and some duck hunting takes place along the shoreline and estuaries. There
are no public beaches; hence few people swim in the South Bay or its associated
streams and sloughs. Commercial activities are limited to the harvesting of
shrimp for live bait and the production of salt.
Pollution Sources and Emissions
Pollutants
The pollutants contaminating the South Bay can be divided into three
general groups. The first group, known as conventional pollutants, are
defined as biological oxygen demand (BOD), total suspended solids (TSS), oil
and grease, coliform bacteria and pH. Non-conventional pollutants, the
second group, include ammonia (NH3), phosphate (P04), and physical and
biological factors (such as temperature, turbidity and bacterial concentration).
The third group is toxic pollutants, which include metals and toxic organic
chemicals. Toxic substances, the focus of this report, can have a negative
impact on the environment and can present a human health threat if people are
-------
u>
Clara
FIGURE
5-1
AREA OF CONCENTRATED STUDY
-------
5-4
exposed to the toxics. Direct exposure to toxics in local surface water can
take place through drinking contaminated water, swimming, or consumption of
contaminated fish. However, in the South Bay exposure appears to be limited
primarily to consumption of contaminated fish, as discussed later in this
chapter.
While conventional and non-conventional pollutants have both potential
health effects (i.e. through bacterial contamination) and ecological impacts
the scope of the IEMP analysis is limited to the human health effects of
toxic pollutants. We present estimates of the loadings of conventional and
nonconventional pollutants for comparative purposes.
Sources of South Bay Pollution
Contamination of South Bay water by toxic pollutants occurs as a result
of discharges from both point sources and non-point sources. Point sources
are specific sites which discharge directly into surface waters. Non-point
sources are dispersed and non-specific sites which discharge into surface
waters through storm drains and streams.
There are five South Bay point dischargers:
The Palo Alto Regional Water Quality Control Plant, which serves Palo
Alto, Mountain View, and Los Altos, with a total population of approximately
175,000, discharges to Guadalupe Slough under a National Pollutant Discharge
Elimination System (NPDES) permit. The rated capacity of the plant is 35 mgd
and it has tertiary treatment capability. In addition to municipal wastes it
treats wastes from a variety of industries including semiconductor and chemical
companies. In 1983, the average daily flow was 32.2 million gallons per day
(mgd).
The San Jose/Santa Clara Water Pollution Control Plant serves an
area of approximately 300 square miles with an estimated population of
1,125,000. The plant has tertiary treatment capability. It treats wastes
from four municipal sanitary districts (Burbank, Cupertino, Milpitas, and
Sunol); the City of San Jose; the City of Santa Clara; County Sanitation
Districts 2, 3, and 4; and several industries, including canneries, semi-
conductors, and metal finishers. The plant is currently undergoing capital
improvements to provide 167 mgd mean peak week wastewater treatment capacity
and to eliminate overloading of unit processes. Effluent is discharged to
Artesian Slough under a NPDES permit. Average flow in 1984 was 115 mgd.
The Sunnyvale Water Pollution Control Plant, serves the City of Sunnyvale
with a population of 106,618, Moffett Naval Air Station, and a number of
industrial facilities, including semiconductor manufacturers, other electronics
firms and metal finishers. The rated capacity of the plant is 29.5 mgd and
it provides tertiary treatment. Effluent is discharged to Guadalupe Slough
under a NPDES permit. The average daily flow in 1984 was 21.6 mgd.
The Gilroy-Morgan Hill Wastewater Treatment Facility is a land
treatment/disposal facility; under normal operations no liquid wastes are
discharged to surface waters. However, because of inadequate capacity
and soil which is relatively impermeable in the winter, it has historically
discharged wastewater into the Llagas Creek when heavy rains or other
problems overwhelm the plant's land disposal capacity. Llagas Creek
drains into the Monterey Bay via the Pajaro River; hence, no discharge to
the South Bay occurs.
-------
5-5
Cerro Corporation, a manufacturer of brass products, has been discharging
to the South Bay by way of Mowery Slough since 1957. Average daily discharge
is 0.07 mgd. Cerro Corporation's NPDES permit sets discharge limits for
suspended solids, oil and grease, chromium, copper, and zinc. Cerro
Corporation is not considered a major point source.
FMC Corporation, which manufactures a variety of phosphate products,
soda ash, caustic soda, and caustic potash, began discharging to the South
Bay in 1929. Industrial wastes containing phosphate are piped to a spray
cooling pond. The overflow, which is regulated under a NPDES permit for
suspended solids and phosphates, is discharged into a drainage ditch that
flows into Plummer Creek. The average phosphate discharge for the period
1977 to 1982 was 18.6 kg per day. FMC is not considered a major point source.
Non-point sources emanate from a variety of sites and sources. Because
of impervious surfaces created by construction and paving in urbanized areas,
the major portion of rain water leaves as urban runoff. Urban runoff can
consist of street runoff, stormwater runoff from storm water collection
systems, and irrigation runoff containing fertilizers and pesticides from
landscaping. These waters will likely contain whatever chemicals, metals,
and particulate matter that have been deposited in the streets and soils of
the city, including toxics. Table 5-1 lists some potentially significant
urban runoff pollutant sources. Figure 5-2 illustrates both nonpoint and
point source pathways into the South San Francisco Bay.
Pollutant Loading Rates
In this section, we present data on the current quantity of emissions
as well as yearly variations, trends, seasonal variations, nonpoint
source loadings and a comparison of loading rates between point and
nonpoint sources. An understanding of source loading rates is important
for identification of possible targets for reduction of discharges.
Table 5-2 lists the major point source pollutant loading rates for
conventional pollutants and total metals. Loadings from the three PCTWs
(San Jose/Santa Clara, Palo Alto, and Sunnyvale) are presented. The
loading rates of FMC Corporation and Cerro Corportion are not presented
because they are considered minimal in comparison with those of the
POTWs. Table 5-3 breaks down the heavy metals emissions into loading
rates for specific metals for each POTW.
The yearly variation in emissions is also illustrated in Tables 5-2
and 5-3. Even in recent years (1982 -1984) the loading rates of different
pollutants and metals fluctuates. Likely, pollutant loading rates for
1995 were estimated using 1984 South Bay Discharger Association (SBDA)
effluent rates and projected 1995 flow rates, and show an expected increase
despite a recent upgrading of treatment capacity.
Historical trends in emissions are presented in Figure 5-3. Figure
5-3 shows that that the point source loading rates for major pollutant
classes have declined dramatically since 1962, in spite of a doubling of
the flow rate, due to increased treatment efficiency. Plant upsets at
the San Jose/Santa Clara POTW occured in 1979 and 1980 due to overloaded
capacity and resulted in higher loading rates. However, the San Jose/Santa
Clara POTW has upgraded its treatment capacity, and upsets due to overloaded
capacity are not expected in the future.
-------
Table 5-1
Potential Significant Urban Runoff Pollutant Sources
Pollutants
Sediment
Oxygen-demand ing
matter
Nutrients
Bacteria
Heavy Metals
Pestidides/
Herbicices
Oil & Grease
Floating Matter
Other Toxic
Materials
Rooftops
Street
Surfaces
X
X
Parking
Lots
Landscaped
Areas
X
X
X
Vacant
Land
Construction
Sites
Other-1
X
X
1 Industrial and solid waste runoff.
SOURCE: Pitt and Bozeman, 1982
-------
Point
Figure 5-2 R>llutant Pathway*
Industrial
Facility
Residence
n
Industrial
Facility
Pratreatrtwnt-
t?TOES Regulated Discharge
»ton-point
Sources
Streets Parking Lots
Vacant Land Agricultural
Areas
-Construction Sites Runoff
\
Sewage Treatment Plant
from Industrial,
Facilities .
NPEeS Regulated Discharge
South San Francisco Bay
-------
5-8
Table 5-2
Major Point Source Pollutant Effluent Emissions^
Flow (million
SJ/SC
SUN
PA
Total
1982
gallons per
109.5
15.5
28.0
153.0
In Santa Clara
1983
day-mgd)
118.8
22.0
32.2
173.0
Valley
1984
115.0
21.6
29.5
166.1
19952
147.0
25.0
33.0
205.0
BOD (pounds per day-ppd)
SJ/SC
SUN
PA
Total
TSS (ppd)
SJ/SC
SUN
PA
Total
NH3-n (ppd)
SJ/SC
SUN
PA
Total
P04 (ppd)
SJ/SC
SUN
PA
Total
Oil & Grease
Total
Metals (ppd)
SJ/SC
SUN
PA
Total
2500.0
300.0
1300.0
4100.0
1200.0
1350.0
1750.0
4300.0
300.0
150.0
1050.0
1500.0
19000.0
100.0
4100.0
23000.0
(ppd)
2310.0
120.0
25.0
75.0
215.0
4800.0
500.0
1000.0
6300.0
2500.0
2000.0
1700.0
6200.0
1000.0
150.0
1000.0
2150.0
14500.0
300.0
4300.0
19100.0
1580.0
140.0
20.0
70.0
230.0
4050.0
1150.0
850.0
6050.0
1900.0
2500.0
1050.0
5450.0
1150.0
300.0
650.0
2100.0
16900.0
1500.0
4400.0
-22800.0
1790.0
140.0
20.0
60.0
220.0
5150.0
1350.0
950.0
7450.0
2450.0
2850.0
1200.0
6500.0
1470.0
310.0
740.0
2520.0
21600.0
1700.0
4900.0
28200.0
2210.0
180.0
25.0
65.0
270.0
-"-Data taken from SBDA Third Year Monitoring Report 1985. Larry Walker
Associates/Kinnetic Laboratories, Inc.
2 Estimated average concentrations. See Text.
-------
5-9
Table 5-3
Major Point Source Heavy Metals Bnissions^
Cadmium
SJ/SC
SUN
PA
Total
Copper
SJ/SC
SUN
PA
Total
Chromium
SJ/SC
SUN
PA
Total
Lead
SJ/SC
SUN
PA
Total
Mercury
SJ/SC
SUN
PA
Total
Nickel
SJ/SC
SUN
PA
1982
2.3
-
0.7
3.0
19.3
9.3
30.4
59.0
7.0
1.2
2.3
10.5
14.0
0.7
6.4
21.0
0.2
-
0.1
0.3
32.9
8.0
10.8
in Santa Clara Valley
pounds per day
1983
2.0
0.1
1.5
3.6
24.1
6.3
19.4
50.1
6.3
1.3
3.4
11.0
16.2
0.6
7.4
24.2
0.2
-
0.1
0.3
32.0
8.0
19.8
1984
5.82
0.3
1.0
7.1
23.0
3.6
15.0
41.6
9.6
0.9
2.5
13.0
16.3
0.5
5.7
22.5
0.4
0.1
0.5
32.6
4.9
14.8
Total
51.7
56.6
52.3
Continued on next page
-------
5-10
Table 5-3 - continued
1982 1983 1984
Silver
SJ/SC
SUN
PA
1.9
-
1.3
1.8
0.3
2.3
1.9
0.2
1.2
Total 3.2 4.4 3.3
Zinc
SJ/SC 42.2 51.3 48.0
SUN 7.2 8.7 7.6
PA 13.7 14.4 15.7
Total 63.1 74.4 71.3
Total Metals
SJ/SC 121.0 139.0 140.0
SUN 26.0 22.0 19.0
PA 68.0 68.0 58.0
Total 215.0 229.0 217.0
1 Data taken from SBDA Third Year Monitoring Report, 1985. Larry Walker
Associates/Kinnetic Laboratories, Inc.
2 Disputed values. These numbers were derived, in part, from data collected
by the Regional Water Quality Control Board, and the Department of Health
Services. The staff of the San Jose/Santa Clara WPCP has disputed the
apparent increases indicated by these values in separate communications with
Regional Board staff.
-------
5-11
_,
Ftow
180-
HO-
eo-
00-
BOO
10-
NH3
fpckMO »
TS3
Total
»oo-
1070
reeo
FIGURE 5-3 HISTORIC SBDA LOADINGS
Source: Self monitoring reports; Kinnetic Laboratories,
Inc. et al.f 1983.
-------
5-12
Figure 5-4 illustrates the seasonal variation in loading rates for
conventional polllutants. This variation can be important in terms of ambient
impact. It is possible that toxics may have similiar seasonal variability in
loading rates.
Non-point pollutant loading rates are extremely difficult to quantify
and classify because they mirror the myriad commercial and recreational
activities of a city, and vary with the amount of rainfall. Furthermore,
no one is required to record the quantities of runoff generated, nor the
pollutants which it may contain. Data on non-point sources are therefore
not as extensive or reliable as for point sources and controlling non-point
sources of toxic pollutants may prove difficult because loading rates and
sources are not well understood.
However, some estimates of non-point source pollution loading rates are
available and we present these in Table 5-4. These estimates should be con-
sidered very uncertain, especially compared to point source data. The estimates
are based upon a study that took data from representative streams in the
South Bay and then multiplied these figures by average annual streamflow
estimates to arrive at an estimate of annual loading rates.
These estimates allow a rough comparison of non-point discharges
with the respective values for point source discharges as presented in
Table 5-5. Although point source discharges have a higher annual flow
rate than non-point discharges, non-point discharges have a higher loading
rate than point sources for many pollutants. Table 5-6 illustrates this
point. However, it is important to recognize that non-point source loading
rates vary considerably during the year, and peak during periods of high
rainfall. Thus, toxics released during these peak periods are more diluted and
do not necessarily have the same impact as toxics entering the Bay during low
flow periods. Non-point source pollution's relative impact upon water quality
is an area of further research both on the national level and for the San
Francisco Bay through the National Estuaries Program,, This research should
further understanding of the relationships between both point and non-point
source pollution loading, ambient water quality, and the ecological health of
the Bay.
Regulation of Point Sources
Point sources of pollution are currently regulated under the federal
Clean Water Act and state law. According to the NPDES provisions of the
Federal Clean Water Act, all direct discharges of pollutants into waters of
the United States must obtain a discharge permit. Facilities are required to
meet nationwide standards, established by type of facility, for effluent
quality. Dischargers may be required to meet more stringent effluent standards
if necessary to meet water quality objectives. In California, the Federal
government has delegated the responsibility for the administration of permitting
to the State Water Resources Control Board and its Regional Water Quality
Control Boards.
NPDES requirements and the POTVTs' effluent record for 1984 are presented
in Table 5-7. Table 5-7 indicates that the three major POTWs were generally
meeting their discharge requirements during this recent period.
-------
Plow
IUt«
0400)
960-
200-
tao-
JOO-
0-
*!
tftOO-
aooo-
15OCM
10OO-I
6OOH
/BOO
' ' \
S '» N
\ \-
\
Flow
i
i>
O)
]OAO
M J
Month
FIGURE 5-4 MONTHLY LOADING RATES FOR POINT SOURCE DISCHARGERS
-------
5-14
Table 5-4
Parameter
BOD
TSS
Oil & Grease
NH3
N03
P04
Cadmium
Chromiun
Copper
Lead
Mercury
Zinc
Total Metals
Non-point Source Pollution Loading Rates-*-
in Santa Clara Valley
Average Annual
Loading (pounds per day)
23,820
239,000
5,000
80
700
2,380
10
20
30
400
<0.1
180
640
Total Flow
120 mgd
Data taken from SBDA Third Year Monitoring Report, 1985, Larry Walker
Associates/Kinnetic Laboratories, Inc. For a discussion of methodology
used to estimate emissions see text.
-------
5-15
Table 5-5
Comparison of Point and Nonpoint Loading Rates^
Nonpoint^
in Santa Clara Valley
Point3
Ratio of Nonpoint/Point
CONVENTIONAL POLLUTANTS
BOD
TSS
Oil &
Grease
NH3
PO4
HEAVY METALS
Cadmium.
Chromium
Copper
Lead
Mercury
Zinc
Total
Metals
Total (mgd)
Flow
23,820
239,000
5,000
80
2,380
10
20
30
400
0.1
180
640
120
6,050
5,450
1,790
2,100
22,800
7.1
13.0
41.6
22.5
0.5
71.0
220.0
166
3.9
43.0
2.3
0.04
0.1
1.4
1.5
0.7
18.0
0.2
2.5
2.9
0.7
All units are in pounds per day, unless otherwise noted.
Nonpoint estimates are averaged over the entire year.
Point source estimates are for 1984.
-------
5-16
Table 5-6
Comparison of Peak Urban Runoff Rates and Average POTW Effluent Rates
For Selected Parameters1
San Jose/
Santa Clara
Average
Pollutant
BOD
TSS
Cadmium
Chromium
Copper
Lead
Zinc
Peak Runoff (mg/L)
30
850
.04
.04
.09
1.5
.55
effluent (mg/L)
21
26
.002
.016
.081
.001
.087
Ratio of Runoff /POTW
1.4
32.0
20.0
2.5
1.1
1500.0
6.3
1 Source: Anatec, 1985
-------
5-17
Table 5-7
NPDES Limits and POTW Results2
in Santa Clara
Valley
December 1983-November 1984 Results
Parameter
BOD
TSS
Oil & Grease
Settleable
Matter
Turbidity
Cl2 Residual
PH
Toxic ity
Metals
Arsenic
Cadmium
Time Frame Units
30 day avg2 mg/1
Daily max. "
30 day avg.2 mg/1
Daily max "
30 day avg.2 mg/1
Daily max. "
30 day avg.2 ml/l-hr
Instant max. "
Instant max JTU
Instant max. mg/1
Range
Median Survival %
(3 samples)
(monthly average)
6 month
median/daily
max. mg/1
n ii
Total Chromium "
Copper
Lead
Mercury
Nickel
Silver
Zinc
ii n
n n
n n
ii n
n n
ii n
NPDES
Limits
10, 203
20, 303
10, 203
20, 303
5
10
0.1
0.2
10.0
0.0
6.5-8.5
SJ/SC Sunnyvale
5.2 9.8
10 12
4.4 23.4
19.4 28.7
1.6 4.1
4.5 5.9
0.0 ND4
4.5 9.0
0.86 0.0
6.5-7.4 6.4-7.9
90 95-100 95-100
0.01/0.02
0.02/0.03
0.01/0.02
0.2/0.3
0.1/0.2
0.001/0.002
0.1/0.2
0.02/0.04
0.3/0.5
0. 003/0. 0076 0.004/0.012
0.006/0.018 .0006/.001
0.010/0.016 .005/.014
0.024/0.029 .020/.04
0.017/0.026 .003/.012
Palo Alto
3,0
10
5.0
10
3.2
3.2
ND
7.9
G.O
6.4-7.6
90-100
.008/.01
,004/.OQ.J
.01/.018
.061/.075
.023/.038
0.0004/0.001 .0001/.0017 .0004/.00
0.034/0.038 .027/.07
0.002/0.002 .001/.002
0.050/0.052 .042/.063
.06/.07
.005/.009
.064/.071
continued on next page
-------
5-1,
Table 5-7 (continued)
December 1983-November 1984 Results
Parameter Time Frame Units
Cyanide
Phenolic
Compounds
NPDES
Limits
0.1/0.2
0.5/1.0
SJ/SC Sunnyvale
.01/.02 .04/.06
.0039/.0046 .02/.05
Palo Alto
.04/.06
.05/.05
Total
Identifiable
Chlorinated Hydrocarbons
.002/.004 .001/.001
ND
Total median MPN/100 ml 23 2-5.7 2-105
Coliform (monthly avg.) "
Organisms Maximum 5000 2400 2400
.001/.001
2-21
140
1 SOURCE: SBDA Third Year Monitoring Report, 1985. Larry Walker Associates/Kinnectic
Laboratories, Inc.
2 Maximirn month data.
3 Corresponding values for Sunnyvale POTW are 20 mg/1 and 30 mg/1.
4 Not detected.
-------
5-19
For POTWs, the Clean Water Act requires the facilities to establish
plans for controllingor pretreatingwastes introduced into municipal
sewer systems. Pretreatment requirements are designed to reduce the
flow of toxic chemicals - heavy metals and organic substances - from
industrial sites to the POTW. POlWs are designed to control conventional
pollutants and are not equipped to reduce toxic pollutants. The EPA has
developed national regulations establishing allowable discharge limits
for various industries; local pretreatment programs must meet or exceed
these standards. All three local POIWs have established limits on toxic
chemicals in water discharged to the POTWs by industrial facilities and
have programs in place to assure compliance.
Ambient Environmental Conditions
Water pollution frcm point and non-point sources can be damaging to the
ecology of the South Bay. The potential effects include:
o the presence of toxic contaminants in sediments, water column, plants
and fish/shellfish tissue;
o changes over time in animal and plant communities near discharges;
o reduced reproductive rates and increased morbidity and mortality
among plant and animal species;
Bear in mind that the relationship between pollution loading and
toxics found in the water column, sediments and fish/shellfish tissue is
extremely complex and not well understood. This is especially true for
the South Bay because it is not a closed system. Toxics may travel
through the water column from sources elsewhere in the Bay, toxics in
sediments have undetermined origins, and toxics in fish tissue may have
been picked up anywhere in the fish migratory path. While it is reasonable
to assume that local point and nonpoint sources have an impact on local
ambient environmental conditions, the exact nature of the impact is not
always clear.
There are many measures of ambient conditions of levels of toxic
contamination in a water body. Our analysis presents data on ambient
concentrations of toxics in the water column, sediments, and fish tissue.
These measures allow comparison to other situations and we present
information on historical levels as well as toxic levels in other geo-
graphic areas.
Table 5-8 presents the concentrations of metals in the water column
in the South Bay and compares them to historical South Bay data, EPA Criteria
for Salt Water Aquatic Life, and California Ocean Plan Limits. Wnile
these standards are not ambient water quality standards for San Francisco
Bay, they can provide a crude basis for comparison. The EPA Criteria for
Saltwater Aquatic Life are contaminant concentration levels designed to
protect human health and support a balanced indigenous population of
plant and animal life. The State Water Resources Control Board sets the
California Ocean Plan Limits. Both the EPA and the State limits apply to
saltwater aquatic organisms in the open coastal waters and so are not
-------
TABLE 5-8 1984 METALS AND TRACE CONSTITUENTSWATER COLUMN
(all concentrations in ug/1)
Metal
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Silver
Zinc
Dumbarton Brdg.
Monitoring
Station
ave. (max.)/l
9.4 (30)
12.3 (30)
16.6 (30)
51.8 (185)
<0.1 (0.2)
50.8 (165)
4.3 (9)
20.1 (28.5)
South Bay
Monitoring
Station
ave. (max.
9.2 (30)
12.3 (30)
12.8 (20)
<18.5 (57)
<0.1 (0.
46.9 (155)
6.2 (10)
14.5 (20)
Artesian Slough
Monitoring
Station
) ave. (max.)
9.5 (30)
<8.8 (20)
14.0 (25)
9.9 (25)
1) <0.5 (1.55)
23.8 (35)
3.9 (8)
32.4 (55)
Guadalupe Slough
Monitoring
Station
ave. (max.)
6.7 (20)
8.0 (15)
10.6 (16.5)
<12.0 (24.5)
<0.1 (0.2)
30.4 (60)
4.5 (7)
23.6 (34)
EPA Criteria
Saltwater
Aquatic Life
Daily Instant
ave. max.
4.5 59.0
18.0 1260.0
4.0 23.0
25.0 668.0
0.1 3.7
7.1 140.0
2.3
58.0 170.0
CA Ocean
Plan
Limits/2
Daily Instant
ave. max.
3.0 30.0
2.0 20.0
5.0 50.0
8.0 80.0
0.14 1.4
20.0 200.0
0.45 4.5
20.0 200.0
Ln
I
/I Maximum for any quarterly sample, averaged over depth (typical),
/2 Not directly applicable to enclosed bays of California.
Adapted from: Larry Walker Associates/Kinnetic Laboratories (1985)
-------
5-21
directly applicable to the South Bay. A comparison of South Bay levels
with the two sets of criteria shows that levels of toxics found in the
South Bay are higher than many of the criteria; this is not surprising
because the areas surrounding the South Bay are heavily populated and
industrialized.
Table 5-9 presents the concentrations of metals in South Bay sediments
and compares them to historical South Bay data. The historical data may
not be directly comparable because the sampling stations are different.
Nevertheless, they do provide some basis of comparison and show a variability
in levels over time for different metals. Levels of cadmium, chromium,
mercury, and silver appear to be increasing slightly. However, levels of
copper, lead, and zinc have no conclusive trend.
Data on concentrations of toxics found in fish tissues in the South Bay
compared with levels of toxics in other areas show that the South Bay is
more contaminated than at least some other industrialized areas. Table
5-10 compares the metal concentrations found in mussels taken from the
South Bay with those from a pristine area (Tomales Bay) and an industrialized
area (Los Angeles Harbor). All of the trace metal concentrations were sub-
stantially higher in the South Bay than in the Tomales Bay. Silver, copper,
cadmium, and chromium levels were observed to be elevated for the South
Bay in comparison with the L.A. Harbor.
Data from the California Mussel Watch Program, which identified 56
synthetic organic hydrocarbons in mussels in 1978 and 1979, provides an
additional comparison. Table 5-11 presents a comparison of selected toxic
organic contaminant concentrations in mussels between San Francisco Bay
(Redwood Creek), Los Angeles Harbor, and San Diego Harbor. San Francisco
Bay appears to be have higher degrees of contamination for most of these
substances than the other areas.
In summary, significant levels of toxics in the ambient environment
have been detected in the South Bay's water and sediment and in shellfish
tissue samples taken from the South Bay. Limited comparison of South Bay
toxics levels with those of a few other areas indicates that local metal
and organic contamination levels are clearly higher than those found in
pristine areas, and that some toxics appear to be higher than levels found
in other industrial areas as well.
Human Health Risks
There are several possible exposure pathways by which toxics from the
South Bay could affect human health. Plants and animals with toxic chemicals
in their tissues can be consumed by humans, or humans can swim in or drink
contaminated water. However, because South Bay water is not used as a source
of drinking water and there is little swimming or diving in the South Bay
due to limited access and unfavorable conditions, it appears likely that
exposure to toxic contamination occurs almost exclusively through the
ingestion of contaminated fish.
-------
TABLE 5-9 1984 METAL CONSTITUENTS OF SEDIMENTS
(all values in mg/kg\)
1984 Results
Metal
Cadmium .
Chromium
Copper
Lead
Mercury
Nickel
Silver
Zinc
Artesian
Slough
2.0
50.0
32.0
29.0
1.5
75.0
1.0
94.0
Guadalupe
Slough
1.5
43.0
26.0
38.0
1.2
63.0
1.0
78.0
South Bay
0.9
31.0
17.0
13.0
1.6
43.0
0.9
54.0
Dumbarton
Bridge
1.8
49.0
22.0
19.0
1.6
51.0
0.7
53.0
South Bay
Streambed
Sediments
(1981-82)71
<1.0
30.0
18.0
40.0
0.3
62.0
Previous
South Bay
Data
(1970's)/2
0.8
50.0
29.0
0.25
1.8
167.0
Ln
NJ
/I Average of 1981-82 data from Coyote Creek,.Guadalupe River, and Los Gatos Creek for metals recoverable
from the stream bottom.
Adapted from: Larry Walker Associates/Kinnetic Laboratories (1985).
-------
Station
TABLE 5-10 CONCENTRATIONS OF TRACE METALS IN TRANSPLANTED MUSSELS
Trace Metal (UR/R)
Silver
0.12
1.08
1.40
1.13
1.86
4.35
Alum.
375
350
641
689
660
750
Arsenic
5.25
6.26
6.18
7.57
7.69
6.66
Cadmium
4.4
10.6
9.4
12.1
18.9
12.6
Chromium
1.5
1.7
2.6
2.8
3.0
3.6
Copper
6.4
12.1
10.5
12.2
13.9
15.6
Mercury
0.17
0.26
0.28
0.39
0.44
0.56
Manj>an.
21.2
15.6
29.4
23.1
26.3
28.9
Lead
0.8
3.7
2.3
2.8
2.6
2.8
Selenium
1.98
2.15
2.57
3.79
2.92
3.11
Zinc
62
116
137
160
198
162
Tomale.s Bay (reference) 0.12
San Francisco Bay
Fort Baker
Treasure Island
Hur.cer's Point
San Mateo Bridge*
Redwood Creek*
Los Angeles Harbor
(reference) 0.69 403 9.4 1.7 6.7 0.328 18.1 12.5 163
Southern San Francisco Bay
SOURCE: Modified from Anatec, 1985.
Ln
I
ro
-------
Table 5-11
Comparison of Pesticide Contamination of Mussels in Three California Bays, 1979
Parameter
Aldrin
alpha-BHC
cis-Chlordane
trans-Chlordane
DDE
Dieldrin
Heptachlor
trans-Nonachlor
PCB
Los Anqeles Harbor
______
ND
16
19
790
5.9
0.73
16
370
( in ppb)
San Francisco Bay 2
2.7
56
50
94
36
0.68
70
850
San Dieqo Harbor
ND
ND
10
18
210
8.3
ND
23
1,400
1 SOURCE: Anatec, 1985
2 Native species taken from Redwood Creek
-------
5-25
Human health risks from toxic contamination in the South Bay depend
upon the following factors:
o Exposure to toxics fron fish/shellfish consumption
o Concentration of toxics in fish tissue
o Potency of the toxic chemicals consumed
In this section, we summarize the data we used in analyzing these factors
and the limitations of the analysis. Vfe then present and interpret the
results of our health risk calculations.
Ideally, to account properly for consumption we would need data on the
types of fish eaten, the amount consumed and the amount of the population
consuming contaminated fish. None of this information is available. As a
result, the assumptions that we make about fish consumption are very uncertain
and serve only to illustrate the level of estimated health risk if that
amount of fish was consumed. There are two types of fishing in the South
Bay - commercial fishing and sportfishing. Commercial fishing is known to
focus almost exclusively on shrimp. However, a series of PGIV7 upsets in
1979-1981 almost devastated the shrimp population. Currently, the shrimp
that are caught commercially are used primarily for bait, although some
ethnic groups have been known to consume these shrimp. Little is known
about sportfishing in the South Bay.
Because we lacked data on the types of fish consumed, the amount
consumed, and the amount of the population consuming fish caught in the
South Bay, our analysis used data on toxics in three species of fish and
applied two different assumptions on quantity of fish consumed. Using
these assumptions we examine the range of potential risk to those hypothetical
individuals who may be consuming significant quantities of local fish. We
discuss the limitations of these assumptions below.
Our analysis used data on concentrations of toxics in the tissues of
three species of fish/shellfish - shrimp, mussels and striped bass. We
were unable to calculate the risk from consumption of other fish species
fron the South Bay due to a lack of data on fish tissue concentrations of
toxic substances. However, the species for which we have data may be
a reasonable indication of the likely health risks from the regular
consumption of other contaminated fish from the South Bay.
Mussels, according to the California Mussel Watch Program, best reflect
environmental conditions in the Bay and estuarine waters. The use of
mussels as indicators of water quality is advantageous because of their
widespread distribution, their tendency to concentrate pollutants normally
found only in small concentrations in the water, their stationary nature,
and the ease of collection. Mussels are not ccnmonly eaten, but their
levels of contamination may be similiar to other shellfish such as oysters
and clams. Shrimp may be consumed by certain ethnic groups. Striped bass
are the only non-shellfish species for which we had tissue concentration
data, and this concentration data may be more similiar to other free swimming
species than shellfish. Because sampling stations are not located south
of Dumbarton Bridge, our bass data are for bass caught in the San Francisco
Bay Delta area.
Assumptions about the health risks from all fish species based on a
few representative species must be made with care. Numerous factors,
-------
5-26
including pH, temperature, salinity, particle size influence how much of a
toxic substance is ingested by the different species of fish/shellfish.
Accumulation of toxic substances in the cells of living organisms (bio-
accumulation) is also poorly understood. Scientists do not fully understand
the factors which influence why one plant or animal will bioaccumulate
large amounts of a substance and another very little. Also, metals can
take many forms; one form may be toxic to humans while another is not.
To calculate health risks, the IEMP was limited to analyzing chemical
substances for which there were available monitoring data. The tissue
concentrations of toxics in shrimp and mussels represent actual monitoring
data from the South Bay. We use the mean value for shrimp and mussel
tissue concentrations in our health risk calculations. The tissue
concentrations for bass are for bass caught throughout the San Francisco
Bay/Delta area. Because a suitable mean value for bass taken from the
South Bay does not exist, we have calculated health risk estimates for
the lowest and highest concentration of a substance found in bass tissues.
We use two estimates for daily fish consumption. The low estimate is
a standard EPA estimate of 6 5 grams of fish per day, and is equivalent to
eating approximately 5.2 pounds of fish per year. This estimate reflects the
amount of fish that the average American consumes. People who regularly eat
fish probably consume substantially larger quantites. Therefore, we also use a
higher estimate of 65 grams per day, or 52 pounds of fish per year, to
estimate the potential risk faced by persons consuming large amounts of
fish. This high estimate of fish consumption is most likely an extreme
upper-bound value and estimates of potential health risk should be interpreted
with this in mind.
Potential Health Risks,
We do not calculate aggregrate cancer risks because we lacked reliable
data on the overall consumption of fish from the South Bay. However, it
seems likely that the number of people eating fish from the South Bay is
very small due to limited accessibility and hence the aggregrate risk is
correspondingly small. We do calculate risk estimates for the hypothetical
individuals who regularly consume fish from the South Bay.
In Table 5-12 we present the carcinogenic potency values, strength of
evidence classifications and information sources for the suspected carcinogenic
substances found in the tissues of mussels, shrimp, and bass from the San
Francisco Bay. The potency values are from the EPA's Carcinogen Assess-
ment Group (CAG) and represent plausible upper-bound values. Arsenic is
the only proven human carcinogen and the other substances are considered
probable human carcinogens. However, the evidence for arsenic's carcinogenicity
via ingestion is the subject of some dispute. Some scientists believe that
arsenic may not be harmful, and may even be beneficial at low levels.
EPA's Office of Drinking Water has concluded that arsenic is not harmful at
low levels, and has proposed a Maximum Contaminant Level of 50 micrograms
per liter in drinking water. Because of the uncertainty regarding arsenic's
carcincgenity, we have estimated arsenic's risk under two different
assumptions - no carcinogenicity, and using the CAG score.
-------
5-27
TABLE 5-12
CARCINOGENIC POTENCY VALUES AND STRENGTH OF EVIDENCE
Substance
Arsenic-
Potency Value
[mgAg-dayJ(-l)
1.6 x 101
Source^
CAG, 1986
Weight of Evidence2
Chlordane
1.6
CAG, 1985
B2
DDT
3.4 x 1CT1
CAG, 1986
B2
Heptachlor
3.4
CAG, 1985
B2
PCB
4.3
CAG, 1984
B2
Potency estimate derived by EPA's Carcinogen Assessment Group (CAG).
The weight of evidence of carcinogenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's
categorization of levels of evidence of carcinogenicity, A = proven human
carcinogen; B = probable human carcinogen; C = possible human carcinogen;
D = not classifiable; and E = no evidence.
The Carcinogen Assessment Group has calculated a potency score for arsenic
based on epidemiological evidence of "Blackfoot Disease", a form of skin
cancer. However, some scientists believe that arsenic may not be carcinogenic
and may be a necessary element at low levels.
5-37
-------
5-28
Table 5-13 presents cancer risk estimates for a hypothetical individual
who regularly consumes fish fran the South Bay contaminated with carcinogenic
substances. Because of significant uncertainties in the underlying data
and assumptions, these estimates of individual cancer risk are rough approx-
imations of actual risk and should only be used in the context of the IEMP
analysis. While we have attempted to err on the side of overestimating
cancer risk to the exposed individuals, one possible way that we have left
potential risks out of the analysis is that there may be other unidentified
carcinogenic substances in the fish from the South Bay for which there are
no monitoring data.
PCB's in bass contaminated with the highest tissue concentrations
found in the San Francisco Bay/Delta area represent the highest calculated
cancer risk. An individual consuming 6.5 grams of bass per day, with the high
level of contamination, would have about a 1600 in a million chance of
contracting cancer. An individual consuming 65 grains of bass per day at
this level of contamination would have about a 16,000 in a million chance
of contracting cancer. The use of the lowest level of PCB's found in bass
tissues throughout the San Francisco Bay/Delta area would lower the
estimated risk to 80 in a million for individuals consuming 6.5 grams
per day, and 800 in a million for individuals consuming 65 grams per
day. The wide range in estimated risk under different assumptions - 80
to 16,000 chances in a million - illustrates the uncertainty about possible
exposure and suggests caution in interpreting risk estimates. Nevertheless,
the high end of the range implies a relatively high cancer risk.
Arsenic in shrimp represents the next greatest calculated cancer
risk. An individual consuming 6.5 grams of shrimp per day over a lifetime
would have an estimated increased probability of 0 to 610 in a million
of contracting cancer. An individual eating 65 grams of shrimp per day
would face an increased probability of contracting cancer ranging from 0
to 6,100 in a million. These figures are similiar to the estimated range
of skin cancer risks resulting form arsenic in drinking water. The upper
end of the range is relatively high in comparison with other IEMP risk
calculations. However, substantial uncertainty exists as the toxicity
of arsenic at low levels, as discussed previously in this chapter. The
other substances in fish are estimated to be much less significant in
terms of cancer risk than arsenic or PCB's.
In Table 5-14 we present estimated thresholds for possible non-cancer
health effects for selected substances found in the tissues of mussels,
shrimp, and bass. The threshold values represent the values below which
the adverse effects are assumed not to occur. These thresholds are based
on both EPA-developed reference doses (RFDs; also referred to as Acceptable
Daily Intakes or ADI's) and additional health effects thresholds estimated
by IEMD for this study. For a more complete discussion of non-cancer
effects, refer to Chapter 2.
In Table 5-15 we present a comparison of the estimated dose of toxics
found in contaminated mussels, shrimp, and bass with the lowest presumed
human threshold. Using our upper-bound,'high consumption scenario of 65
grains per day (i.e. one Ib of fish per week) non-cancer thresholds are
exceeded for cadmium in mussels and PCB's in bass (using the high concen-
tration data). Using the low consumption rate of 6.5 grams per day,
non-cancer thresholds are also exceeded for PCB in bass (using the high
concentration data). Possible non-cancer health effects for arsenic
-------
5-29
TABLE 5-13
LIFETIME RISK OF CANCER FOR INDIVIDUALS CONSUMING FISH/SHELLFISH
CAUGHT IN SOUTH SAN FRANCISCO BAY
Fish
Tissue
Individual Lifetime
Cancer Risk
Substance
Arsenic 5
Chlordane
DDT
Heptachlor
PCB
Species
Shrimp
Mussels
Mussels
Mussels
Mussels
Mussels
Bass (low)
(high)
Concentration
(mg/kg wet
weight)2,3,4
4.1 x KT1
3.3 x 10-1
2.8 x 1CT3
4.7 x 10-3
3.4 x 1(T5
4.3 x ID"2
2.0 x 10'1
4.0
IF
6. 5 grams
consumed daily
0 - 6.1 x 10-4
0 - 4.9 x 10~4
4.2 x 10~7
1.4 x 10~7
1.1 x 10~8
1.7 x 10~5
8.0 x 10~5
1.6 x 10~3
IF
65 grams
consumed daily
0 - 6.1 x 10~3
0 - 4.9 x 10-3
4.2 x 10~6
1.4 x 10~6
1.1 x 10~7
1.7 x 10~4
8.0 x 10~4
1.6 x ID"2
NOTE: Because of significant uncertainties in the underlying data and
assumptions, these estimates of individual risk and disease incidence
are rough approximations of actual risk. They are based on conservative
estimates of potency which are more likely to overestimate risks than
underestimate risks.
1 Exposure estimates - estimates of consumption of fish/shellfish caught in
the South Bay - are uncertain. No data exists on local consumption of fish
from South Bay. See text.
FOOTNOTES CONTINUED ON NEXT PAGE
-------
5-30
2 Concentration data are fron the following sources:
Mussels and Shrimp: Anatec, Evaluation of Risk from Surface Water
Contamination, 1985
Bass: Southwest Fisheries Center, The impact of Estuarine degradation
and Chronic Pollution on Populations of Anadromous Striped Bass in
the San Francisco Bay-Delta, August, 1984
Shrimp and Mussels data are the mean monitored values. Bass data are
low and high monitored values.
Conversion factors from dry weight to wet weight: Mussels 5%, Shrimp 15%,
Bass 20%.
There is considerable controversy regarding the carcinogenity of arsenic
by ingestion. See Text.
-------
5-31
TABLE 5-14
NON-CANCER HEALTH EFFECTS FOR SELECTED SUBSTANCES
IN FISH/SHELLFISH CAUGHT IN SOUTH SAN FRANCISCO BAY
(All values are in rog/kg/day)
Substance
Health Effect
Presumed Human
Threshold Source2'3 Reference
Arsenic
Cadmium
Chlordane
Chrcmium+3^
Chronium+6
Copper
DDT
Liver
Neurobehavioral
Periph. Vascular
Skin
Reproductive
Teratogenicity
Multiple Organ
Arsenic Poisoning
Kidney
Liver
Mutagenicity4
Reproductive
Fetal
Reproduction
Liver
Weight Loss
Fetal
Liver
Kidney
Reproductive
Fetal
Other
Liver
Neurological
Reproduction
3.8 x
n
ti
n
3.8 x
II
It
4.8 x
5.0 x
11 I
1.4 x
2.5 x
1.5
4.9 x
6.0 x
1.0
II
II
4.9 x
4.9 x
4.26
4.8 x
II
3.7
10-3
II
M
n
10-3
n
n
10-2
10-4
1
10-1
10-2
10-5
10-3
10-3
10-3
10~4
n
IEMD
IEMD
IEMD
IEMD
IEMD
IEMD
IEMD/RFDV
IEMD
EPA/RFD1
IEMD
IEMD
IEMD
IEMD
IEMD/RFCMV
IEMD
EPA/RFC^
IEMDUV
IEMD
IEMD
IEMD
IEMD
EPA/RFDV
EPA/RFDV
IEMD
Clement
Clement
Env Health Crit #18
WHO 81
Perry et al., NIOSH 75
Hood et al. 77
Hood et al. 77
Heywood & Sortel 79
Zaldivar 77
Kjellstrom et al. 77
Friberg 50
Scharpf et al. 72
Schroeder & Kitchener 71
Comm. of European
Communities, 1981
EPA, 1985
EPA, 1985
Clement
Clement
Clement
Gale 78
Gale 78
Venugopal et al., 1978
1985
1985
Deichman, 1977
-------
5-32
TABLE 5-14 (continued)
Substance
Health Effect
Presumed Human Source2'3 Reference
Threshold
Heptachlor
Fetal
Liver
Reproduction
8.0 x 10~5
Mercury
Nickel
Fetal
Neurobehavioral
Fetal
Cardiovascular
Skin
Reproduction
3.1 x 10~4
II 11
1.0 x 10-2
EPA/RFD
EPA/RFDV
IEMD
EPA 80
EPA 80
Schroei
& Mitchener
1971
Schroeder et al., 1974
U.S. EPA 1985
Ambrose et al., 1976
PCS
Liver
Neurological
Skin effects
Kidney
Reproduction
3.1 x 10-4
EPA/RFD1
EPA/RFDUV
IEMD
EPA/RFDUV
1985
Kuratsune, 1972
1985
Reatonow, 1973
Battelle, 1985
Selenium
Selenosis
Liver
Neurobehavioral
Kidney
Reproductive
Fetal
3.0 x 10~3 EPA/RFD Venugopal & Lucky 79
" " IEMD Natl Academy Sci 76
EPA 80
IEMD Smith et al. 36
Harr et al. 67
" IEMD Venugopal & Lucky 79
" " IEMD Schroeder & Mitchener 71
NAS 76
IEMD Schroeder et al. 71
Silver
Skin Discoloration 3.0 x 10~3
EPA/RFDV Gaul & Staud 35
Blumberg & Carey 34
East et al. 80
Zinc5
Reproductive
Gastric Problems
7.4 x ID"1 IEMD Schliker & Cox, 68
2.1 IEMD EPA, 80
FOOTNOTES ON NEXT PAGE
-------
5-33
TABLE 5-14 (continued)
For sake of brevity, negative evidence (i.e., where a laboratory test or
epidemiological study has been performed, but no evidence was found of a
health impact) is not reported here. More complete toxicological information
is available from IEMP.
"IEMD" indicates a threshold estimated by IEMD toxicologists and consultants
using existing literature and following EPA procedures.
"EPA/RFD" indicates a threshold derived from an EPA "Reference Cose" (RFD)
or "Acceptable Daily Intake" (ADI) level. ADIs and RFDs are estimated
no-effect thresholds that are intended to protect an individjal from the
most potent non-cancer chronic health effect. EPA has reviewed and verified
some of these thresholds internally, although they have not necessarily
been peer reviewed. Those which EPA has verified internally are indicated
by v; those which are unverified internally are indicated by uv.
Mutagenic effects cannot be expressed meaningfully in quantitative terms.
Therefore, we note positive evidence of mutagenicity by listing it (on this
table) as an effect.
Zinc and Chromium +3 are considered necessary elements at low doses.
-------
5-34
TABLE 5-15
COMPARISONS OF TOXIC DOSE LEVELS FROM CONTAMINATED FISH/SHELLFISH
CAUGHT IN SOUTH SAN FRANCISCO BAY WITH LOWEST PRESUMED HUMAN
THRESHOLD FOR NON-CANCER EFFECTS
(all units are in mg/kg/day, unless otherwise noted)
Dose Dose
Fish 1,2 Concentration 3 Concentration 3
Tissue IF IF
Concentration (6.5 grams (65 grams
Substance Species (mg/kg wet weight) consumed daily) consumed daily)
Lowest Presumed 4
Human Threshold
Arsenic
Cadmium
Chromium+3
Chromium"1"6
Copper
Chlordane
DDT
Heptachlor
Mercury
Mussels
Shrimp
Mussels
Shrimp
Bass (low)
(high)
Shrimp
Shrimp
Mussels
Mussels
Mussels
Mussels
Mussels
Shrimp
Bass (low)
(high)
0.33
0.405
0.63
0.085
0.036
0.26
1.11
0.72
0.78
0.0028
0.0047
0.000034
0.028
0.087
0.012
0.320
3.1 x 10-5
3.8 x ID'5
5.9 x 10~5
7.8 x 10~6
3.3 x 10~6
2.4 x 10-5
1.0 x 10~4
6.7 x 10-5
7.2 x 10~5
2.0 x 10~7
4.0 x 10~7
3.2 x 10~9
2.6 x 10~6
8.0 x 10~6
1.1 x 10~6
3.0 x 10-5
3.1 x 10-4
3.8 x ID'4
5.9 x ID"4**
7.8 x 10-5
3.3 x 10-5
2.4 x 10~4*
1.0 x 10-3
6.7 x 10~4
7.2 x 10-4
2.0 x 10~6
4.0 x 10~6
3.2 x 10~8
2.6 x 10~5
8.0 x 10-5
1.1 x 10-5
3.0 x ID'4*
3.8 x 10-3
5.0 x 10-4
1.0
4.9 x ID'3
4.26
4.9 x ID'5
4.8 x 10-4
8.0 x 10-5
3.1 x ID"4
-------
5-35
TABLE 5-15 (continued)
Substance
Dose
Fish Concentration
Tissue IF
Concentration 6.5 grams
Species (mg/kg wet weight) consumed daily
Dose
Concentrat ion
IF
65 grams
Lowest Presumed
Human Threshold
Nickel
PCB
Selenium
Silver
Zinc
1
Bass (low) 0.1 9.2 x 1CT6 9.2 x 10~5 1.
(high) 0.4 3.7 x 10~5 3.7 x 10~4
Shrimp 0.3 2.8 x 10~5 2.8 x 10~4
Bass (low) 0.20 1.9 x 10~5 1.9 x 10~4* 3.
(high) 4.00 3.7 x 10~4** 3.7 x 10~J**
Mussels 0.0425 3.9 x 10~6 3.9 x 10~5
Mussels 0.155 1.4 x 10~5 1.4 x 10~4 3.0
Shrimp 0.0795 7.3 x 10~6 7.3 x 10~5 3.0
Bass (low) 0.2 1.9 x 10~5 1.9 x 10~4 7.4
(high) 13.2 1.2 x 10~3 1.2 x 10~2
Mussels 8.1 7.5 x 10~4 7.5 x 10~3
Shrimp 8.7 8.0 x 10~4 8.0 x 10~3
Concentration data are from the following sources:
Mussels and shrimp: Anatec, Evaluation of Risk from Surface
Contamination, 1985.
Bass: Southwest Fisheries Center, The Impact of Estuarine
0 x ID"2
1 x 10~4
x 10~3
x ID"3
x ID'1
Water
Degradation and Chronic Pollution on Populations of Anadromous
2
Striped Bass in the San Francisco Bay-Delta, August, 1984
Data for shrimp and mussels are the mean monitored value. Data for
striped bass are the low and high monitored values.
Dose concentrations noted with * approach the lowest presumed human
threshold; those with ** exceed lowest presumed human threshold.
See Table 5-14 for all estimated thresholds for non-cancer effects.
-------
5-36
consumption include liver, neurobehavioral, peripheral vascular, skin,
reproductive, teratogenic, and multiple organ effects. Possible non-cancer
health risks from PCB exposure include liver, neurological, skin, kidney
and reproductive effects.
In addition, the threshold levels for cadmium and mercury are approached
in bass using the high consumption and tissue concentration assumptions.
For PCB's, the threshold is approached using the low contamination in
combination with the high consumption rate. All other values are below
the threshold levels.
Although our risk estimates are conservation (i.e. they are more
likely to overestimate risk rather than to underestimate risk) and
uncertain (given the lack of fish consumption data) our findings are
generally consistent with the concern over health risks from consumption
of fish caught in the San Francisco Bay. For example, shellfish from the
San Francisco Bay have been a recognized health risk for many years.
Shellfish beds have been closed since the 1940's due to bacterial con-
tamination, although some shellfish beds may be reopening on a trial
basis. In addition, a Department of Health Services health advisory
presently cautions pregnant women against eating striped bass because of
mercury contamination.
Conclusions
1. The South Bay has elevated levels of toxics substances in its
ambient environment. Ihe concentrations of toxics in water, sediments
and fish tissues appear to be higher than comparable values for pristine
environments and a few other industrialized areas.
2. The aggregate human health risks from toxics in the South Bay
appear to be very low. The main exposure route to toxics is probably the
consumption of fish from the South Bay. Although we do not have information
on the number of people who eat fish caught in the South Bay, the amount
of fishing that takes place is small due to limited accessibility, and
hence the total number of people exposed to toxics from the South Bay is
likely to be very small.
3. Regular consumption over a long period of time of fish caught
from the South Bay may pose significant health risks. Although we
lacked data on consumption of fish in the South Bay, the health risk
estimates derived from the concentrations of toxics in mussels, shrimp
and bass indicated that long-term consumption of large amounts of fish
(65 grams daily or one Ib a week) caught from the South Bay may pose a
relatively high risk of cancer. Such consumption may also expose people
to a number of non-cancer effects. In order to assess more accurately
the present level of risk, information is needed on the number of people
consuming fish from the South Bay, and the amount and species of fish
that they are consuming.
-------
REFERENCES
Anatec Laboratories Inc. Evaluation of Risk from Surface Water Contamination
July 1985. '
Larry Walker Associates and Kinnetic Laboratories, Inc. South Bay Dischargers
Authority Water Quality Monitoring Program, Third Year Technical Report
December 1983 - November 1984, August 1985
National Marine Fisheries Service, Southwest Fisheries Center Hie Impact
of Estuarine Degradation and Chronic Pollution on Populations of
Anadromous Striped Bass in the San Francisco Bay-Delta, California.
August 1984. Administrative Report T-84-01.
-------
CHAPTER SIX
CONCLUSIONS
-------
CHAPTER SIX
CONCLUSIONS
I Introduction 6-1
II Limitations and Uncertainties 6-1
III Results of Stage I Risk Assessment 6-4
IV Next Steps 6-22
-------
CHAPTER SIX
CONCLUSIONS
Previous chapters have presented background information on the Santa Clara
Valley IEMP, on the comparative risk assessment methodology used in this study,
and on exposure and risk estimates in air, drinking water, and the regular
consumption of fish caught in the South Bay. This chapter summarizes and
compares estimates of risk from different chemicals, sources, and exposure
pathways, and discusses overall conclusions.
The IEMP analysis initially identified about 50 chemicals of potential
concern in the Santa Clara Valley. For this report, we estimated exposures and
risks for 41 pollutants all those for which we could find sufficient evidence
of exposure and for which we had toxicological data.
Sources of data and analytic methods varied, depending upon available
monitoring data, expected changes in regulation, and other factors. In many
cases, we have examined alternative assumptions for key, but uncertain, variables;
this results in ranges, as well as point estimates, of risks.
We have estimated several types of health risks. For exposures that could
increase the risk of cancer, we made conservative (i.e., pessimistic) estimates
of: (1) possible incidence, or the number of additional people who may contract
the disease; (2) average individual risk, or the increased likelihood that a
typical individual will contract the disease over the course of a lifetime as a
result of a particular type of exposure; and (3) risks to highly - or most -
exposed individuals (MEIs). For health effects other than cancer, we compared
estimated exposure levels with an estimated reference dose, or no-effect threshold,
below which exposure is judged to be safe.
In this chapter we present a summary of the more important conclusions,
and some recommendations on issues to pursue in the next stage of the IEMP.
(For a more detailed discussion of conclusions, see Chapters 3 through 5.)
Anong the more important ways in which we present our estimates of risks is by
comparing risks by pathway (i.e., air vs. drinking water); by pollutant (i.e.,
benzene health risks vs. risks from arsenic); and by type of source (i.e.,
motor vehicles vs. industrial point sources). Breaking down risks estimates
in this way increases insight into the nature of environmental health threats,
and allows us to define problems appropriately (i.e., should we be concerned
about benzene, industrial sources or drinking water?).
We first discuss the limitations and uncertainties in this analysis. This
includes limitations in the scope of the IEMP, and in the analysis. We also
discuss how the results are best interpreted given the goals of Stage One.
LIMITATIONS AND UNCERTAINTIES
An understanding of the uncertainties and limitations that underlie the
IEMP analysis is critical to a proper interpretation of its results. Limitations
in the scope of what was studied, and uncertainties in both the exposure and
toxicology data, argue against taking the estimates too literally. Nevertheless,
-------
6-2
decision-makers must often act now to protect against health threats from
toxic chemicals and cannot afford to wait for scientific certainty. The IEMP
analysis uses the best information available today to estimate health risks
from toxics so that decisions that cannot wait will be as informed as possible.
Limitations in Scope
The reader should recall that this analysis does not directly examine
disease incidence in the local population and attempt to link it with environmental
exposure. Because the analysis is not an epidemiologic study, it is not intended
to and does not answer questions such as what may have caused a statistically
higher rate of birth defects in the Los Paseos area. Instead, the IEMP attempts
to evaluate what health effects might result from current, and future environmental
exposures.
This analysis does not atterupt to estimate the health risks from all chemicals
that individuals may be exposed to in their daily lives. The IEMP did not
estimate risks from indoor air contaminants, nor those frcm occupational exposures.
(The IEMP has commissioned a scoping study on occupational exposures, which is
now in progress.) Similarly, risks frcm contaminants in food are not estimated.
Quitting analysis of these routes of exposure does not imply that they are
unimportant; indeed, it is quite possible that risks frcm any of these exposure
pathways could exceed risks frcm the exposures that we did examine. The IEMP
decided not to assess these exposure routes because of resource limitations
and because they are outside EPA's traditional purview and area of expertise.
The IEMP chose not to analyze exposure to and risks from conventional
pollutants in air and water (such as ozone and oxides of nitrogen and sulfur
in air, and oxygen-depleting substances and oil and grease in water) because
EPA believed it could make a more significant contribution by concentrating on
less well understood and less regulated toxic chemicals (largely organic chemicals
and heavy metals thought to be potentially hazardous at low levels).
Finally, the IEMP did not estimate risks from possible infrequent, accidental
releases of toxic chemicals, such as a major release of a toxic gas. (The
study did estimate risks from more frequent and predictable accidental releases,
such as tank leaks and chemical spills.) The probability and magnitude of
such an accident is very difficult to estimate, and the likely risk frcm such
an event is therefore difficult to quantify. The omission of such events from
this analysis does not imply that possible accidents are not an important
environmental and public health concern.
In sum, it should be clearly understood that this report is not an analysis
of health risks from all possible exposures to potentially dangerous chemicals
in the Santa Clara Valley.
Limitations in Exposure Data
Beyond these intentional limitations in scope, the study's exposure and
toxicological estimates are uncertain in a number of potentially important
ways. On the exposure assessment side, one limitation of the analysis is that
it did not exhaustively examine all sources and pollutants. While the IEMP
has tried to identify and assess risks frcm the most significant sources and
pollutants, it was unable to estimate exposure to some chemicals, such as
-------
6-3
arsine and phosphine, because of a lack of data. In reality, of course, chemicals
not included in the study may also pose some health risk.
Even where exposure data were available, those data varied significantly
in their quality. Thus, the resulting exposure estimates vary in their reliability.
Those based on extensive monitoring, such as for trihalomethanes and inorganic
substances in water, are probably fairly good. Those based on less extensive
monitoring, such as for metals in air (based on a single long-term monitoring
station) are somewhat less reliable.
Exposure estimates derived from modeling also vary in their reliability.
Estimates of exposure to toxic organic chemicals in air, calculated using a
dispersion model, are primarily dependent on the ouality of the emissions
estimates and other factors such as meteorological data. The range of possible
error for most pollutants is probably well under an order of magnitude. The
analysis of the future risks from groundwater contamination, which relies
heavily on engineering assumptions and modeling of future events, is more
uncertain. Vtfiere there are significant uncertainties, such as in the groundwater
exposure analysis, we have attempted to make assumptions that are likely to
err on the side of overestimating possible health impacts. In addition, we
have performed sensitivity analysis of particularly important variables, such
as possible chemical reactions and the effectiveness of regulatory actions.
Limitations in Toxicology Data
Estimates of the potential health effects of particular chemicals are
designed to be conservative (i.e., more likely to overestimate toxic health
effects than to underestimate them) in several ways. Health effects observed
in laboratory animals are assumed to be a reasonable indicator of potential
effects in humans. In converting the animal data to predicted human responses,
and in extrapolating from high doses to low doses, EPA uses models that yield
a plausible upper-bound estimate of potency rather than a "best guess" estimate.
On the other hand, many substances of potential concern have never been
evaluated scientifically, or have not been evaluated in sufficient detail to
allow estimation of effects on humans. For example, lead (present in air,
water, and dust) is thought to pose a health risk to children at ambient levels;
currently, however, EPA has no established way of estimating individual risks
or numbers of possible cases. EPA is likely to be aware of the dangers from
many of the most potent chemicals, since the evidence for their toxicity will
typically be the most obvious; however, it is possible that some chemicals
about which we currently know little may someday be demonstrated to be toxic.
Because of the many uncertainties and potential omissions, it is impossible
to say whether the total risk estimates presented here are over- or underestimates
of total toxic health risks from pollutants in air and drinking water. For
those chemicals for which the IEMP was able to make quantitative estimates of
exposures and risks, it is more likely that risks are overestimated than under-
estimated. To the extent that toxic chemicals about which we currently know
little have been left out, risks may be underestimated. The value of the IEMP
methodology is that it allows an evaluation and comparison of the health risk
fron chemicals and pollution sources about which we know something. Management
-------
6-4
of these risks, based on the best current information, can proceed, while
research continues on the effects of chemicals about which little is currently
known.
RESULTS OF STAGE I PISK ASSESSMENT
A brief summary of estimated environmental exposures is presented below,
followed by a presentation of estimated health risks from environmental toxics.
Exposure
Detailed estimates of exposure to toxic chemicals are too lengthy to
present in this summary; the interested reader is referred to the full report.
In air, dispersion modeling of toxic organic substances indicates that pollutant
concentrations are generally highest in the northern part of the study area,
which is more industrialized and more heavily populated. Monitoring data for
toxic metals indicates that they are present in Santa Clara Valley's outdoor
air at low levels - in some cases, the lower end of the range of concentrations
is below the detection limit for the analytic eguipment used.
Overall, concentrations of air toxics modeled or monitored in Santa Clara
Valley appear to be similar to or lower than pollutant concentrations typical
of urban areas. Estimated average concentration levels for irost chemicals
examined were below five micrograms per cubic meter (ug/m^).
Estimated exposures to most-exposed individuals (MEIs) near sources of
air toxics (such as semiconductor facilities, dry cleaners and traffic inter-
sections) were typically five to one hundred times higher than the average
concentration levels. The difference between average and MET exposures was
greater for chemicals such as ethylene oxide and chloroform whose emissions
were dominated by a few point sources, and less for chemicals such as xylene
and toluene, which are emitted by many dispersed sources.
About half the population in Santa Clara Valley is exposed to trihalo-
methanes in treated drinking water, at levels that are fairly typical for
disinfected water (about 20-80 micrograms per liter (ug/1)). Highly-exposed
individuals are estimated to be exposed to THM levels at the high end of this
range, but below the 100 ug/1 standard.
Thirty-six public wells and about 56 known private wells have been affected
by industrial chemicals. In the majority of cases where a source has been
identified, the pollution has resulted from leaking underground tanks or
associated chemical spills. Some operating public wells are serving water
containing 1,1,1-trichloroethane, perchloroethylene, 1,1,2-trichloro-l,2,2-
trifluoroethane, carbon tetrachloride, and a few other chemicals in the low
parts per billion, well below current state drinking water standards. (The
highest concentration level recorded at an operating public well is seven parts
per billion, or ug/1). About 129,000 people are currently drinking water from
public wells with low levels of contamination.
Recent testing by the County Health Department of 171 private wells found
that about 8% of the wells were affected by detectable contamination by synthetic
-------
6-5
organic chemicals, and that almost 40% were affected by bacteriological contam-
ination (i.e., they were unsanitary). Wells were not selected by random sanpling,
so these figures are not necessarily representative of other private wells in
the Valley.
Modeling of possible future drinking water contamination under conservative
(pessimistic) assumptions yielded estimated exposures significantly higher than
current levels, and included some pollutants (such as gasoline constituents)
not yet seen in drinking water wells. Concentration levels were estimated to
be significantly higher at private wells than at public wells, because public
wrlls benefit from greater regulatory and natural hydrogeologic protection.
The IEMP estimated that future exposure to contaminated groundwater sources of
drinking water could affect 15% of the population, in addition to those already
affected.
In the southern part of the county, several public wells and an unknown
number of private wells contain levels of nitrates that are above state and
Federal standards. Large systems exceeding standards are under order to comply
by 1988. Little evidence was found of pesticide contamination of drinking
water, either in local groundwater or in imported surface water.
Health Risks
1. Overall Cancer Risk; EPA's findings suggest that the estimated cancer
risks from the toxic chemicals and sources studied are apparently a small
proportion (well below one percent) of total cancer cases in the Valley.
Since any level of exposure to a carcinogen is assumed to pose some risk, all
1.4 million residents of the Santa Clara Valley are projected to face some
level of increased cancer risk as a result of environmental exposure. EPA
estimated that exposure to the pollutants and sources examined may be responsible
for about four cases of cancer per year; an estimated 3,600 cases of cancer
occur annually in Santa Clara County.1 This finding, although tentative,
provides an important perspective on health risks from toxic chemicals in
the outdoor air and drinking water in the Santa Clara Valley as compared to
other possible means of exposure to toxic substances, such as smoking, diet,
occupation, and indoor air. However, it is important to keep in mind that
this study examined a relatively small nunber of known toxic chemicals; exposure
to many thousands of chemicals in the air and drinking water about which we
know little may also be a source of significant, although currently unknown,
health risk.
Average individual cancer risk estimates for typical individuals exposed
to toxics in both air and drinking water indicate a potential increase in
cancer probability of about 200 in a million ov°r a lifetime. This estimate
of increased risk is the projected cumulative risk for exposure to all sources
and pollutants examined. Projected individual cancer risk is a small proportion
of the total lifetime cancer risk for an average person of about one in four
(250,000 in a million). Of course, individuals who are particularly highly
1 Ratio of estimated cancer cases to estimated cancer deaths, from 1983
national data from the American Cancer Society: 1.93. Cancer deaths in Santa
Clara County, 1984: 1,879 1,879 cancer deaths X 1.93 cases/death = 3,626
estimated cancer cases in Santa Clara County.
-------
6-6
exposed to chemicals, by virtue of their proximity to a source or for some
other reason, may face significantly higher-than-average cancer risk. (Risks
to such highlyexposed individuals are discussed below.)
2. Non-cancer Risks; EPA estimated that a about 10% of the population in
Santa Clara Valley may be exposed to chemicals at levels high enough to pose a
risk of effects other than cancer. The population estimated to be at risk of
non-cancer health effects due to exposures above no-effect thresholds is shown
in Table 6-11-
The IEMP estimated that exposure to benzene in the air could pose an
increased risk of lowered blood cell counts to about 100,000 people in the
Santa Clara Valley. This exposure is the most widespread exposure, at a level
above an estimated no-effect threshold, of any chemical examined. Benzene is
released primarily by vehicles.
In the southern part of the Santa Clara Valley, nitrate contamination of
yioundwater supplies of drinking water is above threshold levels estimated to
pose an increased risk of methemoglobinemia, or blue baby syndrome, to infants.
The IEMP estimates that up to 50 or 100 babies may, at any one time, be exposed
to nitrate levels high enough to pose risk.
In addition, under some assumptions about the way groundwater contamination
may affect drinking water supplies, the IEMP projects that several hundred
people who drink from private wells could be at increased risk of a variety of
effects - including birth defects and neurobehavioral, cardiovascular, liver,
blood and kidney effects - from industrial contaminants from tank leaks and
spills. Concentration levels in public well water are projected to remain
below no-effect thresholds, even under conservative assumptions.
1 TCA has not demonstrated any teratogenic potential in published studies
conducted using rodent species. Therefore, the IEMP base-case analysis assumes
that exposure to TCA poses no risk of fetal effects. An unpublished study.
which has not undergone scientific peer review, reports fetotoxic effects
(cardiac malformations) in rat pups exposed in utero to TCA (Dapson et al.,
1984). In order to assess the importance to Santa Clara Valley residents of
further research on this issue, the IEMP uses the Dapson study to examine the
possible impact of TCA under the alternative assumption that exposures above an
estimated threshold based on that study's results could pose the risk of fetal
effects. THE SENSITIVITY RESULTS SHOULD NOT BE INTERPRETED AS INDICATING
WHETHER OR NOT A RISK IN FACT EXISTS; EPA RECOMMENDS AGAINST USING THIS INFORMATION
FOR RISK MANAGEMENT DECISION-MAKING OR REGULATORY ACTION. Under this alternative
assumption, the IEMP projects that about 3,000 people, mostly those using
private wells, could be exposed to levels of TCA in their drinking water that
exceed the estimated threshold. In addition, most-exposed individuals downwind
of an industrial facility are projected to be exposed at levels above the
estimated threshold in the air. These findings suggest that more research is
appropriate, both on actual levels of exposure and on TCA's potential adverse
effects. The National Toxicology Program has commissioned a project to repeat
the limited Dapson study; results are expected in Fall of 1986.
-------
6-7
TABLE 6-1
POPULATIONS ESTIMATED TO BE AT RISK OF NON-CANCER HEALTH EFFECTS
IN SANTA CLARA VALLEY, BY POLLUTANT
POLLUTANT
EXPOSURE POTENTIAL
PATHWAY HEALTH EFFECTS
POPULATION
EXPOSED ABOVE
NO-EFFECT
THRESHOLD
PRIMARY
SOURCE(S)
VOLATILE ORGANIC
CHEMICALS
Benzene
1,1 Dichloro-
ethylene
Methylene
Chloride
1,1,1 Trichloro-
ethane 2
Trichloro-
ethylene
Vinyl
Chloride
Air Blood
Water Liver, kidney
Water Liver, fetal
Water Liver, neuro-
behavioral
Water Liver, neuro-
behavioral
Water Liver, kidney,
cardiovascular
100,000 Motor Vehicles
20 - 340 Underground Tanks
0-50 Underground Tanks
10 - 100
0-10
0-10
Underground Tanks
Underground Tanks
Underground Tanks
METALS AND
INORGANIC SUBSTANCES
Nitrates
Water
Blue baby
syndrome
50 - 100 Fertilizer, Septic
Tanks
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF THE POPULATIONS POTENTIALLY AT RISK OF DISEASE ARE ONLY ROUGH
APPROXIMATIONS. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND CHEMICAL
TOXICITY. See Text. UNLIKE CANCER RISK ESTIMATES IN THIS REPORT, THESE ARE
ESTIMATES OF POPULATIONS EXPOSED AT LEVELS THAT MAY POSE HEALTH RISK; THEY ARE
NOT ESTIMATES OF POSSIBLE NUMBERS OF CASES OR PROBABILITY OF DISEASE.
1 In many cases, risks for most-exposed individuals (MEIs) were estimated
without estimating populations involved. Such risk estimates are presented
in table 6-5.
2 The IEMP conducted sensitivity analysis on TCA for possible fetal effects. See
footnote to text.
-------
6-8
Substantial evidence exists that lead may cause toxic effects, including
blood effects and decreased IQ, particularly in children, who are roost sensitive
to it. Lead is present in air, dust and water, as a result of combustion of
leaded gasoline, use of lead solder in pipes, and other sources. The IEMP was
unable to calculate risks from lead exposure in this analysis because of a
lack of an accepted EPA method for doing so. The IEMP hopes to estimate health
risks from lead in Santa Clara Valley as a part of follow-on work in Stage II.
It is important to note that exposure to toxic chemicals in the air or
drinking water may pose some health risk at levels below estimated thresholds
if exposures from other sources - such as diet or occupation - are significant.
Even in this instance, comparisons with estimated thresholds provide a useful
indication of the significance of the portion of exposure due to outdoor air
or drinking water. Environmental exposures at or near estimated thresholds
are likely to pose a more significant added risk than exposures well below a
threshold.
3. Risks by Exposure Route; outdoor air and imported surface supplies
of drinking water appear to be the major exposure routes by which toxic
contaminants in the ambient environment are likely to affect most people. The
estimated breakdown of cancer risk by exposure route is shown in Table 6-2.
Estimated health risk through different exposure routes (e.g., air or
drinking water) generally reflects the extent of exposure to toxic chemicals
through those routes. Exposure to air toxics is the most widespread; everyone
breathes the air and all 1.4 million Santa Clara Valley residents are estimated
to be at some increased health risk from toxic air pollutants. Not surprisingly,
toxic chemicals in the air are estimated to pose the most significant health
risks among the exposure routes studied, over two estimated additional cancer
cases per year.
Cancer risks from imported drinking water supplies are estimated to be
somewhat lower than those from air sources - slightly over one additional case
per year. This estimated risk results primarily from exposure to disinfection
by-products, to which half the Valley's population is exposed. (See Conclusion
5 for more details).
One of the more striking findings of this study is that overall risks
from consumption of contaminated groundwater are estimated to be low ( about
1-2% of the cancer risk among the sources examined in this study, or about one
additional cancer case every 15 to 30 years). Estimated cancer risks from
current levels of exposure at public wells are lower: one estimated additional
cancer case every 800 years. The primary reasons for this finding of relatively
low risk from groundwater are that natural hydrogeologic protection and a
number of regulatory programs and voluntary actions in effect or soon to go
into effect are expected to limit most people's exposure. The IEMP estimates
that no more than about 25% of the people in the Santa Clara Valley are likely
to be exposed to groundwater contamination, compared to about 100% to air
contaminants and about 50% to trihalomethanes in surface water. It is important
to note that while our analysis of future groundwater contamination has substantial
uncertainties, this conclusion of relatively low health risks holds up under a
wide range of alternative assumptions and appears fairly solid.
-------
6-9
TABLE 6-2
ESTIMATED INCREASE IN CANCER INCIDENCE
IN SANTA CLARA VALLEY, BY EXPOSURE PATHWAY
EXPOSURE PATHWAY
POINT ESTIMATE
OF ANNUAL
INCREASE IN CANCER
INCIDENCE (Range)
WEIGHT OF EVIDENCE
FOR CARCINOGENICITY l
Air
2.2 (0.8 - 7.8)
A-B2
Drinking Water
Surface Water
Groundwater
1.3 (1.3 - 8.3)
0.06 (0.04 - 0.3)
A-B2
A-C
TOTAL
3.6 (2.1 - 16.4)
A-C
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS
OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND
POTENCY, AND ARE THEREFORE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE
THEM. See text.
1. The weight of evidence of carcinogenicity for the compounds included in
the analysis varies greatly, from very limited to very substantial. According
to EPA's categorization of levels of evidence of carcinogenicity, A = proven
human carcinogen; Bl = probable human carcinogen (limited human evidence);
B2 = probable human carcinogen (insufficient human evidence but sufficient
animal evidence); C = possible human carcinogen D = not classifiable; E= ;
evidence.
-------
6-10
The key hydrogeologic factor is the presence of an aquitard, or clayv
layer, over much of the Valley protecting public drinking water sources.
While this clay layer has, for the most part, prevented contamination near the
surface fron reaching deep drinking water supplies, there is concern that such
contamination could occur either through abandoned wells that may function as
conduits, or through faults in the confining layer itself. The recent discovery
of deep groundwater contamination in Mountain View (which has not yet affected
public drinking water wells) provides the first strong evidence of contaminant
transfer through conduit wells in the Santa Clara Valley. This finding is
consistent with the IEMP analysis, which suggests that conduit well transport
is likely to be more significant than contamination through the major clay
confining layer itself, and that a number of public wells may eventually be
affected in this way.
One important set of regulatory programs estimated to reduce groundwater
contamination and human exposure are the local Hazardous Materials Management
Ordinances, which have became models for hazardous materials control in other
areas. These ordinances reduce contamination at the source by requiring ground-
water monitoring near underground tanks, improved tank construction standards,
and better chemical handling processes. Other important regulatory and response
actions include clean-up actions at existing contamination sites, public drinking
water well monitoring for a broad range of organic chemicals (to be required
annually), and a policy of closing any public well contaminated above state
action levels. Voluntary actions taken by firms, such as underground tank
replacement and improved handling procedures, are also likely to reduce future
groundwater contamination.
Analysis of the effectiveness of all of these programs indicates that, in
combination, they may reduce health risks by roughly a hundred times (e.g.,
risks with these programs in place may be only 1% of what risks would have
been without them). Other programs, including efforts to seal abandoned wells
that may act as contaminant conduits, and efforts to monitor and protect private
wells, are also likely to reduce health risks from groundwater contamination.
Exposure through the outdoor air and drinking water is direct, as people
take in pollutants through breathing or drinking. Contamination affecting the
San Francisco Bay and local surface streams, by contrast, was judged to be
only indirectly related to human exposure, largely through body contact or
fish consumption. Exposure through these routes appears to be relatively
small by comparison with air and drinking water exposure.
Most hazardous wastes are exported from the Santa Clara Valley for recycling
or disposal elsewhere, and thus pose little local risk. Those local risks we
could identify from hazardous waste storage and handling appear to be primarily
through groundwater contamination, and were analyzed under that exposure pathway.
Accidental releases, such as those resulting frcm transportation accidents,
also have the potential to affect soil and groundwater.
Although the IEMP explicitly examined a number of potential issues of
pollution transfer from one medium to another, none appeared to be very significant
in terms of public health risk in the Santa Clara Valley. For example, the
study estimated the possible toxic organic chemical air emissions from sewage
treatment plants, groundwater aeration/clean-up sites, and sanitary landfills.
Air emissions from these sources were estimated to be fairly small in comparison
to other sources of toxic organic gases.
-------
6-11
4. The toxic environmental contaminants posing the most significant
health risks in Santa Clara Valley are, for the most part, the same as those
found in the other urban environments. A relatively small number of toxic
chemicals, including the trihalomethanes (primarily in drinking water), and
benzene, gasoline vapors, carbon tetrachloride, benzo(a)pyrene, chromium and
arsenic (primarily in air), account for about 92% of aggregate cancer risk
estimated in this stduy. National studies and data from other areas show that
estimated exposure levels in the Santa Clara Valley are similar to, and in
some cases lower than, ambient concentrations found in other urban areas. It
should be noted that the less-developed southern Santa Clara Valley has a
contamination problem typical of many agricultural areas: high nitrate levels
in the groundwater. A summary of estimated cancer risks by pollutant is presented
in Table 6-3.
As a class, volatile organic compounds account for the majority (about
58%) of the cancer risk estimated in this study. Heavier organic chemicals,
such as benzo(a)pyrene, comprise an estimated 20% of aggregate cancer risk.
Metals and inorganic substances account for about 22% of total estimated cancer
risk.
Arsenic risk in drinking water is a significant question mark in this
analysis. Arsenic accounts for as little as 0 to as much as 66% of total estimated
cancer risk, depending on assumptions about its toxicity. Some evidence exists
that arsenic in drinking water may cause a form of skin cancer known as "Blackfoot
Disease." Applying EPA's standard risk estimation techniques, the IEMP would
estimate up to about seven additional cancer cases a year from exposure to the
levels of arsenic found in Santa Clara Valley water (these levels are fairly
low in comparison to those found in many areas). Substantial disagreement
exists as to the careinogenicity of low levels of arsenic in drinking water,
however, and EPA's Office of Drinking Water believes that the levels of arsenic
found in Santa Clara Valley water are well within safe limits. This uncertainty
does not affect the estimate of lung cancer from airborne arsenic; the evidence
for this effect is much stronger.
The cancer risk from chromium in the air is another significant uncertainty
in this analysis. Monitoring data do not distinguish between hexavalent chromium
(thought to pose a risk of lung cancer) and other forms (not considered carcinogenic)
Depending on assumptions about the proportion of chromium that is hexavalent,
estimated cancer risk ranges from none to four additional cases per year.
Based on studies conducted elsewhere and a tentative identification, of local
sources, this study conservatively assumed that about 10% of airborne chromium
was hexavalent.
-------
6-12
TABLE 6-3
ESTIMATED INCREASE IN CANCER INCIDENCE
IN SANTA CLARA VALLEY, BY POLLUTANT
POLLUTANT POINT ESTIMATE OF ANNUAL INCREASE
(WEIGHT OF EVIDENCE)1 IN CANCER INCIDENCE (Range)
VDLATILE ORGANIC
CHEMICALS
Trihalone thanes (B2)
Benzene (A)
Carbon Tetrachloride (B2)
Gasoline Vapor (B2)
1,1, Dichloroethylene
Perchloroethylene (B2)
Ethylene Oxide (Bl)
Other (A-C)
TOTAL, VOCs
ORGANIC PARTICULATES
All Exposure
Pathways
1.3
0.3 (0.3- 1.2)
0.2
0.1 (0 - 0.4)
0.04
0.04
0.03
0.03
2.1 (1.9- 3.5)
Surface
Air Water
<0.01 1.3
0.3 (0.3- 1.2) <0.01
0.2
0.1 (0 - 0.4)
0.03
0.03
0.02 <0.01
0.7 (0.6- 1.9) 1.3
Groundwater
<0.01
<0.01
<0.0001
0.04
<0.01
0.01
0.06
(0.03 - 0.3)
BENZO(A)PYRENE
GROUP (B2)
0.7 (0.01-1.3)
0.7 (0.01-1.3)
METALS AND INORGANICS
Chromiun (A) **
Arsenic (A) ***
Cadmium (Bl) **
Other (A-B2)
TOTAL, METALS
0.
0.
0.
0.
0.
4
3
07
03
8
(0
(0.
(0.
(0
(0.
- 4.
2- 7.
04-0.
- 0.
2-11.
0)
4)
1)
07)
6)
0.4
0.3
0.07
0.03
0.8
(0
(0.
(0.
(0
(0.
- 4.
2- 0.
04-0.
- 0.
2- 4.
0)
4)
1)
07)
6)
0
0 (0-7)
0
0
0 (0-7)
0
***
0
0
0
TOTAL, ALL CHEMICALS
STUDIED
3.6 (2.1-16.4)
2.2 (0.8- 7.8)
1.3 0.06
(1.3 - 8.3) (0.03 - 0.3!
-------
6-13
FOOTNOTES TO TABLE 6-3
NOTE: BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS
OF ACTUAL RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND
POTENCY, AND ARE MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM.
See text.
1 The weight of evidence of carcinoqenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's categori-
zation of levels of evidence of carcinogenicity, A = proven human carcinogen;
Bl = probable human carcinogen (limited human evidence); B2 = probable human
carcinogen (insufficient human evidence but sufficient animal evidence);
C = possible human carcinogen; D = not classifiable; E = no evidence.
* The weight of evidence identified for trihalomethanes is that for chloroform
only.
** Neither (hexavalent) chromium nor cadmium is thought to be carcinogenic
in water. See chapter 4.
*** There is some dispute over the carcinogenicity of arsenic in water. See
text. Arsenic exposure listed under surface water is for combined exposure
to surface water and groundwater.
-------
6-14
5. The pollution sources posing the most significant overall health
risks appear to be similar to high-risk sources identified in other urban
areas. However, the sources of some of the most important environmental toxics
are uncertain. Identifying sources is important for risk management, since
pollution control decisions aimed at reducing risk must be directed at known
sources of risk. Table 6-4 presents a preliminary breakdown of cancer risk by
source type, making some assumptions about the sources of chemicals whose
origin is not well understood.
Most (about 77%) of the estimated risk from air exposure, particularly
for the toxic metals and organic particulates, is from sources that are only
tentatively identified. The AQMD does not maintain emissions inventories for
the metals and organic particulates, as it does for toxic organic gases.
Since consideration of control actions requires a knowledge of the sources of
risk, this report has identified the collection of data on the sources and
emissions of these substances as an important research need. In Stage II, the
AQMD, with assistance from EPA, plans to compile a metals emissions inventory.
The IEMP has done some preliminary analysis of the possible sources of many of
the substances of concern. However, more definitive source identification
would be required in most cases before risk management control actions can be
taken.
Preliminary analysis of likely sources suggests that the primary sources
of toxic chemicals in the air may be dispersed area sources, such as residential
heating and motor vehicles. These sources appear to emit the bulk of the
benzene, gasoline vapors, benzo(a)pyrene, and metals that are projected to
cause most of the air toxics risk. Industrial point sources do not appear to
be significant contributors to aggregate risk, with the possible exception of
the coal-burning cement plant. This pattern of many small and dispersed sources
suggests that it may be difficult to control major contributing sources so as
to reduce risk.
Surface water risks are dominated by hazards posed by trihalomethanes
resulting from water disinfection. The presence of trihalomethanes in treated
drinking water involves a trade-off of one form of risk for another: while
chlorination introduces chloroform and other potential carcinogens into drinking
water supplies, it protects the population from the otherwise much greater
risk of infectious diseases such as cholera and typhoid.
Although discontinuing disinfection is not a viable option, there are
other disinfection methods that reduce the formation of trihalomethanes. The
Santa Clara Valley Water District (SCVWD) has recently implemented one such
treatment method, chloramination, at its two local water treatment plants.
The IEMP projects that this change may reduce potential risks substantially.
In addition, the SCVWD has recently commissioned a study of still other disin-
fection techniques, such as ozonation, that might reduce risks further. Given
the apparent importance of these chlorinated organic chemicals relative to
other sources of toxic health risk, such analysis of alternatives may be appro-
priate, both in Santa Clara Valley and elsewhere.
"Background" contamination is contamination not linked to any known current
source. Such contamination may be from natural sources, such as minerals in
the soil, or from prior agricultural or industrial activities. Background
contamination, largely from persistent levels of carbon tetrachloride in the
-------
6-15
TABLE 6-4
ESTIMATED INCREASE IN CANCER INCIDENCE
IN SANTA CLARA VALLEY, BY SOURCE TYPE
SOURCE TYPE
(EXPOSURE PATHWAY)
[WEIGHT OF EVIDENCE]
ESTIMATE OF ANNUAL
INCREASE IN CANCER
INCIDENCE
PERCENT OF
TOTAL ESTIMATED
CANCER INCIDENCE
Drinking Water
Disinfection
(Surface Water) [B2]
1.3
36%
Fuel Combustion
for Residential
Heating
(Air) [A-B2]
0.63 - 1.1
18-31%
Motor Vehicles
(Air) [A-B2]
0.63 - 0.67
18-19%
**
Cement Plant
(Air) [A-B2]
0 - 0.5
0-14%
unknown Sources/Back-
ground Contamination
(Air) [A-B2]
0.2
6%
Other Area Sources
(Air) [A-B2]
0.15
4%
Other Point Sources
(Air) [A-B2]
0.1
3%
Underground Industrial
Tanks
(Groundwater) [A-C]
0.05
1%
Underground Fuel Tanks
(Groundwater) [A-B]
TOTAL, ALL SOURCES
STUDIED
<0.01
3.6
100%
-------
6-16
FOOTNOTES TO TABLE 6-4
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND ASSUMPTIONS,
THESE ESTIMATES OF DISEASE INCIDENCE ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL
RISK. THEY ARE BASED ON CONSERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND ARE
MORE LIKELY TO OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. See text.
1 The weight of evidence of carcinogenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's categori-
zation of levels of evidence of carcinogenicity, A = proven human carcinogen;
Bl= probable human carcinogen (limited human evidence); B2 = probable human
carcinogen (insufficient human evidence but sufficient animal evidence);
C = possible human carcinogen; D = not classifiable; and E = no evidence.
Chloroform is considered a probable carcinogen. The upper end of this
range reflects the possibility that other THMs are also carcinogenic.
See chapter 4.
Source identification for residential heating and cement plant is
preliminary and uncertain. See chapter 3.
This point estimate derives from an estimated range of 0.2 to 7.4 annual
incidence. The point estimate does not include the potential risk from arsenic
in drinking water. There is substantial disagreement as to the carcinogenicity
of low levels of arsenic in drinking water. Conservative assumptions of carcino-
genicity, developed by EPA's Office of Research and Development, suggest that
the levels in the drinking water in the Santa Clara Valley could result in up
to about 7.2 additional cases per year. However, EPA's Office of Drinking Water,
which is responsible for setting standards, believes that low levels do not
pose risk.
-------
6-17
air, is estimated to account for about 5% of total cancer risk from sources
and pollutants studied. However, if pessimistic assumptions about the qarcino-
qenicity of arsenic in drinking water are correct, the risk from background
contaminants increases to well over half of all estimated cancer risk in the
Santa Clara Valley.
The major groundwater contamination sources examined, underground fuel
and solvent tanks, are estimated to account for about 1-2% of the total cancer
risk among sources examined.
6. Some individuals, who live near pollution sources or are highly
exposed for other reasons, face toxic health risks that appear to be signifi-
cantly higher than average. Estimates of potential risk to these most-exoosed
individuals (MEIs) are shown in Table 6-5.
In contrast to the estimates of low overall risks from groundwater contam-
ination, people drinking from private wells appear to he vulnerable to potentially
significant levels of exposure and risk as a result of leaks from underground
tanks. Individuals who obtain water from private wells are more vulnerable to
risk because these wells are shallow (and thus not protected from surface
contamination by an intervening clay layer) and are not typically monitored.
Risks to individuals who may be highly exposed to industrial chemicals in
their private wells were estimated to be potentially higher than risks from
any other source examined. The estimated risk to the most-exposed individual
drinking from a private well is guite uncertain and should not be interpreted
literally. However, the potential vulnerability of this group is clear, and
this is the important conclusion for risk management. Current efforts by the
County Health Department to monitor some private wells appear to be a useful
first step in addressing this problem.
Persons living near a highly congested intersection were estimated to face
the highest individual cancer risk from exposures to toxic air contaminants such
as benzene. The risk facing individuals living near hospitals and exposed to
the sterilant ethylene oxide (ETO) was estimated to be nearly as large. These
potential risks from high exposures to air toxics are substantially lower than
the estimated risks for highly exposed individuals at private wells. The
comparatively high estimated risk near intersections reinforces the importance
of vehicles as a source of air toxics risk - both to the general populace and
to highly exposed individuals. Ethylene oxide from hospitals, on the other
hand, is not projected to be a major source of risk for most people but nevertheless
appears to pose comparatively high risks near the source. Because of uncertainties
about ETO emissions, and the possible reactivity of the chemical once released,
estimated emissions and exposure levels should be confirmed before control
actions are taken. Since use of ETO as a disinfectant is not unique to the
Santa Clara Valley area, this finding, if confirmed, may have implications for
other areas.
Although we lack actual consumption data, estimated risks to a hypothetical
individual regularly consuming significant quantities of contaminated fish or
shellfish caught from the South San Francisco Bay appear to be fairly high.
Concentrations of PCBs, pesticides, mercury, and other metals in shrimp, mussels,
and striped bass may pose a significant risk. Possible effects include cancer,
neurobehavioral, reproductive, kidney, and liver effects (estimated thresholds
for non-cancer effects are exceeded only under a "high" consumption estimate of
-------
6-18
TABLE 6-5
ESTIMATED HEALTH RISK TO MOST-EXPOSED INDIVIDUALS
IN SANTA CLARA VALLEY
EXPOSURE PATHWAY
& SOURCE TYPE
INCREASED
LIFETIME
CANCER RISK
(CHANCES IN
A MILLION)
POLLUTANT
(WEIGHT OF EVIDENCE)2
POTENTIAL
NON-CANCER
HEALTH EFFECTS3
AIR:
TRAFFIC
INTERSECTIONS
HOSPITALS
PHARMACEUTICAL
MANUFACTURER
300
200
100
COMPUTER EQUIPMENT 40
MANUFACTURER 4
INDUSTRIAL 30
FACILITY 4
Benzene (A)
Benzo(a)pyrene(B2)
Cadmium (Bl)
Ethylene Dibromide(B2)
Ethylene Oxide (Bl)
Ethylene Oxide (Bl)
Benzene (A)
Methylene Chloride (B2)
Methylene Chloride (B2)
Benzene (A)
Blood, fetal
Blood, fetal
FUEL PIPELINE
DRY CLEANERS
SEWAGE TREATMENT
PLANTS 4
GAS STATION PUMP
GROUNDWATER
AERATION 4
20
10
5
0.2
Benzene (A)
Perchloroethylene (B2)
Chloroform (B2)
Benzene (A)
Methylene Chloride (B2)
Perchloroethylene (B2)
Trichloroethylene (B2)
Benzene (A)
Gasoline Vapors (B2)
Methylene Chloride (B2)
Trichloroethylene (B2)
-------
6-19
TABLE 6-5 (cont.)
ESTIMATED HEALTH RISKS TO MOST-EXPOSED INDIVIDUALS
IN SANTA CLARA VALLEY
EXPOSURE PATHWAY
& SOURCE TYPE
INCREASED
LIFETIME
CANCER RISK
(CHANCES IN
A MILLION)!
POLLUTANT
(WEIGHT OF EVIDENCE)2
POTENTIAL
NON-CANCER
HEALTH EFFECTS3
GROUNDWATER:
UNDERGROUND TANKS 5
(AT PRIVATE WELLS)
20,000
1,1 Dichloroethylene (C)
Vinyl Chloride (A)
Perchloroethylene(B2)
Ethylene Dibromide (B2)
Methylene Chloride (B2)
Chloroform (B2)
Benzene (A)
Trichloroethylene (B2)
1,1,1-Trichloroethane
Liver, kidney,
Liver, kidney,
cardiovascular
Liver, fetal
Liver, neurobehavioral
Liver, neurobehavioral
*
FERTILIZER, SEPTIC
TANKS
Nitrates
Methemoglobinemia
(Blue baby syndrome)
SURFACE WATER;
DRINKING WATER
TREATMENT
BACKGROUND
100**
Trihalcmethanes (B2)
0 - 7,000*** Arsenic (A)
SOUTH SAN FRANCISCO BAY;
SHRIMP CONSUMPTION 7 0 - 6,000*** Arsenic (A)
MUSSEL CONSUMPTION 7 20 - 200 PCB (B2)
Chlordane (B2)
DDT (B2)
STRIPED BASS
CONSUMPTION 7
80 - 16,000
PCB (B2)
Cadmium
Mercury
Liver, neurobehavioral
kidney, reproductive
Kidney, reproductive,
liver, birth defects
-------
6-20
FOOTNOTES TO TABLE 6-5:
NOTE; BECAUSE OF SIGNIFICANT UNCERTAINTIES IN THE UNDERLYING DATA AND
ASSUMPTIONS, THESE ESITMATES OF INDIVIDUAL RISK AND DISEASE INCIDENCE
ARE ONLY ROUGH APPROXIMATIONS OF ACTUAL RISK. THEY ARE BASED ON CON-
SERVATIVE ESTIMATES OF EXPOSURE AND POTENCY AND ARE MORE LIKELY TO
OVERESTIMATE RISKS THAN UNDERESTIMATE THEM. See text.
1 Except in the case of underground tanks at private wells, estimated cancer
risk is for all pollutants combined from given source. For underground
tanks, estimate is for pollutant posing the greatest cancer risk. In each
case, pollutants for each source are listed in decreasing order of estimated
cancer risk.
2 The weight of evidence of carcinogenicity for the compounds listed varies
greatly, from very limited to very substantial. According to EPA's
categorization of levels of evidence of carcinogenicity, A = proven human
carcinogen; Bl = probable human carcinogen (limited human evidence);
B2 = probable human carcinogen (insufficient human evidence, but sufficient
animal evidence); C = possible human carcinogen; D = not classifiable;
E = no evidence.
3 Non-cancer health effects are reported only if exposures are above estimated
thresholds for such effects.
4 If TCA is assumed to be carcinogenic, total estimated cancer risk is at or
slightly above level presented.
5 Estimated impacts are for "high" release, base case; see chapter 4.
6 Estimated taste and odor threshold is very slightly below the estimated
threshold for blood effects.
7 Estimated risks for fish consumption are for a hypothetical individual
who regularly consumes contaminated fish or shellfish caught in the South
Bay. Assumed consumption is 5 to 52 pounds of fish per year. Note that
the IEMP has no actual data on the number of people eating fish from the
South Bay, although we believe that number is small.
8 Estimated exposure value for mercury in striped bass is just slightly
under the lowest estimated human threshold.
* The IEMP conducted sensitivity analysis on TCA for possible fetal effects.
See footnote to text.
** Risk for system with highest estimated average risk; some individuals
may be exposed to higher risks.
*** Considerable controversy exists as to the carcinogenicity of arsenic by
ingestion. See text.
-------
6-21
one pound per week of contaminated fish). We stress that these exposure estimates
are conservative, and that we have no data on the number of people consuming
contaminated fish from the south Bay. Nevertheless, these estimates suggest
that regular consumption of fish or shellfish from the south Bay may be unwise.
This finding is consistent with a health advisory issued by the State Department
of Health Services, warning pregnant women not to eat striped bass.
7. One of the more important implications of this analysis is that
groundwater contamination may be an economic and natural resource issue as
well as a risk issue. IEMP estimates of future risk depend on many actions
that we assume will be taken in the future. For example, the study assumes
that public drinking water wells will be closed when contaminated above action
levels and replacement supplies obtained; it also assumes that the Hazardous
Materials Management Ordinances will be implemented, although this has not yet
fully occurred. While IEMP projects that these actions will be largely successful
in controlling risk, they could be extremely expensive. The direct economic
costs of contamination prevention and response include the costs of tank replacement,
clean-up, monitoring and well closure. In addition, groundwater contamination
causes a potentially significant indirect natural resource cost: the loss of
clean, local groundwater.
The IEMP analysis illustrates the difference between drinking water health
risk and groundwater resource impacts. Under the rather pessimistic assumptions
used in this study, health risks to people drinking groundwater from public wells
are projected to be comparatively small. Yet, about 78 public wells serving
200,000 people are projected to be affected by fuel or industrial contamination,
with 20% to 30% of the wells contaminated above state action levels.
Contamination above action levels requires well closure or treatment. In
some cases, contamination below action levels has led to removing a well from
service. Clean-up of contaminant plumes can also have a significant impact on
the groundwater resource, as large quantities of groundwater are pumped, cleaned
and discharged to the Bay. This water must be replaced with recharge water
imported from the Sacramento Delta. While the IEMP estimates of the number of
wells likely to be affected are intentionally pessimistic, they clearly indicate
the importance of examining the natural and economic resource impacts, as well
as the health effects, of groundwater contamination and programs to address it.
Thus, the low aggregate risk estimates presented in this draft report do
not imply that groundwater contamination is not an important environmental
management issue. Despite the comparatively low estimated aggregate risks, it
may be appropriate to assess groundwater control and treatment options in
terms of their potential risk, cost, and resource impacts.
8. This study identified many scientific uncertainties and data gaps
that may be appropriate research priorities for regulatory agencies or others.
A few of the most important include:
0 Hydrogeology: A better understanding of Santa Clara Valley hydrogeology-
in particular the effectiveness of the major clay confining layer or
aquitard - would improve the ability to protect the groundwater resource
effectively.
-------
6-22
0 Pollutant transport and transformation: In particular, better under-
standing of the speed with which fuel contaminants degrade after release
into the environment is critical to determining the importance of. leaking
fuel tanks as a groundwater contamination source.
0 Monitoring data: Two of the more critical uncertainties in this Stage
1 analysis - levels of organic chemicals in the ambient air and in
private wells - are being addressed by local agencies. Local data on
organic particulates in air would be valuable also; the IEMP plans to
sponsor the collection of such data as a part of Stage II.
0 Source data: Better information on sources of metals and organic parti-
culates in air is needed to assist in the development of risk management
strategies. The Bay Area Air Quality Management District, with EPA
support, will be compiling a metals emissions inventory in Stage II.
0 Non-Cancer effects: Development of a method of estimating possible
disease incidence for effects other than cancer would allow a more
complete analysis of toxic health risks. Some key chemicals of concern
in the Santa Clara Valley have been identified in this report. These
issues are being pursued within EPA and by scientific peer review groups.
Next Steps
This Stage I Report for the Santa Clara Valley Integrated Environmental
Management Project presents the results of the lEMP's comparative analysis of
toxic environmental health risks. Potential health effects were analyzed, and
exposure pathways, pollutants and sources compared in terms of health risk. A
draft of this report has been reviewed widely by EPA's two advisory committees,
the Intergovernmental Coordinating Committee and the Public Advisory Committee,
and by other interested agencies, scientists and individuals. It is now undergoing
scientific peer review by a group of scientists at Rutgers University.
The Stage I Report findings are intended to provide the basis for, Stage
II of the IEMP, which will focus on managing risks: identifying priority issues,
analyzing control options for dealing with those problems, and implementing
solutions. Stage II will also expand and improve upon the problem definition
developed in Stage I.
EPA, in consultation with its IEMP advisory committees, has developed a
Stage II workplan to guide the project's future work. This workplan identifies
risk management priorities, taking into account public concerns and ongoing
programs. It outlines research priorities, analyses of pollution control
options, and a management strategy that EPA and its local partners hope will
lead to discussions and actions that protect public health and the environment
more effectively.
-------
APPENDICES
-------
APPENDIX A: PUTTING RISKS IN CONTEXT
How can we decide whether or not the risks estimated for exposures to
these toxic chemicals are cause for concern? One way of putting into context
the estimated risks in Santa Clara Valley is to compare those risks to similar
risks in other areas. As we have already mentioned, exposures to (and hence
risks from) key toxic chemicals in Santa Clara Valley appear to be consistent
with, and in some cases lower than, exposures in other urban areas. Trihalo-
methane levels, for example, are consistent with those of other areas using
chlorinated water. Benzene levels, which result primarily from vehicle use
and fuel distribution, are expected to be similar in most urban areas.
Ambient concentrations of benzo(a)pyrene may be lower in Santa Clara Valley
than in many areas, because there is less combustion for heating purposes.
Some trace metals also appear to be on the low end of the typical range
because there are relatively few heavy industrial point sources burning
fossil fuels, which contain metals. A comparison of the ambient concentrations
of common pollutants in different urban areas is presented in Tables 3-36
and A-l.
Several studies have estimated cancer risks from toxic air contaminants
nationally or in a number of urban areas. Comparisons of the risks estimated
in these studies with those estimated for Santa Clara Valley are presented
in Table A-2. National estimates of the risks from indoor air exposure can
also be compared with risks from outdoor air exposures in Santa Clara Valley.
Such a comparison is presented in Table A-3.
Another way of putting these estimated risks in context is to compare
them with statistics on mortality and sickness from a variety of other causes.
In making this comparison, one should remember that the mortality reference
statistics are known to be accurate, while the estimates in this study are
conservative estimates of potential impact. The estimated 1.4 (range 1.3 -
8.6) additional cancer cases which may occur due to exposure to environmental
toxics in water, and the 2.2 (range .8 - 7.7) additional cases from toxics
in air can be compared to an estimated 3,600 cases of cancer that occur
annually in Santa Clara County.* It can also be compared to the overall
nunber of deaths in Santa Clara County in 1984, about 8,000. About 1,900 of
these deaths were from some form of cancer. The American Cancer Society
estimated in 1983 that 33% of cancer deaths are related to smoking.
The estimates of disease incidence made in this report thus provide
important information on the risks from toxic chemicals. However, the public
may not perceive the relative risks from different sources, both in this
report and otherwise, as being only a matter of the death or disease that
result. Their perceptions of risk may be colored by other factors, such as
the voluntariness, controllability, and familiarity of the risks, and by the
immediacy pf the consequences. In addition, one must realize that risks
* Ratio of estimated cancer cases to estimated cancer deaths, from 1983
national data from the American Cancer Society: 1.93. Cancer deaths in
Santa Clara County in 1984: 1,879. Thus, 1,879 cancer deaths x 1.93
cases/death = 3,626 estimated cancer cases in Santa Clara County.
A-l
-------
A-2
often accrue in connection with some congruent benefits that result from
the action being undertaken. Thus, all risks are not necessarily equal.
By implication, policy responses to situations involving risk may
not be simply a matter of addressing those risks that appear the largest.
Perceptions of risk, feasibility of control, available resources,
compensating benefits, and other factors in addition to estimated risk
may appropriately influence policies aimed at protecting public health.
-------
A-3
TABLE A-l
COMPARISON OF AMBIENT LEVELS OF SELECTED WATER TOXICS
IN SANTA CLARA VALLEY WITH LEVELS IN OTHER CITIES
All Values in Micrograms/Liter
SUBSTANCE
Tr i ha 1 ome thane s
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Zinc
SANTA CLARA
COUNTY !
4-80
0.3 - 15
<100 - <1000
0-10
1-8
1-40
0-6
1 - 10
1 - 20
1 - 90
Baltimore 2
50 - 54
10
20 - 30
1-2
1
10
1
2-5
1
na
National
Studies 3
na
10 - 1000
0 - 172
na
0 - 112
na
0.1 - 1.8
0-10
na
na
1 SOURCE: Table 4-11, Exposure to metals and minerals
2 SOURCE: Baltimore IEMP, 1984 memo from Versar
3 SOURCE: National Organics Monitoring Survey, 1977
-------
A-4
TABLE A-2
:OMPARISON OF ANNUAL CANCER INCIDENCE ESTIMATES FDR AIR TOXICS
IN SANTA CLARA VALLEY WITH ESTIMATES FROM OTHER STUDIES 1
Annual Incidence per Million Population
National Risk
POLLUTANT
Benzene
Chromium 3
Cadmium
Arsenic
Trichloroethylene '
Perchloroethylene
Ethyl ene Oxide
Carbon Tetrachloride
Chloroform
1,1 DCE
SANTA CLARA i
VALLEY
0.23
0.31
0.05
0.23
0.002
0.02
0.02
0.15
0.001
N/A
NESHAP 2 .1
Study
0.14
1.43
0.04
0.02
0.04
0.01
0.21
0.06
<0.01
<0.01
Ambient Air 2
Quality Study
1.02
1.05
0.06
0.26
0.08
0.10
N/A
0.19
0.07
0.27
35 County 2
Study
0.39
0.29
0.02
0.02
0.15
0.14
N/A
0.004
0.002
N/A
1 SOURCE: Table 3 -36, Estimated Annual increase in cancer incidence
from toxic Pollutants in Santa Clara Valley's Outdoor Air
(adjusted for population)
2 SOURCE: Elaine Haemisegger, et al., "The Air Toxics Problem in the United
States: An Analysis of Cancer Risks for Selected Pollutants,"
U.S. EPA, Office of Air and Radiation, Office of Policy, Planning
and Evaluation, May, 1985.
3 Methodologies may be inconsistent in their treatment of the amount of Cr+6
as a proportion of all chromium.
-------
A-5
TABLE A-3
CANCER RISKS FROM INDOOD AIR EXPOSURES
COMPARED TO RISKS FROM OUTDOOR AIR EXPOSURES
IN THE SANTA CLARA VALLEY
Pollutant
Incidence from Indoor
Air Exposures per
(per million people)-*-
Incidence from outdoor
Air Exposures in
Santa Clara Valley
(per million people)^
Radon
Passive smoking
Formaldehyde
Carbon Tetrachloride
Benzene
Chloroform
Perchloroethylene
Trichloroethylene
4 to 91
2.2 to 22
0.7
1.5
2.2
1.1
0.9
1.0
0.15
0.23
0.001
0.02
0.002
Source: Elaine Haemisegger, et al., "The Air Toxics Problem in the United
States: An Analysis of Cancer Risks for Selected Pollutants,"
U.S. EPA, Office of Air and Radiation, Office of Policy. Planning
and Evaluation, May, 1985.
Source: Table 3-36, Estimated Annual Increase in Cancer Incidence from
Toxic Pollutants in Santa Clara Valley's Outdoor Air
(adjusted for population).
-------
A-6
TABLE A-4
NUMBER OF DEATHS IN SANTA CLARA COUNTY
RESULTING FROM SELECTED CAUSES, 1984
NUMBER OF
CAUSE DEATHS
TOTAL- all causes 8045
Circulatory 3605
Cancers 1879
0 Trachea/bronchus/lung 455
0 Digestive,
incl. stomach 470
0 Breast 164
0 Leukemia 71
Accidents/Poisoning/Violence 665
0 Traffic accidents 219
0 Suicide 153
0 Homicide 74
0 Surgical misadventures 2
Cirrhosis of the liver 174
and other chronic liver disease
Diabetes 132
Tuberculosis 8
SOURCE: State of California, Department of Health Services, Center for
Health Statistics Deaths from Selected Causes, 1984
-------
APPENDIX B: NON-CANCER HEALTH EFFECTS METHODOLOGY
As referenced in the text, we have prepared additional information on the
non-cancer health effects methodology has been prepared to assist the reader
in this area of the assessment.
I. IEMD non-cancer health effect methodology - a general overview.
II. Application of the IEMD methodology.
III. FPA's Office of Research and Development's concerns on the limitations
of the IEMD approach.
IV. Summary data on exposures and observed or estimated concentrations
for selected chemicals being considered for Stage II.
V. Sample dose-response curves for non-cancer health effects for air
and water exposures.
-------
B-l
IEMP METHODOLOGY FOR NON-CANCER HEALTH EFFECTS; A GENERAL OVERVIEW
The Nature of the Problem
EPA is charged with controlling the levels of exposure to thousands of
chemicals in the environment. Etecisions about which chemicals to control are
based, in large part, upon the toxicity of individual chemicals. This is
because we want to control first those problems that pose the greatest risk,
and because the amount of resources we will apply to a problem should bear
some relationship to the health threat it presents.
EPA must therefore analyze the toxic effects of many chemicals. In practice,
this has been done in two quite different ways, depending on whether or not
the chemical has been identified as a carcinogen, a substance that may cause
cancer in animals or people. Toxicological research has led the Agency to
assume that exposure to any amount of a carcinogen is associated with the
possibility that some person in the exposed population may develop cancer.
For many other health effects, in contrast, the Agency assumes that there
is some level of exposure that causes virtually no harm. This level of exposure
is called a "threshold." In almost all people, exposures below the threshold
should not cause adverse health effects, while exposures above the threshold
may lead to such effects in some exposed individuals.
For each of these groups of chemicals, the Agency has adopted a standard
process for estimating how potent a substance might be. For non-carcinogens,
the Agency uses the concept of the threshold to determine Risk Reference Doses,
or RfDs (previously called ADIs).
For carcinogens, the process is more complex. The EPA Cancer Assessment
Group (CAG) is responsible for defining the risk from exposure to carcinogens.
The problem is that most of the information we have about the effects of carcinogens
relates to high doses, in animal studies or human occupational exposures.
Almost all environmental exposures occur at much lower levels. For years, the
scientific community has struggled with the problem of extrapolating effects
at low doses frcm information about high doses. Also, for many substances,
the only information we have is from animal studies. Here a further uncertainty
arises because scientists are uncertain about how to apply information from
animal experiments to predict the effects of carcinogens on people.
CAG must make assumptions based on the best available scientific evidence.
It uses complex mathematical models that are based on scientific understanding
of biochemical mechanisms to "fill in" the missing low-dose data. Intrinsically,
the assumptions and application of such models in developing estimates of low
dose effects result in some.uncertainty. In general, when making a choice
among such assumptions, concern for the public health drives Agency policy to
select the one that generally turns out to be "conservative" or pessimistic,
i.e., the one most protective of human health. The Agency, therefore, has
chosen to define the CAG scores as a "plausible upper bound." That is, if the
application of the so-called "CAG numbers" predicts, say, three cancers frcm a
certain exposure, we state that it is r.ot likely to be more than that, but may
be much less, or even zero.
-------
B-2
CAG numbers undergo substantial review within the Agency. In a similar-
fashion, the Agency has established a review process for RfDs. CAG numbers
are often reviewed further by EPA's Science Advisory Board. In this process
they are often modified as a result of additional research findings. Over-
time, CAGs and RfDs come to represent a consensus about what EPA and the broader-
scientific community think about the ability of various substances to produce
adverse health effects. As such they are used as the scientific underpinnings
of regulatory policy making, and although risk assessment is not perfect, it
is the best quantitative decision aid we have.
The Integrated Environmental Management Method
The purpose of the IEM method is to give decision-makers some way of
assessing the efficiency, or cost-effectiveness, of various pollution control
options. Essentially, the method converts a set of exposures to toxic chemicals,
in a particular geographic area, or stemming from a particular kind of source,
into a set of health effects related to those chemicals. These health effects
are expressed as the predicted incidence of particular diseases. We can then
estimate the impact that different control policies will have on the nature
and distribution of these health effects.
We use existing EPA risk assessments and models that can be appropriately
applied. For carcinogens, we use the plausible upper-bound CAG numbers to
convert a given exposure to a conservative estimate of potential "cases." We
also use the Agency RfDs as "thresholds'1 for non-cancer health effects. But
in addition, we attempt to define the magnitude of risk associated with exposure
to chemcials at levels above the RfD so that we can compare the effects of
many different chemicals, and thus say something about the relative effectiveness
of controlling them in terms of protecting public health.
In the IEM model, the use of the CAG and RfD values to estimate risks
or levels of concern about risk is relatively straightforward. These indicators
of risk have been used throughout the Agency for many years. The point of
departure from standard EPA approaches to risk assessment is our novel procedure
for estimating the incidence of non-cancer health effects. Just as the Agency's
regulatory programs associate levels of exposure with probable risk of cancer,
the IBM model associates levels of exposure with probable incidence of non-cancer
effects.
Again, it must be emphasized that this is an experimental approach for
estimating potential levels of risk resulting from exposure to toxic chemicals.
We hope that eventually such a technigue will enjoy broad consensus within the
Agency. The Office of Research and Development (ORD) considers this part of
our method a good first attempt to synthesize non-cancer health effects data,
but believes that the toxicology of many chemicals has not progressed sufficiently
to allow the guantitative prediction of human non-cancer health effects from
animal data. (A more detailed discussion of ORD's position is included below.)
Despite this objection, IEM staff believe that the approach is valid enough
to make broad comparisons of chemical toxicity when used in the kind of policy
planning that the IEM models were designed for. Thus we are in the first staqe
of what we expect will be an extensive dialogue with the scientific conmunity.
In fact, some of the more recent scientific literature addresses the feasibility
-------
B-3
of doing what the IEM model does. Without doubt, both our method and the status
of scientific consensus will evolve in the future.
Because scientific consensus does not yet exist as to how to estimate
possible incidence of non-cancer effects, this Report does not present estimates
of such effects. Instead, the Report presents estimates of people exposed
above RfDs and similar, estimated no-effect thresholds. This Appendix presents
lEMP's experimental dose-response curves for those chemicals estimated to be
above thresholds.
lEMP's dialogue with the scientific community has been formalized through
the formation of two different scientific panels that are conducting independent
peer reviews related to the IEMP methodology. Unfortunately, the timetables
for these review efforts - completion at the earliest by late fall 1986 - makes
it impossible for the Santa Clara Valley IEMP to incorporate the dose-response
non-cancer methodology in this revised Stage I x-eport.
The first review group, EPA's Science Advisory Board's Integrated Environmental
Management Subcommittee, has been charged with reviewing the general methodology
used by all lEMPs, including Santa Clara Valley. Within this broader review,
the subcommittee will critioue the methodological correctness of using dose-response
curves for non-cancer effects. This group, which includes health professionals
affiliated with universities, medical schools, EPRI, EPA, Environmental Defense
Fund, the Conservation Foundation and Decision Focus Inc., met for the first
time in April 1986. The second review group is from Rutgers University and
will first review the Santa Clara Valley IEMP. Included in the scope of the
Rutgers review is an evaluation of all dose-response information for certain
chemicals to which people are exposed in the Santa Clara Valley. The chemicals
include: 1,1,DCE; 1,1,1 TCA; Chloroform; Methylene Chloride; Lead; Vinyl
Chloride; Benzene; Arsenic; Cellosolve; Methyl Cellosolve; TCE; Chloramines;
BaP; and Ethylene Oxide.
-------
B-4
APPLICATION OF THE IEMP METHODOLOGY FOR NON-CANCER HEALTH EFFECTS
The Stage I Report and the previous section of the appendix discuss the
general approach to the risk evaluation methodology. This appendix will try,
as directly as possible, to explain the basic application of the approach for
estimating the probable increased annual incidence of non-cancer health effects
for an exposed population.
Calculation of dose-response curves for non-cancer effects is slightly
different from the calculation of curves for cancer; however, the principle is
the same. In the case of non-cancer effects, it is assumed that the dose-
response curve has a threshold (see Figure 1). Thus, from zero exposure up to
some threshold level, the body is able to cope with the chemical exposure and
repair any resultant damage. At the threshold exposure level, the body can no
longer cope with the damage and increasing exposures above this threshold would
result in "observable" damage to an organ system. The term "observable" has
to be stressed since data are collected on effects that are seen in exposed
laboratory animals or in exposed humans. Obviously, the exposure level at
which the threshold is defined is highly dependent upon the ability to "see" an
adverse effect and upon the existence of test data examining relevant doses
and effects.
EPA Reference Doses
EPA has calculated RfDs (Reference Doses, previously called Acceptable
Daily Intake Levels, or ADI's) for a large set of chemicals. These RfD's
represent an estimate of the threshold below which we expect no adverse health
effect. The RfD values are based on animal and human dose-response data. By
using the RfD as a benchmark exposure level, one can estimate whether or not a
specific environmental exposure level should be a cause for concern. For example,
as the environmental exposure level approaches and/or exceeds the RfD, our
level of concern about potential adverse health effects also increases.
In calculating the dose-response curves for non-cancer effects, we use the
EPA RfD values as surrogates for the threshold, since there is no way of knowing
what the "true" threshold would be for a population. Since the RfD is
another critical part of the IEMD model, an explanation of how the RfD is
calculated is in order.
When the RfD is calculated, all the animal and human data available are
reviewed and the dose levels at which different effects are seen are noted.
The incidence of the effect is not of concern at this point; we are interested
simply in whether or not an effect was observed. The next step is to identify
the dose levels at which effects are observed, as indicated in Figure 2, and
then select the NOEL (no observed effect level) or NOAEL (no observed
adverse effect level), which represents the highest dose tested that did not
produce observable results. It is also desirable to define the LOEL (lowest
observed effect level) or LOAEL (lowest observed adverse effect level),
which is the lowest dose tested at which some type of adverse effect was seen.
-------
B-5
Chemical X
Incidence
RfD
Exposure
Estimated or
Monitored
Exposure
FIGURE B-l: Dose-Response Curve for Non-Cancer Effects
-------
Incidence
(% of
population
affected)
B-6
RfD
NOEL LOEL
Dose (milligrams per day)
Other
Other FEL
Figure B-2: Examples of Thresholds for a Single Substance
RfD (Reference Dose): EPA's no-effect threshold = NOEL divided by safety factors.
NOEL (No Observed Effect Level): highest dose at which no effect is seen.
LOEL (Lowest Observed Effect Level): lowest dose at which effect is seen.
Other other adverse health effect seen at this dose (e.g., liver or reproductive).
FEL (Frank Effect Level): more serious health effect seen at this dose.
-------
B-7
Bear in mind that there could be many NOELS and LOELs for each chemical.
It all depends on the doses that happened to be selected by the researchers.
For example, assume that the "true" NOEL =10 mgAg/day, but that no one has
tested 10 mg/kg/day. Instead, data exist for 0-5 mgAg/day and 30-80 mgAg/day.
In this case, assuming no effects were seen at 5 mgAg/day and that some adverse
effects were seen at 30, one would call the 5 mgAg/day does the NOEL and the
30 mgAg/day dose the LOEL. Also, research is very expensive and no experiment
can be exhaustive, so effects may be missed. For example, assume that at 15
mgAg/day the critical effect is subtle liver damage that can only be diagnosed
with enzyme tests and complex histological work. If the researcher did not
look for liver effects or simply used organ weight as a measure of liver effects,
one would have called 15 mgAg/day a NOEL.
To calculate the RfD, we select the lowest reliable NOEL among all health
effects divided by appropriate safety factors (SF). The SFs selected are based
on the nature of the study from which the NOEL was derived. Most SFs are
multiples of 10, with each one representing an extra degree of uncertainty for
1) use of animal data to predict human effects, 2) use of subchronic data to
predict chronic effects, 3) differing sensitivities among human populations,
etc. For example, assume the NOEL from a chronic animal study was 5 mgAg/day.
Then the RfD would be 5/100, where the safety factor of 100 represents 10 for
extrapolation from animal data to human effects times 10 to account for differing
sensitivities among humans. Different effects will be associated with different
thresholds (Figure 2). The RfD is designed to protect against all adverse
effects and as such, is set below all the effect-specific thresholds.
IEMD Effect-Specific Thresholds
It is critical to the screening phase of the IEMD methodology to set
priorities. One way to do this is to develop the dose-response curves for
exposures above the RfDs. To develop such dose-response curves we must first
establish thresholds for all the potential adverse health effects as reported
in the literature.
These effect-specific thresholds are calculated in exactly the same manner
as RfDs, in that we define an effect-specific NOEL and divide by appropriate
safety factors. (In cases where we list a threshold that is greater than the
RfD, it is specific to the noted health effect category.) Where we simply do
not have adequate data, we use the Agency RfD as a surrogate for the threshold.
If there were complete data on the adverse health effects from a single
chemical, one could plot dose-response curves for all the known effects and the
plot might look like Figure 3.
Estimating non-Cancer Risks Quantitatively
Calculating the individual lifetime risk for non-cancer health effects is
different than for cancer effects because of the threshold. The crucial
difference is that risk is estimated only from the level of exposure over the RfD.
Therefore we subtract the RfD from the estimated exposure level (E) over the
RfD to estimate the exposure. This time the potency (the risk per unit of dose, or
the slope of the dose response curve) provides the individual lifetime risk.
-------
B-8
100%
Incidence
0%
Data for Chemical X
RfD
lung
effects
RfD
liver
effects
RfD
kidney
effects
RfD
reproductive
effects
Lung RfD based on a NOEL with higher doses causing lung effects.
FIGURE B-3
-------
B-9
R= (Ec - RfD) X Potency
R= The increased probability than an exposed individual will experience
a particular health effect in his lifetime.
Ec= The estimated exposure level. For air exposure,this is expressed in
micrograms per cubic meter (mg/m); in water, it is expressed in micrograms per
liter (mg/1).
RfD = The reference dose, which is an EPA number that indicates the level
below which we expect no specified adverse health effect from
exposure to that particular chemical.
Potency = The slope of the dose-response curve for the specific health effect
and chemical of concern. Much of the available data used in
estimating the dose response curve comes from animal data.
After the individual risk is calculated, the annual increase in incidence
for a specific effect for exposures above the Rfd would be calculated as follows:
Annual Incidence = R X Exposed Population
70
Annual Incidence = The projected increased annual incidence of the
specific health effect.
Exposed Population = The number of people exposed to the chemical
at the estimated concentration.
70 = We divide the results by 70 years (which is
the average individual lifetime used to calculate
individual lifetime risk) in order to estimate
annual incidence.
The purpose of the IEMD methodology is to calculate health effects incidence,
then use the results for setting priorities among different types of pollution
problems. We believe this methodology is reasonable for use in a screening
exercise for comparing risks; there is however, no scientific consensus for
its use in determining absolute risk levels.
-------
B-10
NOTE: This is a very brief summary of the Office of
Research and Development's concerns about the
IEMD methodology.
Office of Research and Development Concerns with IEMD Methods
The methods described by the IEMD represent a good first attempt at the
synthesis of health effects data at doses in excess of threshold. With many
chemicals, however, the toxicology is not sufficiently understood to adequately
extrapolate the quantal responses in animal experiments to humans. Nor are the
variabilities of human responses often sufficiently well understood to confidently
use an adjusted animal dose-response curve even if the general magnitude of the
extrapolation is correct. We reccrrmend that the IEMD staff consider several of
the technical issues raised by the ORD in order to be consistent with previous
and evolving EPA methods. (See for example, U.S. EPA, 1984 a and b; Stara et
al., 1984 a and b; Dourson et al., 1985.)
I Extrapolation of animal response to humans.
One end point of toxicology is the determination of an animal model in
order to study and predict the most likely toxic effect in humans both
qualitatively and quantitatively. Thus, toxicity is studied in many animal species
to address several questions:
1) Is the critical toxic effect similar across species?
2) Is the magnitude of response similar across species?
3) Can differences in these areas (if any) be explained on the basis of
pharmacokinetic differences among species?
[For example, even within the same strain of rats a two-fold
difference in the toxicity to the organophosphate insecticide,
parathion, is observed between males and females (males are less
susceptible) due to the greater activity of the hepatic enzymes
that detoxify this chemical (Casarett and Doull, 1980 p.67).]
4) Can differences in these areas (if any) be explained on the basis of
inherent attributes of the species? (For example, rats cannot be
used as a basis for gall bladder disease in humans because rats do not
have this organ.)
When sufficient animal toxicity data and comparative pharmacokinetic data
are available, dose-response curves can be drawn with some confidence. An
example of this in humans is flurosis and mottling of teeth with fluoride
exposure. More often than not, however, such data are lacking. In these
cases no effect levels are generally a better basis on which to estimate "safe"
human exposures than adverse effect levels (or animal dose-response curves),
-------
B-ll
since based on the work of Heywood (1981, 1982) the types of effects seen are
often different across species.
In general, probability functions are difficult to extrapolate frcm animal
to man unless it can be shoWn that consistent qualitative and quantitative
responses to toxic insult exist among animal species, and that the pharmacokinetics
are not different between experimental animals and man, or if different, that
they can be quantified.
II. Sources of variability in human response (assuming the human response is
qualitatively similar to animals).
When sufficient animal toxicity data and comparative pharmacokinetic data
are available such that dose-response curves can be drawn with confidence,
sources of variability in human response must then be considered. Source of
variability in human response are due to:
1) natural heterogeneity and unique biological and behavioral make-up
for each human being;
2) lifestyle factors such as diet, smoking habits or alcohol consumption;
3) existing diseases and conditions such as metabolic efficiencies.
Given the above factors, the human response is expected to be a broader
probability range than the animal data (i.e., the dose-response curve is likely
to be less steep).
-------
B-12
TABLE B-l POTENTIAL NON-CANCER HEALTH EFFECTS
Chemicals monitored or modeled above estimated thresholds
POLLUTANT PATHWAY
Benzene Air
1,1,1 Tricloro- Air
ethane
Water
Trichloro- Water
ethylene
1,1 Dichloro- Water
ethlyene
Methylene Water
Chloride
Vi nyl Water
Chloride
POTENTIAL
HEALTH EFFECTS
Blood effects
Fetal effects
Fetal effects6
Liver
Neurobehavorial
Fetal effects6
Liver
Neurobehavioral
Liver
Kidney
Liver
Fetal effects
Liver
Card iovascular
Kidney
POPULATION
EXPOSED !'
100,000
MEI
0 - 100
0 - <30
0 - <10
0 - <10
0 - 240
0 - 340
<50
<50
<:s
ESTIMATED
2 THRESHOLD
2.5 ug/m3
41 ug/m3
979 ug/1
12,500 ug/1
260.0 ug/1
260.0 ug/1
310.0 ugl v
25 ug/1
2100.0 ug/1
2100.0 ug/1
45.5 ug/1
246.0 ug/1
246.0 ug/1
ANIMAL AVERAGE
POTENCY ANNUAL
( SLOPE ) 3 CONCENTRATION4
6.23xlO~8 0.2 - 2.6
1.13xlO~6
4.67xlO~8 0.1 - 110
ND
ND 0.2 - 130
9.9xlO-7
1.04xlO~5 0.4 - 143
1.14xlO~6
6.52xlO~8 1 - 1,591
1.58xlO~7
6.78xlO~6
4.40xlO~8 0.2 -130
3.8xlO~8
MAXIMUM
CONCEN-
TRATION5
27.6
18 - 32,000
5 - 500
6 - 800
27 - 3,200
2 - 17200
Nitrates
Water Metheglobinemia 50
(blue baby syndrome)
- 100 45,000 ug/1
5.1xlO~6
Not Available
-------
B-13
TABLE B-l (continued)
POLLUTANT
PATHWAY
POTENTIAL
HEALTH EFFECTS
POPULATION
EXPOSED I'
ESTIMATED
THRESHOLD
ANIMAL
POTENCY
(SLOPE)3
AVERAGE
ANNUAL
CONCENTRATION4
MAXIMUM
CONCEN-
TRATION5
Cadmium
Fish
Kidney
Liver
MEI
S.OxlCT4
No Data
5
Not Available
PCB
Fish
Liver
Neurological
Skin
Kidney
Reproduction
MEI
3.1xlO~4
3.1xlO~4
3.1X10"4
3.1xlO"4
3.1xlO~4
5.6xlO-2
4.5
13
2.7
3.8
Not Available
(1) For the same eftects, the population at risk may not be the total exposed population, but some sub-group
(e.g., pregnant women). Thus, we caution against using these figures to attempt to project aggregate
incidence of the effects.
(2) MEI refers to estimated exposures for the most exposed individuals. For air exposures, we would expect very
few people to be exposed at these levels. For fish, the exposure estimate is for a hypothetical
individual consuming 65 grams a day (one pound per week) of contaminated fish caught in the South Bay.
It would be inappropriate to assume widespread exposure at these levels.
(3) The appropriate units for the slope for air are (ug/m3)~l; for water (ug/l)~l; for fish (mg/kg/day)-!.
(4) We caution against using these estimated average concentrations to calculate individual risks or incidence for
these non-cancer effects. The relevant dose above the estimated threshold is not necessarily the same as
the average concentration above the threshold.
-------
B-14
TABLE B-l (continued)
(5) For drinking water, these estimates reflect the highest level of estimated exposure at any one time. For fish,
these estimates are for a person consuming 65 grams a day (one pound per week) of contaminated fish caught
inthe South Bay.
(6) TCA has not demonstrated any teratogenic potential in published studies conducted using rodent species.
Therefore, the IEMP base-case analysis assumes that exposure to TCA poses no risk of fetal effects. An
unpublished study, which has not undergone scientific peer review, reports fetotoxic effects (cardiac
malformations) in rat pups exposed in utero to TCA (Dapson et al., 1984). In order to assess the
importance to Santa Clara Valley residents of further research on this issue, the IEMP uses the Dapson
study to examine the possible impact of TCA under the alternative assumption that exposures above an
estimated threshold based on that study's results could pose the risk of fetal effects. THE SENSITI-
VITY RESULTS SHOULD NOT BE INTERPRETED ASP INDICATING WHETHER OR NOT A RISK IN FACT EXISTS; EPA
RECOMMENDS AGAINST USING THIS INFORMATION FOR RISK MANAGEMENT DECISION-MAKING OR REGULATORY ACTION. Under
this alternative assumption, the IEMP projects that about 3,000 people, mostly those using private wells,
could be exposed to levels of TCA in their drinking water that exceed the estimated threshold. In
addition, most-exposed individuals downwind of an industrial facility are projected to be exposed at levels
above the estimated threshold in the air. These findings suggest that more research is appropriate, both
on actual levels of exposure to the MEI and on TCA's potential adverse effects. The National Toxicology
Program has commissioned a project to repeat the limited Dapson study; results are expected in Fall, 1986.
-------
B-15
Calculated IEMD Dose-Ftesponse Curves
For
AIR and WATER Exposures
Based on a review of animal dose-response data
NOTE;
Exposure data must be converted to mg/kg/day
for Air 2.86E-4 x X ug/m3 = X mg/kg/day
for Vfeter 2.86E-5 x X ug/L = X mg/kg/day
-------
B-16
AIR AIR AIR AIR AIR AIR AIR AIR AIR AIR AIR AIR
-------
ARSENIC
w
§
o.
DOSE in mg/kg/dojr
BENZENE
LJ
Ct
FETAL
O.2
O.4 O.6 O.S
COf3E in mg/kei/c*ay
-------
CHLOROFORM
n
cc
LU
cc
O.S
0.7 -
O
iNHALATlQN EX5SE in mg/kg/cia;/
DICHLOROBROMOM ETHANE
UJ
Li
O
a
»
UJ
O.O4 O.O6
INHAJJXT1QN COSE in mg/kg/day
O.1
-------
METHYLENE CHLORIDE
0.2 0.4 0.6 O.S
INH-AJJOION DOSE in mg/Kg/ctay
METHYLENE CHLORIDE
Ul
Ul
fclODD
INHALATION DQ5E in mg/hg/doy
-------
TRICHLORO ETHANE
UJ
M
O
a
bi
LU
ce
0.1S -
O.14 -
O.1.3 -i
O.12 -
0.11 -
0.1 -
O.03 -
0.03 -
O.O7 -
o.oe -
O.OS -
O.G4. -
O.O3 -
O.Q2 -
O.O1 -
O
' r 1 1 1 r r 1 T
O.2 O.4. O.6 O.S 1
DDSE in mg/k^/do-y
O.4S
O.4 -
O.2 -
LU
»
S O.
O
a
O.2 -
O.1S -
O.1 -
O.C6 -
TRICHLORO ETHANE
O.OO4 O.OOS O.O12 O.O1C O.O2
INHALATION DOSE in mg/kg/ctay
O.O24
-------
VINYL CHLORIDE
O.O4. O.OS O.12 O.ie O.2:
ir-4HALAT]Qh4 IXlSE in mg/kg/dcay
O.24
LU
»
6
a
Crt
LJ
ct
VINYL CHLORIDE
BLOOD
10
INHALATION DOSE in mg/kg/d=y
-------
WATER WATER WATER WATER WATER WATER WATER WATER WATER WATER
-------
ARSENIC
I
o
%
LU
0.002 O.QO4. o.ooe
OPJJ- DOSE in mg/kg/doy
O.OOS
0.01
ARSENIC
1.2
LU
a
H
LU
ct
1.1 -
1 -
O.9 -
o.s -
O.7 -
o.e -
o.s -
O.4 -
O.3 -
0.2-
O.1 -
O.043
1 1'
O.O47 O.O43
ORAL DOSE in ing/Kg/ctay
-------
O.OC04S
O.OOQ4. -
o.ocojs -
O.COO3 -
LJ
n
i'.OCOSS -
o
O.OOO2 -
O.OC01S -
O.COO1 -
O.OCOC6 -
O
ARSENIC
,
O.OO2
-^1
O.OO4
T 1 T
o.ooe
o.cos
ORAL CCSE in mg/kg/doy
O.OO5
BENZENE
PETW.
BLOOD
O O.COCCO.OQO4O.OOO6O.COaSO.OO1 O.OO12O.CO14O.OO16O.OO1S O.OO2
ORAL. COSE in mg/kg/day
-------
CHLOROFORM
LU
UJ
O.OO4 O.COS O.O12
O.O2 O.O24 O.O2S
DQ3E in mg/kjg/doy
DICHLOROBROMOM ETHANE
O.4
O.O2 O.04
ORAL DOSE in mg/kg/doy
-------
METHYLENE CHLORIDE
o.ce
O.O7 -
O.OB -
o.ce -
n. O.OA -
tl
LU
tt
^ 0.03 H
O.Q2 -
0.01 -i
O
uvee.
FETAL.
O.O2 O.O4 O.O6
OFWJ_ COSE In mg/kg/'dqy
O.OS
0.1
METHYLENE CHLORIDE
LU
tn
a
n
LU
ORAL DOSE in mg/kg/ctay
-------
O.O7
o.oe -
o.os -
g 0.04 -
ce O.03 -
O.O2 -
O.O1 -
METHYLENE CHLORIDE
ORAL COSE !n mg/kg/etay
O.COO3
O.OOO23 -
O.OCO2S -
O.OCO24 -
O.OOO22
O.OOO2 -
'.OO014. -
O.CC01 -
o.ooocs -
O.OCOG4 -
0.00302 -
o
TRICHLOROETHANE
O.OOO2 O.COO4 O.COCC O.ODCS O.OD1 O.OD12 O.OO14 O.CO16 O.OD1S
ORAL DOSE in mg/Kg/ciay
-------
TRICHLORO ETHANE
O.O24 -
O.O22 -
O.O2 -
O.O1S -
^ o.oie -
tn
§ 0.014 -J
a.
B 0.012 -|
cc
^ O.O1 -
o.ooa -
o.oos -
O.OO4 -
O.OCG -
O.OOCC: O.OOCul O.OOOB O.OOOS O.CO1 O.OO12 O.OO14 O.OO16 O.OO1
OFlul- COSE in mg/kg/doy
O.OO4
TRICHLOROETHANE
O.O2S O.Q32 O.Q36
O.O2
O.O4
OFWJ_ COSE in mg/kg/doy
O.O44 O.CU13
-------
VINYLIDENE CHLORIDE
LU
Ul
III
Ul
0.22
0.2 -
O.1S -
o.ie -
O.14 -
0.12 -
O.1 -
o.os -
O.OB -
0.04 -
O.O2 -
o
uvat
0.002 o.ood. o.ooe o.oos o.oi 0.012 0.014
CX3SE in mg/kg/etajr
Ul
LI
§
a.
VINYLIDENE CHLORIDE
o.co2 O.OO4 o.ooe
OPWL DOSE in
o.ooa
O.O1
-------
VINYL CHLORIDE
UJ
0.12
O.t1 -
0.1 -
o.co -
o.oa -
O.O7 -
o.oa -
o.oa -
O.04. -
o.oa -
O.02 -
O.O1 -
o
1
O.OOO2
1 -
O.OOO4.
COSE in mg/kg/etay
VINYL CHLORIDE
O.O3
O.O2S -
O.O26 -
O.O24 -
O.O22 -
0.02 -
o.o 1 a -
O.O16 -
O.O14. -
O.012 -
O.O1 -
o.ooa -
o.ooe -
O.OG4 -
o.occ -
o
till I T T I I 1 1
O.ODCut O.COGS O.O012 O.C016 O .CO2 O.OD2A
DOSE in mg/kg/cfcjy
-------
YL CHLORIDE
O.O1S
0.02 0.04 o.oe o.oa
ORAL COSE in mg/kg/doy
0.1
------- |