&EPA
            United States
            Environmental Protection
            Agency
            Office of the Administrator
            Science Advisory Board
            Washington DC 20460
SAB EG88-Q40A
September
Final Report
Appendix A:
Strategies for Sources,
Transport and Fate Research
                        '•W
            Report of the Subcommittee
            on Sources, Transport and Fate
            Research Strategies Committee

-------
                            NOTICE
     This report has be^n -written as a part of the activities
of the Science Advisory Board, a public advisory group providing
extramural scientific information and advice to the Administrator
and other officials of the Environmental Protection Agency.
The Board is structured to provide a balanced, expert assessment
of scientific matters related to problems facing the Agency.
This report has not been reviewed for approval by the Agency;
hence, the contents of this report do not necessarily
represent the views and policies of the Environmental Protection
Agency or of other Federal agencies.  Any mention of trade
names or commercial products do not constitute endorsement or
recommendation for use.

-------
            U. S. Environmental Protection Agency
                    Science Advisory Board
                Research Strategies Committee
           Sources,  Transport and Fate Subcommittee
       George Hidy
         Electric Power Research Institute, Environmental Division
         3412 Hillview Avenue
         Palo Alto, CA   94303
Members
   Dr. Anaers Andren
         University of Wisconsin, Water Chemistry Laboratory
         660 N. Park Street
         Madison, Wisconsin  53706

   Dr. Jack Calvert
         National Center for Atmospheric Research
         1850 Table Mesa Drive
         Boulder, Colorado  80303

   Dr. Yoram Cohen
         University of California at  Los Angeles,
            Chemical Engineering  Department
        'Boelter Hall
         Los Angeles, CA  90024

   Dr. Richard Conway
         Union Carbide Corporation, South  Charleston  Technical  Center
         3200 Kanawha Turnpike  (Bldg.  770)
         South Charleston, WV   25303

   Dr. Robert  Huggett
         Virginia  Institute of  Marine Science,
            School of Marine Sciences
         9  Raymond  Drive
         Seaford, Virginia  23696

   Dr. Donald  O'Connor
         Manhatton College,
            Environmental  Engineering and  Science Program,
         Riverdale,  New  York   10471

   Dr. Barbara Walton
         Oak Ridge  National  Laboratory,  Environmental Sciences Division
         Post  Office  Box X
         Oak Ridge,  TN   37831-6083

   Dr. Herbert Ward
         Rice  University,
             Department  of  Environmental Science and Engineering
         61QQ  South  Main
         Houston,  TX  77005

-------
Support Staff
   Dr.  Donald G.  Barnes*
   Acting Staff Director,  Science Advisory Board
   U.S. Environmental Protection Agency

   Ms.  Joanna Foellmer
   Secretary to the Staff  Director,  Science Advisory  Board
   U.S. Environmental Protection Agency
* Original draft prepared through the support of
    Dr.  Terry F. Yosie,  former Director,  Science Advisory Board

-------
         Strategies for Sources,  Transport and Fate Research

                       Appendix to FUTURE RISK

                           TABLE OF CONTENTS
                                                            Page
1.0   Executive Summary                                       1

2.0   Importance of Sources,  Transport and Fate Research      4

  2.1    The Role of Sources,  Transport and Fate Research      4

  2.2    Key Elements Needs in Sources, Transport,             5
            and Fate Research

3.0   Strategy for Sources, Transport and Fate Research       7

  3.1    FIRST STRATEGIC ELEMENT:
          Reduction of Uncertainty in Estimating
            Environmental Concentrations of Pollutants        8

    3.1.1  Modeling and Model Validation                      8

    3.1.2  Source Characterization                           10

      3.1.2.1  Chemical Characterization                     11

      3.1.2.2  Release Rates                                 11

      3.1.2.3  Episodic Releases                             11

      3.1.2.4  Source Characterization by Medium             12
       3.1.2.4.1  Air                                        12
       3.1.2.4.2  Surface Water                              12
       3.1.2.4.3  Ground Water                               13
       3.1.2.4.4  Soils and Sediments                        13

-------
    3.1.3   Transport  Processes                               14

      3.1.3.1   Surface  Water                                 15

      3.1.3.2   Ground Water                                  15

      3.1.2.3   Water-Underlying  Bed  Interactions             16

      3.1.2.4   Soils                                         16

    3.1.4   Fate Processes                                    17

  3.2    SECOND STRATEGIC ELEMENT:
          Early Detection  of  Environmental  Problems          17

   3.2.1   New  Stressor  Identification:
            The Need  for Early Warning                       17

   3.2.2  Early Warning Data  Sources                        18

    3.2.2.1 Chemical,  Biological  and Physical Monitoring   18

    3.2.2.2 Societal,  Economic  and  Technological Changes   21

    3.2.2.3 Literature Reviews  and  Expert  Workshops        22

  3.3   Implementation at EPA                                 22

4.0  Recommendations                                         23

  4.1   Recommendation I: Emphasis  on STF Models             23

  4.2   Recommendation II:  Leadership by
                              Risk Assessment Council        24

  4.3   Recommendation III: Establishment of an
                             Early Warning Group             24

-------
       STRATEGIES FOR SOURCES,  TRANSPORT AND FATE RESEARCH


1.0  Executive Summary

     Sources, transport and fate (STF)  research explores the
interconnections between sources of environmental pollutants,
their transport and transformation through the environment,  and
their ultimate fate.  These research findings allow measurement
or prediction of pollutant concentrations at points distant  from
the source.  These exposure data are coupled with toxicity
information to assess risk.  In other cases, STF research can  be
used to identify sources of environmental risks.  For example,
previously unsuspected pollution sources fortuitously have been
identified through field measurements (e.g., chlorinated  •
dibenzo-p-dioxins from pulp and paper mills), and mathematical
models have successfully related suspect source emissions to
particular environmental findings  (e.g., stratospheric ozone
depletion).  In addition to risk assessment purposes, STF
research is being looked to as a generator of "early warning"
information on potential, emerging, and/or escalating
environmental problems.

     In order to meet these growing demands, STF research
strategy in the 1990's should have two major elements, which are
central to this report:
      a. Strengthening EPA's capability for predicting
            environmental  form and concentration of pollutants,
            with a  known level of  uncertainty,  through
            measurements and modelling.
      b. Utilizing  STF knowledge to provide an  early warning
            vehicle  for anticipating issues that are likely to
            become  priority concerns for  EPA.

The  first  element of the strategy  calls  for expansion of  the
knowledge  base  on transport and  transformation  processes  in order
to develop and  validate models  needed  in  the  assessment  and
management of environmental risks.  The  second  element  is
designed to  raise Agency and public awareness  of environmental
problems at  a stage early  enough to permit  adoption  of  a
cost-effective  approach  to risk  reduction.

     Regarding  the  first  strategic element,  much of  the success
of STF  research depends  upon the development and validation of
mathematical  models, specifying their  degrees of uncertainty.
While  the  basic principles applicable  to many of these  models are
known,  site-specific conditions,  process data needs  and differing
scale  requirements  (e.g.,  local vs.  regional vs. global, or
short-term vs.  long-term)  limit the  current successful
application  of  these models and introduce uncertainties which

                                — 1 —

-------
need to be identified, quantified,  and narrowed.
         •
     Models are predicated on mass  conservation and,  necessarily,
require data on source characterization, media transport and
chemical conversion processes, and  deposition or media removal
processes, which are sometimes called "fate" terms.  A broad
range of known and potential sources in all media (e.g., air,
surface water,* ground water, and soils) should be characterized
through a core program examining the chemical characterization,
release rates, -and frequency of releases from the sources.  This
source information, coupled with fundamental knowledge of
transport and transformation processes, should feed into the
mathematical models which predict the behavior and ultimate fate
of the pollutant(s) in various media and generate estimates of
exposure.

     Regarding the second major element of the strategy -- the
need for early warning --, great benefits can be derived from
early identification of problems; i.e., reasoned risk reduction
actions can be implemented to correct a situation before it
requires a costly crisis response.   The ability to detect
problems before they would traditionally appear is related to
foresighted collection and judicious use of key environmental
data.  Chemical, biological and physical monitori-ng activities, a
major source of such data, need to be strengthened considerably.

     In addition, a shift in  strategic  thinking is needed in  the
basic approa-ch to environmental monitoring.  Currently, the
Agency focuses on a limited number of selected pollutants,
adopting a "feedback" strategy.  That is,,  if certain pollutants
are found in excess of some existing standard or limit, the
information is fed backhand regulatory  action is taken.   In the
future, the Agency should adopt a "feed forward" strategy that
involves monitoring a much broader range of compounds and other
environmental  stressors of potential interest, many of  which  do
not have regulatory standards.  The resulting data would  provide
an increasingly realistic and complete  estimate of the  total
toxic burden in the environment and a context within which  to •
determine more easily the extent to which  chemically transformed
products or new, unregulated  compounds  enter  the environment.   It
would also highlight  situations in which the  distribution of
chemicals change,  perhaps indicating  significant changes  in
environmental  conditions; e.g., global  air (climate) warming  from
the increased  presence of radiation- absorbing gases  in the
atmosphere.  Such  information would be  fed forward and  analyzed,
possibly  leading.to the development of  a regulatory  or  other  risk
reduction response.

     Careful  consideration  of societal, economic and
technological  trends  could  also be  helpful in anticipating—and
possibly  avoiding—environmental problems.  Measurements alone
will not  suffice  to provide an anticipating framework.   A forward
looking  analysis also will  be needed.   The Agency  should assign
the  task  of  achieving this  "early  warning" goal  to a group

                                -2-

-------
charged with discerning the implications of emerging observations
and knowledge in the context of past knowledge.  This group would
submit an annual report on their findings and projections, or a
special alert of findings as necessary.

     This Report contains three specific recommendations:

      a. Recommendation I: Improved STF Models
         EPA should strengthen its research on STF model
            development, evaluation and validation, as a means  ,
            for reducing  uncertainty in risk assessment and risk
            management.

      b. Recommendation II: Leadership by the Risk Assessment
            Council
         The Agency's Risk Assessment Council should take steps
            to insure that STF research is integrated into EPA's
            approach to exposure assessment analyses,

      c. Recommendation III: Establishment of Early Warning Group
         The Agency should establish a group of senior scientists
            and engineers to identify potential, emerging, and/or
            escalating environment health and ecological  problems
            using systematic,  long term measurements and  their
            interpretation.
                                -3-

-------
2.0  Importance of Source,  Transport and Fate Research

2.1  The Role of Sources,  Transport and Fate Research

     The study of the sources,  transport and fate (STF)  of
pollutants is an essential part of environmental research.  This
type of work has served three major* roles in environmental
assessment,  resulting in the abil-ity to estimate exposure
concentration levels and to relate excessive exposure levels to ^
sources needing emission reduction.  The roles are:
      a. Generation of fundamental knowledge about the physical
            and  chemical characteristics of emissions from
            pollution sources.
      b. Clarification of the nature of transport, conversion and
            media loss processes (e.g. deposition and absorption)
            that lead to exposure.
      c. Highlighting of chemical conversions leading to.
            pollutants  that differ from the direct emissions;
            e.g., the conversion of sulfur dioxide into sulfuric
            acid in the atmosphere.

     In the first instance, an expanding inventory of compounds
or radioactive substances has emerged that potentially affect the
environment.  In some research the interpretation of temporal and
spatial distributions of ambient measurements has led to the
identification of sources otherwise not considered (e.g.,
polychlorinated dibenzo-p-dioxins from pulp and paper mills).
Accurate source identification information is important for.the
development of cost-effective control strategies.

     Second, STF research has provided the principal basis  for
developing mathematical models to relate source emissions to
ambient conditions, yielding exposure estimates.  Such models
provide a critical element of exposure estimation in space  and
time to assess the impact of both existing and new sources.
Particularly notable accomplishments of modeling  that have  not
only provided regulatory tools but have advanced  the
understanding of source-receptor relationships include multiple
source air dispersion models and chemical fate models for ground
water contamination.

     Specially designed measurement programs have been required
to provide data to investigate environmental processes and  to
verify and test the reliability of  such models.   These
measurements are distinct  from monitoring and surveillance  for
existing regulatory requirements.

     A third role of STF research  involves  understanding
large-scale environmental  phenomena,  using  basic  knowledge  of
chemical processes, results  from  laboratory  experimentation,
development of- mathematical  models, and  critical  measurements  in
the  field.  Examples of major  contributions  in  this  category

                                — 4~

-------
include:
      a.  Tropospheric ozone prediction schemes employing
            meteorological factors and highly complicated
            chemistry.
      b.  An explanation of the role of chlorofluorocarbons  in
            modifying stratospheric ozone,
      c.  The ^discovery of organic chlorine  compounds in treated
            drinking water.
      d.  The exploration of complex bio-geochemical factors
            affecting the speciation of heavy metals such as    *
            mercury, chromium and selenium.

     An additional emerging role of transport and fate research
concerns the identification, analysis and interpretation of
certain long-term trends that can alert policy makers to
significant environmental issues in the future.  Those have
included inferences from long-term monitoring that emerged
through ecological and biological effects,  acid deposition and
long--range transport, the build-up of greenhouse gases to produce
climate change, surface hydrogeological concerns in the storage
and land disposal of wastes, and identification of global
contamination of the oceans from certain pesticides.

     The past successes of STF research in raising scientific and
public awareness on a wide range of environmental  issues
indicates that the public  investment  in this work  is well
founded.  The area should  continue to be an important component
of EPA efforts in  research and development.


2.2  Key Elements Needed  in Sources,  Transport and Fate Research

     An  important factor  in risk  assessment  is knowledge of  the
sources  of pollutants and  the processes that  subsequently
govern the dispersal  of  a  pollutant  into the  environment.   This
relationship can be  conceived as  links  in  a  chain  beginning  with
pollutant emissions  followed  by  transport,  transformation  and
media  removal  (deposition  or  sorption).  Understanding  these
relationships  contributes  scientific insight as  to how
pollutants,  through space  and time,  reach  humans and  other
receptors.   In addition,  definition.of  source-receptor
relationships  aids  decisionmakers in targeting specific sources
for  risk reduction  efforts.   Failure to address  these questions
can  result  in environmental policies and regulations  that  are
cost-inefficient,  spending too little on some problems and too
much on  others.

      EPA's  research has contained a substantial  component of STF
work in  the past.   However, the effort frequently has been poorly
defined  and, consequently, has not always  focused on issues
central  to the Agency's needs.   For the 1990's,  STF research
 strategy should have two major elements:
       a. Strengthening the capability of predicting environmental
             form and concentration of pollutants, with a known

                                      -5-

-------
            level of uncertainty-,  through measurements and
            modeling.
      b. Utilizing STF knowledge to provide an early warning
            vehicle for anticipating issues that are likely to
            become priority concerns for EPA.

     The first strategic element would expand the base of
knowledge on transport and transformation processes in order to
develop and validate mathematical models for assessing* and
managing environmental risks through exposure estimation and
identification of significant sources and their relative
contributions.  The second goal would use a combination of
measurements, theory and analysis to contribute to raising
Agency and public awareness of issues potentially harmful to
public health and the environment at a stage early enough to
permit adoption of a cost-effective approach to risk reduction.

     Section 3 discusses these key elements of the STF research
strategy.  Section 4 offers specific recommendations.
                                -6-

-------
3.0  Strategy for Sources,  Transport and Fate  Research

     Environmental risk assessment and management requires
reliable means for estimating exposure in space and time, as  well
as estimating the contribution of sources to those exposure
patterns.  Exposure to humans and to ecosystems can occur through
respiration, ingestion, direct contact, or the food chain.
Exposure generally is defined in terms of an ambient
concentration or bioaccumulation concentration over time, a
deposition rate to a collector, or a total medium burden (amount
of material in a defined volume).  In the absence of direct
measurements of exposure, estimation of concentration,  deposition
or burden can be carried out through interpolation or
extrapolation of results of mathematical calculations based  on
the principles of mass and energy conservation.  Continuing
research is needed on both direct measurement of exposure and
predictive models, as discussed in the report of the Exposure
Assessment Subcommittee.  In general, projections of source
contributions to exposure at a receptor  (source-receptor
relationships, or SRRs) presently can be done only through
mathematical models.

     Inherent in either the interpretation of field observations
or mathematical modeling are uncertainties in exposure or SRRs
that are seldom known.  A scientifically supportable risk
assessment requires an accurate and precise exposure estimate.
Therefore, determination of uncertainty  in estimates.is as
important as the estimate itself.  This  is also  true in risk
management where a balancing among cost, technological
effectiveness and reliability, and other factors  is often
required or  used  for  selecting among determining  emission control
options.  Inherent in the improvement  of risk methodology is
reduction of uncertainty in estimates  of the attribution of
exposure to  specific  sources.

    Uncertainty  in exposure estimation derives  from  two factors.
The first concerns a  mismatch  in  the  spatial  or  temporal scale  of
calculations relative to receptors.   The second  uncertainty  stems
from errors  in  the models themselves,  the  input  data to the
models,  and  computational errors  inherent  in  numerical  techniques
which  are employed  in the execution of the model.

     Models  calculate concentrations  or burdens  in a relatively
coarse or macro-scale.  An  individual receptor generally is  much
smaller than this resolution and often is mobile.  These  factors
make  it necessary to use population mobility  and statistical
factors to  relate measurements at fixed stations and model
calculations to receptor exposure.   Little research has been done
to reduce  uncertainty in these factors or in estimation of  target
tissue doses resulting from environmental exposure.  This  subject
 is more fully discussed in the report of the Exposure Assessment
Subcommittee.

                                -7-

-------
     Work has been done to define uncertainties in model
calculations but, in general,  models and interpolation schemes
are not well tested for reliability or validated for the quality
of their simulations.

     The validation of theoretically based models is often
defined at two performance levels.  The first is determined by a
suitable comparison of its predictions with ambient concentration
measurements for given physical or chemical input conditions.
However, close correspondence between model prediction and a few'",
selected environmental measurements does not necessarily
constitute adequate model validation.  Thus, a second validation
criterion is needed that tests the model for its integrity in
simulating media processes that link source emissions to ambient
conditions.  Testing models at this level is far more demanding
than the first evaluation, but it needs to be an integral part of
STF research.  Experimentation with models at this level leads to
advances in basic knowledge as well as added confidence in the
model performance.


3.1    FIRST STRATEGIC ELEMENT:
       Reduction of Uncertainty in Estimating Environmental
          Concentrations  of Pollutants.

3.1.1  Modeling  and Model Validation

       Effective risk assessment  begins with an attempt to
estimate the environmental concentration of contaminants wiChin
an acceptable margin of error, followed by exposure assessment
that assumes a reasonable level of confidence in estimates of
contact with these environmental  pollutant concentrations.   Three
factors significantly affect the  accuracy of estimates  of
environmental concentration:
      a.  Specification of source location, chemical
             characteristics and emission rates.
      b.  Description of  transport and  chemical conversion
             processes.
      c.  Description of  fate  or  removal processes.
Information  on these  factors is used  to construct  models  to
represent the phenomena believed  to be  involved  in the  movement
and transformation of  chemicals  in  the  environment, and computer
codes are developed  to  facilitate what  are  frequently very
complex  calculations.

     The EPA requires  reliable data on  concentrations in
environmental media  in  order to  determine  exposures to  target
organisms and populations for  risk assessment.   The Agency also
requires data on the contribution of  specific  sources to
pollutant concentrations  (i.e.,  SRRs) as  a  means of identifying
which  sources to target for  reducing  risk.   In the absence of
sufficient  data  on exposures,  source  data  may  be combined with
transport  and transformation models to provide estimates  of

                                -8-

-------
exposure.  In such cases,  it is critical that the models be
capable of providing exposure estimates within acceptable (or at
least defined) bounds of uncertainty.   Uncertainty is determined
by the suitability of the model (i.e.,  whether it includes
appropriate terms for all of the important variables, such as
dispersion and meteorological conditions) and by the accuracy of
coefficients and other input data associated with each variable
or connecting mathematical term.  In an ideal world, each model
would be fully validated before it is used for risk assessment or
risk management decisions.  However, the high costs and long lead
times required for model validation inhibit the validation effort
in a regulatory agency such as EPA.

     In reality, a model can prove useful, even without a full
validation, provided that it can be empirically verified for a
range of conditions comparable to those in the situation for
which it is to be applied.  Furthermore, as it is applied more
widely and tested periodically against available environmental
data, it can be refined and/or modified, based on operating
experience.  Thus, iterative applications of models have served
the EPA in several ways.  They help to define the bounds of
uncertainty associated with use of the model in risk assessment
or risk management, thereby increasing confidence in the results.
At the same time, the data generated provide a basis for refining
the model and/or extending the  bounds over which it  can be used.

     Although STF models  primarily have been structured, to
address a single medium of the  environment--air, surface water,
ground water or soil--, it is  increasingly apparent  that
intermedia and multi-media models are also necessary to analyze
certain problems.  In any case, the most general and appropriate
structure of a model is based  on  the conservation of mass  or
continuity principles,  incorporating source, transport, transfer
and  transformation components.

     The  basic  principles that underlie  any  modeling effort  are
at  least  qualitatively  understood and  the  numerical coefficients
relating  to  the above mentioned components often can be
estimated.   The application  to a  particular  problem, however,
often  requires  more  detailed  qualitative descriptions  of
transport and  transformation  processes;  e.g.,  resuspension of
aerosols from soils  and the  role of wet scavenging  of  reactive
organics.   In addition,  more accurate  quantitative  measures of
the coefficients  which  model  these processes may be required in
order  to obtain projected estimates that are of practical use to
the risk manager.   Thus,  there arises  the need for  model
calibration and validation specific to the problem and region,
the degree and extent of which should  be guided by the
significance of the question and the environmental and economic
consequences.   EPA should continue to  establish a systematic
procedure and a specific schedule to validate key environmental
models.   This effort should include documenting underlying
 assumptions, updated modeling procedures and protocols, and
 estimating uncertainties in prediction capability for a range of

                                -9-

-------
conditions.

     The verification of models to a defined uncertainty requires
a combination of special .data acquisition,  including source
emission and field tests, laboratory experiments,  and theoretical
or mechanistic studies of media processes.   These  generally
involve progressive and incremental design  considerations based
on a continuing improvement in our knowledge.

     The model components to be quantified  are source
characterization, media transport and conversion processes, and *
ultimate disposition ("fate") processes.  Each of  these
components and their associated uncertainties are  discussed in
the following sections.


3.1.2.  Source Characterization

        In deriving estimates of environmental concentrations of
pollutants,  quantification of sources, their strengths, and
interactions is potentially one of the larger sources of
uncertainty.  Because of legislative mandates, source inventory
and characterization have been directed toward release into
specific environmental media such as air, surface water, ground
water and soil.  Great strides have been made on emissions
estimation in the last 15 years; however, source characterization
should continue to be high in priority because
      a. Historical sources such as abandoned waste pits and
            dumps have been inadequately characterized as to the
            presence of particular pollutants or to releases.
      b. More recently acknowledged sources, such as contaminated
            sediments, present additional assessment problems
      c. Advances in emission control technology and evolution of
            industrial processes and activities require
            progressive re-evaluation of emissions  inventories
      d. Multi-source and multi-media interactions  have been
            inadequately characterized.

     The sources  to be  studied will change  according to the
prioritization  of current and  projected environmental  problems
and introduction  of new  technology.  Within a given problem area,
the sources studied  should not be  limited to those  addressed by
current Federal  regulations.   Rather, the outlook should be as
comprehensive as  possible to  define the magnitude of current and
emerging problems.   In  studying  ground  water pollution sources,
for example, municipal  landfills should be  included as well as
RCRA Subtitle D facilities and the use  of agricultural chemicals.
Source research  should  address area,  as well as point,  sources
and  both mobile and  stationary  sources.  A continuing core
research program in  this area  is recommended both to  develop
generic methodology  and  to apply it  to  critical environmental
problems.   Three aspects of  a  core program  are  addressed  in  this
report: chemical characterization, release  rates, and  episodic
releases.   Emphasis  also should  be placed on emerging

                               -10-

-------
technologies and new chemicals entering the environment.


3.1.2.1   Chemical Characterization

     The objective of chemical characterization is to develop and
apply efficient methods that adequately define problem sources
and  point to solutions.  Accuracy,  precision, detection  limits,
matrix effects, cost, and time are all critical factors.   Besides
the identity and concentration of chemical constituents,  tests
are needed to predict the mobility of materials under various
scenarios and to provide data for selection/design of control
techniques.


3.1.2.2  Release Rates

     Emissions from point sources often can be directly measured,
while the flux of contaminants from various area or diffuse
sources into the environment is estimated by applying a
mathematical model to either source characterization data or
ambient monitoring data.  Each approach is associated with levels
of uncertainty that need to be established and then reduced when
greater accuracy and precision is required.  For example, many
exposure estimates assume a nominally steady discharge of a
pollutant, when in fact, variation in emissions rate may be
critical to an accurate estimate of exposure.  Also, there is a
need to develop approaches which utilize all available data
sources, be they NPDES  reports, air permits or RCRA Part B
applications.  Improved release rate models need  to be soundly
conceived a_nd  adequately verified for a variety of applications.


3.1.2.3  Episodic Releases

     In many situations EPA may emphasize  the  regulation of
emissions  under  stable,  steady conditions  while serious
environmental  and/or human  health problems are caused by  sudden
releases  or  "upset"  conditions.   Formal procedures are needed
for:
      a.  Identifying specific potential hazards;  i.e. situations
             that could result in  a  sudden  release.
      b.  Estimating  probability  of  that hazard occurring.
      c.  Predicting  the magnitude and chemical or physical  form
             of the  release.

      Projections of equipment failure and handling or
 transportation accident rates are needed.   Acceptable standards
 of practice need to be established  and tested.  One approach for
 hazard  identification is to divide  operations into segments and
 compare possible risks against a hazard checklist, including
 combustible mixtures, mechanical stress,  vapor cloud release and
 over-pressurization.  The potential for natural disasters,  like a
 sudden gas release  (e.g. carbon dioxide or hydrogen sulfide) from

                               -11-

-------
a volcanic disturbance or a deep lake sediment should be
investigated.


3.1.2.4    Source Characterization by Medium

3.1.2.4.1  Air

     Although air quality research is said to be more advanced
than research for other media, a continuing effort will be needed
to refine and improve knowledge of emissions for regulatory
decision making.  Over the past decade there has been
considerable effort to characterize emissions of criteria
pollutants and certain hazardous chemicals from stationary and
mobile sources.

     Data acquisition for source characterization will be needed
for air regulatory analysis at a modest level of priority for at
least the next decade.  Continuing work will be required to
maintain and update the inventories.  Characterization of
emissions from new or rebuilt facilities will be required, as
will the estimation of pollutant forms not previously considered.

     Additional research will continue to be needed to provide
improved emission factors and to define the uncertainties and
limitations  in available data.  With such refinements, high
priority should be assigned .to upgrading estimates of emissions
of nitrogen  oxides and volatile organic compound emissions for
use in source-receptor modeling and control strategy analysis of
oxidants and air toxics.


3.1.2.4.2  Surface Water

     In many instances non-point sources are  the major
contributors to  freshwater  surface  problems;  e.g., toxics  in Lake
Superior and Lake Michigan.   Risk reduction efforts will
increasingly turn to  non-point  sources because  of  the  large
fraction of  surface water  pollution problems  they  may  represent
and since point  sources  have  been more effectively controlled.

     Potentially  important non-point  surface  water pollution
sources  include  the following:  run-off and  leachates  from
agricultural and  other land uses, deposition  of wind-borne
volatile organic  chemicals and  heavy  metals,  groundwater  inflow
and sediment releases.   At this time  source models for predicting
organic  loadings  are  much  further developed  than  models  for
inorganic loadings.   Also, agricultural  run-off  is  considered  to
be better characterized  than  is urban run-off.  Research  should
be balanced  between monitoring  (direct measurements  for use  in
identifying/defining  problems,  as  input data  to models,  and  in
model  validation)  and development  of  predictive run-off models.
Specific models  need  couple the source information with
hydrodynamic and process kinetic models,  describing  sediment

                               -12-

-------
transport, and elucidating biologically mediated reactions, metal
speciation kinetics, and hydrophobic compound transport.
Reconstructive models based on concentrations in receptor
organisms, including humans, also are useful.  Balanced funding
'of field measurements and predictive modeling is recommended.


3.1.2.4.3  Ground Water

     Contamination of ground water from human activities
frequently originates from surface impoundments, landfills,
agriculture,  leaks and spills, septic tanks, mining, petroleum
and  gas production, and underground injection of wastes.  EPA's
1977 "Report  to  Congress on Waste Disposal Practices and Their
Effects on Ground Water" (Premier Press, Berkley, CA,  1980)
identified the disposal of wastes at industrial impoundments and
other  solid waste disposal sites as the most important sources of
groundwater contamination.  It estimated that approximately  15%
of the liquid and solid industrial wastes generated in the United
States can be classified as hazardous.  Such 'wastes represent
potential  sources of groundwater contamination, depending  on the
method of  disposal.  Most of  the past  land-disposed wastes were
not  managed by means that comply with  more  recent Federal
regulations,  and, therefore,  they may  threaten  groundwater
quality in many  areas.

      In addition to  industrial wastes,' the  1977 report identified
so-called  secondary  sources of national  importance  including
septic tanks, municipal wastewater, mining,  and petroleum
exploration and  production  residues.   Although  concentrations  of
toxic  material  from  these sources are  generally lower  than from
industrial wastes,  they can be significant  on a regional basis.
In an  area of substantial manufacturing activity  containing  large
numbers of people,  there  exists  a potential for pollution  of
groundwater  resburces,  especially  from products such  as gasoline,
fuel oils, and  solvents.  Areas  where  mining, agriculture,  and/or
petroleum production are  prevalent  are also at  potential risk.


 3.1.2.4.4  Soils and Sediments

      Soils and sediments  can  retain organic and inorganic
 chemicals released to the environment.  Therefore,  they can
 become sources for release  and subsequent contamination of air,
 ground water, and surface waters through resuspension, vapor
 losses, leaching, and removal of particulates containing sorbed
 compounds.   Defensible risk assessments and risk management
 strategies require reliable information on the amount of
 contaminants accumulated in soils at sites and knowledge of how
 to  predict contaminant persistence, transformation, and transport
 to  other media.

       The spatial distribution of chemical contaminants  in soils
 is  often extremely heterogeneous.  Consequently, extensive  core

                               -13-

-------
sampling and/or exhumation to delineate zones of contamination
can be time-consuming and expensive.   In situ and remote assay
equipment and sampling methods are needed to determine
concentrations of chemicals in surface and subsurface soils.
Among approaches that show promise are those that couple recent
advances in laser technology with those in fiber optics in order
to improve the detection  of organics and the development of
portable gas chromatographs for analysis of volatile organics in
the field.  In addition, neutron and  scintillation probes may
prove useful for in situ detection of transuranic and
gamma-emitting radionuclides, respectively.  Development of these
and other techniques to detect and quantify contaminants will
require a significant research effort, but. one that would yield a
high payoff in monitoring capability.

     In order to adequately assess soil sources, there is a need
for appropriate leaching test(s).  Improved methods are needed to
evaluate contaminated soils and wastes to account for variability
in leach rates of constituent chemicals over a long period of
time at specific  sites.  Such methods could be used-to improve
cleanup and closure of RCRA and CERCLA sites, as well as to serve
as guidance for management of land treatment and landfill
facilities.


3.1.3  Transport Processes

     There is general acceptance of the basic principles of the
transport component of  STF jnodels.  The fundamental  equations of
fluid  (air and water) motion  which follow  from these principles
are reasonably well-established and have been utilized  for
calculations.  The principal  limitations of  these models often
lie with certain empirical transfer coefficients that are media
specific.  The mixing and dispersion  associated with the  fluid
motion, although understood,  has less of a research  base  to
support it.   Nevertheless, empirical  relations, based on  field
and laboratory data,  provide  important insight  into  the nature  of
transfer coefficient  variability.  The recent advances  in
turbulence theory  increase  scientific understanding  of  this
important mixing mechanism.   At  the  present  time,  the development
and application of  fluid dynamic models  incorporating  this theory
reside primarily with the  research scientists.  Transferring  this
knowledge to  the environmental analyst is  an important  need  for
the future development  and  application of  air and  water quality
models.   Certain specialized  areas;  e.g.,  underlying sediment bed
failure,  remain as  more fundamental  challenges  for environmental
modellers.

      Analytical and numerical solutions and the associated codes
are available for  surface and ground water transport models.
Stochastic  or Monte Carlo techniques have also  been used to
define the  uncertainty  of the various elements  in these models.
Different  research groups should test a selected group of models
in order  to  determine those which are most appropriate for use in

                               -14-

-------
 environmental  quality  analysis.  This  testing  should be followed
 by  a  comparison  of  model  performance with  sets  of observations
 from  various  air regimes  and  water  systems.


 3.1.3.1   Surface Water

      The transport  of  pollutants in freshwater  bodies  is
 predominantly advective  (moving with the mean  flow), rather  than
 dispersive (associated with the eddying).   Given the knowledge  of
 the hydrologic balance of a drainage area  and  empirical
 correlation of the  dispersion, transport of pollution  in
 freshwater systems  can usually be determined with greater
 accuracy than some  other  components of the mass balance; i.e.
 sources and transformations.  The hydraulic interaction  (e.g.,
 caused by fluctuating  levels) between  the  surface and  ground
 water deserves further attention as it relates to estimating
 contamination.

      The dispersive component in estuarine and coastal systems  is
 more significant than  in  fresh waters.  In the former, the
 effects of density  stratification  on vertical  and  lateral
 dispersion and the  distribution  and disposition of  sediment and
 organic particulates needs to be further developed.  Given the
 intensive computational  manipulations  required to  solve
 multi-dimensional fluid  dynamic/quality models, a  significant
 effort is needed to enable these models to interact for
 long-term, time variable simulations and projections.   Much
 remains to be done on transitions  for coastal  systems both on the
 east-and west coast of the country.  The eddies from the Gulf
 Stream in the Atlantic have marked effects on the transport
 within the region north of Cape Hatteras and notably  in the New
 York Bight.  The Gulf Stream has not been successfully modeled in
 spite of  the advanced state of knowledge of the field and
 aforementioned  developments.

     Wind  and  temperature effects are important factors in
 defining  transport in marine systems,  as well as in lakes and
 reservoirs.  The latter have frequently been modelled in a
 simplistic fashion by assuming complete mixing.  For  long term
 projections  this approximation has been successfully  applied in
 many  cases.  For the  more refined, multi-dimensional  analysis,
—however,  much needs to be done in  applying  turbulence theory to
 the analysis  of transport phenomena in lakes and reservoirs so
 that  detailed concentration  patterns  can  be estimated. In
 addition,  intermedia  transfer phenomena (e.g.,the  role of the
 surface  microlayer in pollutant transport across the
 water/atmosphere interface and critical factors in water/sediment
 transfer) need  additional study.


 3.1.3.2   Ground Water

       As in the  case of  surface water,  the basic equations of

                               -15-

-------
fluid motion in the saturated zone of ground waters are generally
well established and understood.  Qualitatively, fluid transport
in some important ground water media has successfully been
modeled.  Similarly, the dispersive effects, referred to as
dispersivity,  have been empirically defined to s-"~e degree, but
site-specific evaluation of this component is u:   ly required.
By contrast, knowledge of the transport in the u  iturated zone
is inadequate, and further development in this area is needed.  A
major source of uncertainty in groundwater modelling is the
inherent heterogeneity of the soil media and underlying rock
structure, which must be addressed on a site-specific basis.


3.1.3.3  Water-Underlying Bed Interactions

     The exchange between the water and the underlying sediment
bed acts as both source and sink of dissolved and particulate
forms of pollutant constituents.  In some cases, the transfer
rates from or to the bed far exceed the current mass inputs from
point and non-point sources.  Two general constituent categor-ies
are considered: nutrients and toxic substances.  The former
affects the bacterial and algal levels in the water column.   In
many locations  this interaction is the major factor in a
dissolved oxygen budget.  There is a pressing need for the
analysis of dissolved oxygen and eutrophication to better
characterize  the flux of nutrients at the water-underlying  bed
interface.

    Many organic chemicals and  heavy metals partition to
particulate matter, particularly to the organic and clay fraction
of  the  solids.  The degree to which the contaminated particulates
accumulate  in the  sediment depends on the characteristics  of  the
solids  and  the  turbulence and shear at  the  water-underlying bed
interface.  Estuaries,  lakes  or reservoirs  are  net sedimenting
systems and accumulate  these  toxic substances  in  the  bed.   The
broad area  of sediment  interaction with the water column  in all
systems requires a significant  effort  in  order  to understand  the
phenomena of  sediment  transport,  settling,  resuspension and
bioturbation  affecting  water  quality.


3.1.3.4  Soils

     Current  methods  are inadequate  to predict accurately
conditions  in soils (e.g., moisture  and temperature  fluctuations)
or  the  transport  of organic  and inorganic contaminants in this
medium.  Studies  are  needed  to  refine the conceptual models for
organic and inorganic mobility  and to provide for the influence
of  soil heterogeneity and other environmental variables on these
processes.   Structure-activity  analyses should be explored to
improve predictions based on physicochemical properties of
specific  compounds.

      Pollutant transport through soils is often viewed simply as

                               -16-

-------
a single chemical chromatographic process.   This  view fails  to
account for the influence of soil structure (e.g.,  macropores),
rainfall events, contaminant interactions in waste  mixtures,  and
the possibility of movement via colloidal transport and
adsorption to mobile microparticulates.   Studies  are needed  to
delineate the extent to which these additional processes affect
rates of contaminant transport in soils.


3.1.4  Fate Processes                                           "

     The environmental fate of a substance depends  on physical
dispersion processes as well as on its physical,  chemical,  and
biological properties or interactions with substrates.
Information required for future predictions of the  fate of
chemicals in air, soil, or water includes such basic data as
aqueous solubility, vapor pressure, air-water partition
coefficient (Henry's Law constant), molecular diffusivity,  phase
partition coefficient, melting point and absorbtivity.  There has
been progress in acquiring data in pure homogeneous systems.
This fundamental information is needed in order to understand the
effects of cosolvents, unicelles, and colloids on these
properties.  In addition, many chemicals hydrolyze, photolyze or
participate in additional abiotic or biotic degradation
processes, such as electron transfer reactions.  An increased
effort is needed,to produce thermodynamic and kinetic data for
heterogeneous systems as well; e.g., the influence of metal
oxides and microorganisms on the persistence of chemicals in
soils.

     Since economic and  logistical constraints prohibit
laboratory measurements  for all  these properties and  rate
constants, an alternative predictive  tool  is recommended:
investigation  of  chemical structure-activity relationships.
These  estimated parameters  are then adapted for appropriate  fate
models.   Such  an  approach has  been used  with great  success  in
chemical  engineering  to  design unit processes  for  chemical
manufacturing  and  in  pharmacology  to  construct pharmacokinetic
drug transport  models.


 3.2    SECOND  STRATEGIC  ELEMENT:
        Early  Detection of  Environmental  Problems

 3.2.1   New Stressor Identification:  The  Need for Early Warning

     Early identification of potential,  emerging and/or
 escalating environmental problems should take its  place along
 with risk assessment and risk management as a central part of
 EPA's  mission.  The current research program provides no funds
 specifically earmarked toward this objective.  This is
 disappointing in view of the number of issues (e.g., radon,
 stratospheric ozone depletion and global climate change) that
 have only recently risen to priority in EPA's policy agenda but

                               -17-

-------
which have been known to the scientific community for a number of
years.  It is also surprising because of relatively high
perceived risks and the rising priority for these "newer"
problems which are discussed in EPA's February,  1987 report
entitled "Unfinished Business: A Comparative Assessment of
Environmental Problems" (EPA/230/2-87/025a-e).   While admittedly
not a scientific study, the "Unfinished Business" report provides
a rationale for follow-up investigations that need to be pursued,
if only to minimize future surprises and to ensure a better match
between research expenditures and significant sources of public*
health and environmental risk.

    The benefits of early identification of stressors to human
health and ecological systems include:
      a. Cost reduction: more orderly conduct of the research vsc
            expensive crash programs.
      b. Improved regulation: more time is available to develop
            data bases for scientifically supportable
            regulations.
      c. Risk reduction: steps can be taken early to reduce or
            prevent risk either by non-regulatory and/or
            regulatory means.
Initiation of a program to identify new or potential risks, which
can complement the Agency's ongoing efforts to assess known
risks, is strongly recommended.


3.2.2    Early Warning Data Sources

3.2.2.1  Chemical, Biological, and Physical Monitoring

      Many of the environmental stresses that concern this  nation
and the world are caused by anthropogenic chemicals.  Often a
crisis is first detected through the  direct observation  of a
biological effect caused by pollutants rather than  by earlier
prediction or detection of  the release.  There are  numerous
examples of  this pattern,  including:
      a. Kepone in  the James  River detected by the  observation  of
             worker  illness.
      b. Tributyltin  in harbors detected by  the  observation of
             malformed oysters.
      c. Polybrominated biphenyls in  Michigan cattle detected by
             the observation of dead  and dying animals.
      d. Polynuclear  aromatic hydrocarbons  in areas of  the Puget
             Sound  detected by the observation of fish  with
             cancers.

      Often by  the  time a  problem  is  detected, biological damage
has already  occurred  and  remediation is difficult,  expensive  or
impossible  from a  practical standpoint.   In other words, the
anticipatory regulatory  systems  have been  inadequate.

      Improved  anticipation can be promoted through an improved
surveillance system.   Such a system would  continue to include a

                               -18-

-------
chemical monitoring program designed to quantify a preselected
set of compounds already of regulatory interest.  This approach
has been the thrust of most environmental monitoring to date.  An
improved surveillance system should, in addition, provide
qualitative identification of additional chemicals of concern.
This latter approach has been haphazard in deployment, but has
proven important.

    There are advantages and disadvantages to the first, directed
approach.  One advantage is that the qualitative aspects of     f
chemical analyses are simplified.  Analytical methodologies can
be selected or developed for specific compounds, decreasing the
possibility of false identification.  The quantitative outputs of
the analyses are usually more accurate and precise because the
methodologies employed are optimized for the preselected
compounds.  These outputs are particularly important if the
objective of monitoring is to determine compliance with some
regulatory program or permit.

     A disadvantage of the directed approach is  that only  the
preselected compounds are surveyed  even though  other compounds
may also be detected.  The data  for the latter  compounds are
generally  ignored and even lost.  New  compounds,  which  may later
prove to be damaging to human health or the environment, are  not
systematically tracked.  Examples exist where chemical  problems
have been  needlessly overlooked.  Among these are the  impacts of
such organic  chemicals as polychlorinated biphenyls  in  the
1960's,  Kepone and dioxins in the .1970's.   In other  words,
potentially valuable chemical data  have not and are  not being •
utilized specifically for environmental assessment because of a
narrow  focus  on ichemical-specific monitoring.

     Another  way of  describing  most existing monitoring systems
for  toxic  chemicals  is  to  describe  them as  "feedback"  programs.
Such feedback programs  are  keyed by error  signals.  For example,
if a  permit allows a certain amount of a  specific compound in an
effluent,  a concentration  that  exceeds the  permitted level by an
established margin constitutes  an  exceedance;  i.e.,  a violation.
Detection  of  this violation  may feedback,  initiating regulatory
action.   Compounds not  specified in a permit and, therefore,  not
analytically sought, cannot trigger an warning alert even though
these  "new" compounds may be detrimental to the biological
communities in  the receiving media.

      Technologies and expertise now exist to reduce such
oversight through improved design of broad-based chemical
monitoring programs and the use of biological endpoints in
monitoring.  The use of techniques such as gas chromatography
 (GC)  and gas chromatography-mass spectrometry  (GC-MS) provide
 effective means for efficient,  anticipatory monitoring.   Through
 the use of various columns and detectors, these methods yield
 signals for essentially all of the compounds present,  both those
 which are analytical targets and those which are  unexpected.
 Even though many of the output signals are not  essential  to a

                               -19-

-------
feedback system, they can be collected,  stored,  and analyzed
through the use of data handling systems.   This  broad-based
record can be examined historically for  chemicals of possible
concern and for apparent shifts or trends  that may signal an
accumulation of material.

     In such a program, data systems could be linked together to
create networks.  Software could be developed to query the
networks to determine whether new chemicals have appeared between
samplings and whether a compound is increasing or decreasing over
time.  The network could provide efficient access to the areal
distribution of a compound(s) of potential interest.            "

     This alternative to targeted chemical monitoring would
sacrifice some quantitative aspects of the analyses in order to
maximize the qualitative outputs.  That  is, this added
information would be obtained at the cost  of somewhat higher
limits of detection.

     The results of these refinements to chemical monitoring
would be progress toward "feed forward", rather than feedback,
monitoring.  Feed forward monitoring, in this case, is defined as
monitoring designed to determine when new, unregulated or
unselected compounds enter a system and when shifts in the
distribution of chemicals in the environment occur.  Feed forward
monitoring has  the advantage of determining many more compounds,
which in turn provides a much more realistic estimate of the
total toxic burden to which organisms are exposed.  Such
information would be "fed forward" and analyzed, possibly leading
to the development of a  regulatory or other risk reduction
response.

     Although feed forward monitoring programs may  be less
cost-effective  in the  short-term than routine feedback monitoring
for  regulatory  compliance, the  long-term benefits of avoiding a
future kepone-type event justify the costs associated with  the
development and maintenance  of  such an early warning  system.
Further, judicious application  of  knowledge about what biological
and  physical  processes are possible can increase  the  efficiency
with which  feed forward  monitoring is conducted.

     Plants and animals  can  also be used  to gain  important
information about the  sources  and  availability  to  biota  of
chemical contaminants  and  their resulting  effects.   Insight can
result  from analyses  of  tissue that may not be  possible  from
chemical analyses of  abiotic components of  ecosystems.   Another
advantage  of  biological  monitoring is the  extremely high
sensitivity of  certain biochemical endpoints, such as enzyme
induction,  that can  supply  evidence of  the presence of  chemicals
at concentrations below  thresholds for  chemical analyses.
Recently published  studies  have shown that biomedical tests
derived  from  research on mammals are  useful  when applied to
aquatic systems.  The detection of chemical  stresses on aquatic
biota  by utilizing  histopathological  and  immunological techniques

                               -20-

-------
is new possible.  The observation of tumors in fish from Puget
Sound, the finding of lesions and depressed immune systems in
fish from the Elizabeth River, Virginia,  and the determination of
elevations in metallothionein concentrations in fish from
Prickley Pear Creek, Montana, are examples of the use of such
technologies.  Also, non-specific indicators of toxicant exposure
can be valuable monitoring tools in broad-scale screening
programs.  For example, deviations from normal ratios of
single-stranded to double-stranded DNA reflect exposure to a
broad array of genotoxic chemicals.

     There is no doubt that the ability to analyze environmental
samples will improve and become more comprehensive in the future.
There is also no doubt that the need for long-term monitoring
data will increase as technology and human populations expand.
Both of these developments support the concept of collecting and
storing environmental samples to be analyzed in the future as new
techniques .become available or other needs dictate.  The Agency
now participates in such a program, the Environmental Specimen
Bank.  Consideration should be given to expanding the effort.
The availability of documented samples on which to perform
retrospective analyses could be extremely advantageous for
determining temporal or spatial trends.  Similar efforts should
continue with the National Human Adipose Tissue Survey (NHATS).

     Monitoring efforts should also address risks caused by
stresses on humans and ecological systems other than direct
toxicity of anthropogenic chemicals.  These stresses include
global warming, increased UV-B radiation, physical modification
of habitat, radon, pathogenic and engineered organisms, and
natural chemical emissions.


3.2.2.2  Societal, Economic and Technological Changes

     Clues as to potential and emerging public health or
ecological stressors  (risk) can be gained by periodically
examining societal, economic  and technological trends.  For
example, energy conservation  scenarios developed  in  the 1970's
because of rising energy prices could have  predicted the  rising
importance of indoor  air pollution problems heightened by
increasing insulation and resulting decreased ventilation.
Similarly, more recent estimates that approximately  70%~of  the
American people will  live within 50 miles of  a coastal area  by
the  year 2000  strengthen the  urgency  for protecting  estuarine and
marine ecosystems.  Other examples of trends  which  can be studied
are  the  significance  of superconductors, climate  change and urban
population changes.

      Some aspects of  such an  effort were  included in a  1980 ORD
report entitled "Environmental  Outlook  1980"  (EPA-600/8-80-0003,
July 1980).   Potentially useful  procedures  of identifying the
environmental  impact  of trends  in  energy  supply/demand,
demographics,  human activities,  economics,  regulations,  natural

                               -21-

-------
cycles, international activities,  and technology are described.
However, the thrust of the report  was not on identification of
new risks or rapidly escalating risks; rather,  the report was
directed at determining the effects of such trends on existing
efforts to assess and control known risks.  In order to make this
risk-identification effort successful at a reasonable cost,
greater emphasis needs to be on the identification of new,
emerging, and rapidly escalating stressors/risks.


3.2.2.3  Literature Reviews and Expert Workshops
                                                               •7-
     Selected literature should be monitored with the aim of
searching for signals of new stressors.  Also,  workshops should
be held at least annually to solicit the thinking of outside
experts on potentially significant environmental problems.
Possible mechanisms include utilizing units of the National
Academy of Sciences, the National  Academy of Engineering, the
Office of Technology Assessment, professional societies or other
Federal agencies to host or co-sponsor such workshops.  Working
with EPA, these and other institutions; e.g., NIEHS, can organize
leading scientists, engineers, sociologists, economists, and
others to identify potential and emerging ecological and health
stresses.
3.3 •• Implementation by EPA
                                                                •
     Essential to the development and success of an early warning
system is the formation of a. group of people within EPA that
includes, at a minimum, staff drawn from the Office of Research
and Development-and the Office of Policy, Planning and
Evaluation.  The group would prepare analyses and studies of
potential problems, draw upon other Agency expertise, as
appropriate, and fund certain outside studies in the data source
areas cited above.  These people should be experienced
individuals who can discern the implications of existing and new
information and be able to assess its importance.  Inclusion of
visiting scientists from academia, industry or private groups
would assist this effort by adding external inputs to the Agency.

     Each year this group would prepare an annual report to the
Administrator, Deputy Administrator and Assistant Administrators
of new, emerging, and/or escalating health and environmental
problems.  The Assistant Administrator for ORD would develop a
mechanism to ensure that the conclusions, and recommendations of
this group receive formal consideration in the research planning
process.
                               -22-

-------
 4.0  Recommendations

 4.1  Recominendation_JI: Emphasis on STF Models

     EPA  should maintain its research on sources, transport and
_fate  {STF) model development, evaluation and validation, and
 continue  improving  its methods for reducing uncertainty in risk
 assessment.

     To implement this recommendation, EPA should take the
 following actions:
       a.  Continue to  formalize the mechanism and criteria for
             acceptability of STF models for all media, using
             methods such as the current procedures of the Office
             of Air  Quality Planning and Standards  (OAQPS).
       b.  Evaluate  and validate on a priority  schedule widely
             used STF  models  (single medium or  multi-media), using
             a combination of field measurements and  laboratory
             data to determine the level of uncertainties
             predicted by the models and to provide guidelines for
             reducing  these uncertainties.
       c.  Continue  research on media processes to ensure  the
             quality of model input data in order to  improve the
             detection and prediction of chemical transport and
             transformation in environmental media.
       d.  Adopt a systematic review schedule  for STF model
             progress, including  target milestones to achieve
             reduction in predicting uncertainties.
 These  four  actions  will  facilitate the preparation of an  orderly
 and  focused  Agencywide effort to  advance  the  development  and  use
 of STF models  in risk analysis.

      Currently'there  exists  a profusion of numerical codes  that
 are  exposure estimators.  However, in  general, they  are  not
 validated or tested,  nor have they been ascribed  specific
 quantitative uncertainties.  The  methods  adopted  by  OAQPS serve
 as a  useful  Agency  guide  for placing  a more  uniform  certification
 process on  these  types  of models  for  regulatory  analysis.
 Validated models are  essential  for this  use  so that  public
 confidence  in the  reliability of  risk  assessment' results can  be
 increased.   To achieve  the  goal  of  systematic and  continued
 improvement of  STF  models,  research  funds should  be  provided  to
 improve model input data for source  emissions, fluid flow
 estimation,  and physicochemical rate  parameters.   Improvements  in
 these components need to be  assimilated  progressively into models
 to ensure that the models  reflect the current state of knowledge.
 Comparisons between older  and  newer  models should also be
 attempted a regular basis  to evaluate progress in reducing
 uncertainty. These comparisons  should also be incorporated into a
 systematic review process to update the STF models recommended
 for regulatory applications.


                               -23-

-------
4.2  Recommendation II :  Leadership by Risk Assessment Council

     EPA's Risk Assessment Council (Council) should ensure that
STF research is integrated into EPA's approach to exposure
assessment.  Specifically, the Council should:
      a. Initiate the development, of Agencywide guidelines for
            STF model performance criteria and their
            acceptability, following methods adopted by OAQPS.
      b. Endorse and promote the coordination and use of
            interagency STF research as part of an effective
            research strategy for EPA.


4.3 Recommendation III; Establishment of an Early Warning Group

     EPA should establish a formal and continuing group of senior
scientists and engineers who would be drawn from the Office of
Research and Development, the Office of Policy, Planning and
Evaluation, and extra-Agency groups.  These individuals,
representing a number of disciplines, would be charged with
identifying potential, emerging, and/or escalating public health
and environmental problems.  Such a group would, at a minimum,
perform the following functions:
      a. Survey early warning data sources which can be found
            from modest refinements to existing chemical,
            biological and physical data monitoring systems.
            Such refinements lead to feed forward monitoring,
            which can determine when new, unregulated or
            unselected compounds enter a system or when shifts in
            the distribution of chemicals in  the environment
            occur.  Methods for for analyzing  such data should be
            developed.
      b. Identify potential human health and  environmental risks
            that are  currently not classified  as major EPA
            priorities.   The process  for identifying  such risks
            should include an examination of  social,  economic and
            technological changes that can  create new risks,  use
            of existing models and measurement data,  and
            sponsorship of periodic expert  workshops  to survey
            expert judgment on trends and risks.
      c. Prepare an annual report of  potential new problems  to be
            submitted to  the Administrator, Deputy Administrator,
            and the Assistant Administrators.  The Assistant
            Assistant for ORD should  ensure that this report is
           • formally  considered  in each year's research planning
            process.
                               -24-

-------