United States
Environmental Protection
Agency
                    Environmental Research
                    Laboratory
                    Athens GA 30613
 Research and Development
                    EPA/600/S3-85/001   Mar. 1985
Project  Summary
Testing
TOX-SCREEN
B. R. Bicknell, S. r
  TheTOX-SCREE M model, which was
developed recent! r by the Office of
Toxic Substances of the U.S. Environ-
mental Protection
Agency (EPA), is a
screening level, multimedia model de-
signed to provide i
icals to accumulate
lutant fate and tran
release and dispers
deposition; volatil
leaching; and chem
fer coefficients,
parameters. The
  In this study, an
 and  Evaluation of
               Model
. Boutwell, and D. B. Watson
apid assessment of
the long-term tendancy of toxic chem-
n air, surface water.
and soil. TOX-SCF EEN simulates pol-
 port by considering
on of chemicals into
the air, water anf pollutants across
media interfaces is specified by atmos-
pheric deposition vi locities, masstrans-
and mass loading
model assumes a
generic  positionin j of surface water
bodies and conta ninated land areas
with respect to at nospheric pollutant
sources.
evaluation of TOX-
SCREEN to deteimine its ability to
perform reliable sci eening of chemicals
was conducted.  Evaluation included a
review of the TOX-SCREEN processes
and models relative to the current state-
of-the-art  methods, a comparison of
TOX-SCREEN results with observed
data and  with  predictions of other
models, and a sensitivity analysis of
selected input parameters.
  This Project Summary was developed
by EPA's Environmental Research
Laboratory. Athens, GA, to announce
key findings of the research project that
is fully documented in a separate report
of the same title (see Project Report
ordering information at back).

Introduction
  The Toxic Substances Control Act of
1976 requires that EPA evaluate the risks
associated with the manufacture and use
of chemicals. Additionally, implementa-
tion  of  other environmental control
statutes, such as the Clean Water Act, the
Clean Air Act, the Safe Drinking Water
Act, the Federal Insecticide, Fungicide,
and Rodenticide Act, and the Resource
Conservation and Recovery Act, implicitly
require an assessment of health risks
posed to humans and  other life by the
manufacture and use of chemicals. One
of the effects of these statutes has been
to focus attention on multimedia (that is,
air, water, soil, and biota) considerations
in chemical risk assessments.
  The TOX-SCREEN model was developed
by EPA's Office of Toxic Substances to
provide a multimedia,  screening level
model designed to enable rapid assess-
ment of the  long-term tendency of toxic
chemicals to accumulate in air, surface
water, soil, and ultimately in biota. Be-
cause TOX-SCREEN was designed as a
screening tool for evaluating chemicals, it
was simplified to minimize input require-
ments and was  designed to be overly
predictive or conservative. TOX-SCREEN
estimates pollutant concentrations in the
various media based on chemical property
data, climatic and soil data, simple water
body descriptions, and specific chemical
release scenarios.
  TOX-SCREEN simulates pollutant trans-
port and fate in and between air, soil, and
surface water. Point and area sources of
pollution are considered for atmospheric
dispersion. The  soil model simulates
hydrology and pollutant transport  and
fate in a soil column between the ground
surface and the upper saturated soil zone.
Surface water bodies modeled are rivers,
lakes, estuaries, and ocean outfall areas.
The pollutant fate and transport processes
simulated in the model are shown in
Figure 1.

-------
  Implementation of the TOX-SCREEN
program  was accomplished by linking
separate  code sections for air,  soil, and
aquatic processes. The air model uses a
modified  Gaussian plume equation for
dispersion from  a  point source and a
simple box model  for an  area source.
Pollutant degradation is considered. The
soil model uses Level 3 of the Seasonal
Soil (SESOIL) model. SESOIL considers
rainfall, infiltration, surface runoff, evapo-
transpiration, ground-water runoff, and
chemical fate-transport  processes. The
aquatic models(river, lake, etc.) assume a
completely mixed water body, advection
of pollutants with water, pollutant  inter-
actions/decay, and adsorption to  sus-
pended sediments.
  Physical and chemical processes drive
transport of pollutants across air-soil and
soil-water interfaces. These media  inter-
actions occur in  TOX-SCREEN  via such
processes as volatilization from water
and  soil  surfaces, wind resuspension,
atmospheric deposition, and surface and
subsurface runoff. Estimation of pollutant
fluxes across media interfaces also re-
quires the locations of water bodies
relative to atmospheric pollutant sources
and  contaminated  land areas. In  TOX-
SCREEN, a generic positioning of media
is assumed  in which the contaminated
soil area is always  adjacent to any  water
body. Water bodies are assumed to
intercept the atmospheric pollutant plume
almost  immediately downwind of the
maximum concentration point. If no water
body is simulated, the land area immedi-
ately downwind of this point is assumed
to be the contaminated soil area.
  Pollutant concentrations in the three
media reflect both direct input to any or all
of the media from specified sources, and
subsequent interaction via the processes
listed above. TOX-SCREEN simulates a
1-year period during  which  pollutant
source terms and environmental driving
forces (wind speed, velocity, rainfall, etc.)
may vary monthly. During a time step
(month), the source terms are assumed to
be constant.

Testing Procedures
  The procedures  used for evaluating
TOX-SCREEN were generally  motivated
by two concepts.  First, the multimedia
nature of TOX-SCREEN raises  the ques-
tion of testing processes in several media
concurrently or individually. Lack of com-
prehensive multimedia data to test models
of this type is a problem. Separate testing
 of component models, including the inter-
 media transport models, is the alternative.
 Second, the screening nature of TOX-
                                    2
        Source
      (point, area)
                                Atmosphere
                           A dvect ion/Dispersion
                               Degradation
               Deposition
                (wet, dry)
Wind Erosion
                                                    Deposition
                                                     (wet,-dry)
Volatilization  Erosion and
           Surface Runoff
                                              Volatilization
                                                             \
      Direct  '.Vr'.V-  So/7
      Source
      i|^.'! Advection (leaching) Hydrolysis
       :\ Diffusion fair)
       ':'/ Adsorption
       '{,' Degradation
                                                Water

                                                Advection
                                             Direct
                                             Source
                     Cation Exchange \Ground- Water, Degradation
                     Complexation   '.\  Runoff    Adsorption
                                                Acid/base spec/at ion
                                                  '•'. [SedimentBed]* \\\\\'•;'•':':':
              [Ground Water}*
      [  ]* Not Simulated
Figure 1.    Pollutant pathways simulated in TOX-SCREEN.
SCREEN and its "conservatism" compli-
cate comparison testing with observed
data. Generally, environmental screening
level models are expected to be accurate
to within  approximately one order  of
magnitude. In addition to the simplifica-
tions  incorporated  into TOX-SCREEN,
intentional conservatism also was incor-
porated, which may tend to make its
predictions less accurate when compared
to observed pollutant concentrations.
Thus, validation of the model by standard
field testing methods must consider these
caveats.
  As a result of the above problems, TOX-
SCREEN testing procedures were based
primarily on the individual process models
(intramedia and intermedia) rather than
comprehensive multimedia testing and
on review-evaluation of model algorithms.
Comparison with observed data and other
models was performed, but the overall
evaluation relied more on objective eval-
uation of the model components; compar-
isons made in this project to evaluate
TOX-SCREEN were:

 • Review the transport-transformation
   models in detail for each environmen-
   tal medium (air,  soil, water)  and the
   media interaction models (runoff, vola-
   tilization, deposition);  compare  with
                                     other models and state-of-the-art
                                     methods for  the  processes;  identify
                                     major assumptions, and deficiencies.
                                   • Compare predictions  of  individual
                                     media models with observed data
                                     (where  available) and other model
                                     predictions recognizing the temporal
                                     and  spatial simplifications  of TOX-
                                     SCREEN
                                   • Perform sensitivity testing  of selected
                                     input parameters. This was performed
                                     primarily for  the  atmospheric model
                                     and  atmospheric deposition  param-
                                     eters.

                                   Conclusions
                                     The  evaluation led to a  number  of
                                   conclusions regarding the capability  of
                                   TOX-SCREEN  to perform  reasonable
                                   screening assessments of chemicals with
                                   limited data available. In this analysis, an
                                   expected accuracy for screening method-
                                   ologies of one order of magnitude was
                                   considered. Although testing included in
                                   this analysis primarily involved individual
                                   media, conclusions regarding the overall
                                   multimedia capabilities can be made from
                                   those tests as well as sensitivity testing
                                   and subjective review of the model. Given
                                   the lack of multimedia data available to
                                   test this type of  methodology  and the

-------
deliberate overpredictive nature of TOX-
SCREEN, however, actual validation of
the overall system is difficult.
  Specif ic conclusions of this study were:

• The TOX-SCREEN model is currently
   an acceptable  model  for screening
   chemicals that do notadsorbtoa large
   extent. Both  the aquatic and  soil
   models do  not  adequately consider
   adsorption in their transport simula-
   tions. High Kd pollutants in both media
   will tend to be  advected out  of the
   system too rapidly because bed sedi-
   ments are not modeled in water bodies
   and time of travel of soil pollutant to
   ground water does  not depend  on
   chemical properties.
• Testing of individual media  models
   indicated that reasonable choice of
   input data resulted in predictions that
   satisfied the "one order of magnitude"
   criterion for screening accuracy in
   most cases. TOX-SCREEN may be on
   the boundary of acceptability based on
   this criterion.
• The air model requires careful atten-
   tion because of its increased overpre-
   dictive potential (single wind direction).
   Testing results indicate less confidence
   that the  one order of magnitude
   criterion will be satisfied.

• Predicted pollutant concentrations in
   rivers  and  lakes will generally  be
   acceptable,  based on testing  results
   and analysis. Highly adsorbed chemi-
   cals will be questionable because of
   lack of a sediment bed and inappro-
   priate  simulation of suspended sedi-
   ment concentrations.
• The soil model  acceptability  largely
   remains to be determined.  Compari-
   sons between SESOIL predictions and
   observed data were generally accep-
   table, but testing and  analysis suggest
   deficiencies in soil layering and the
   pollutant transport cycle
• Individual process models,  although
   necessarily  simplified,  are generally
   based on  current state-of-the-art
   mechanistic methods forfate-transport
   modeling. Empirical methods for some
   soil hydrologic processes and pollutant
   erosion processes are exceptions.
• All individual media models are rela-
   tively convenient to use, that is, input
   data requirements are minimal and
   typically available. The SESOIL model
   considers  relatively more complex
   chemical  processes,  but these pro-
   cesses and associated input can gen-
   erally be ignored, if desired.
• The code is somewhat confusing, and
  confirmation of "correct" operation is
  difficult. Mass conservation is a critical
  question in TOX-SCREEN; verification
  of mass  balance should be incorpo-
  rated. The documentation is appropri-
  ate.

Recommendations
  A number of improvements in TOX-
SCREEN were identified in this study and
are recommended for further considera-
tion.  Many of  these recommended im-
provements would require significant
effort to implement, and in some cases,
would  slightly increase  required user
input. If TOX-SCREEN is to be applied to a
wide spectrum of  chemicals, however,
more consideration of adsorption (aquatic
and soil) should be included. Recommend-
ed changes considered to be most im-
portant include:

Atmospheric  Dispersion Model
• Implement  modification to buoyancy
  flux calculations.
• Provide guidance to model users for
  avoiding errors  caused  by parameter
  sensitivity. This could include recom-
   mendations to test sensitivity of key
  parameters in all model applications.
So/7 Model
• Add  additional  layers to decrease
  mathematical  dispersion problems.
• Provide  guidance to the  user  for
  avoiding errors caused by sensitivity to
  the soil disconnectivity parameters.
• Eliminate the  threshold that prevents
  "computation" of pollutant processes
  at low concentrations.
• Modify the  pollutant cycle to include
  retardation m the computation of "time
  of travel" to ground water.

Aquatic Models
• Account for key water column-sedi-
   ment  bed  interaction  processes  in
   lakes, rivers, and estuaries.
• Improve computation of  suspended
   sediment concentration by accounting
   for low stream velocities.

Intermedia  Transfer Models
• Provide  additional user guidance for
   estimating deposition parameters in-
   cluding table of values for representa-
  tive and classes of chemicals.

D o cumen t ation
• Improve  user  guidance for defaults
  because screening involves the use of
   uncertain data for new and  poorly
   characterized chemicals.
•  Improve code structure (shorter sub-
   routines, less complicated branching,
   indentation).
•  Modify output of information contained
   in "Monthly Pollutant Concentrations
   and Interaction Terms" to more closely
   resemble SESOIL output tables.

General
•  Investigate the modifications and level-
   of-effort required to incorporate mass
   balance computations.
•  Incorporate a ground-water model.
•  Provide guidance for performance of
   sensitivity  analysis  as part  of  any
   assessment. Provide list of key sensi-
   tive parameters and typical sensitivity
   ranges. This could also be incorporated
   into the code by allowing the user to
   entera range of values fora parameter
   and have  the  model  compute the
   corresponding range of predictions.

-------
    B. R. Bicknell, S. H. Boutwell. andD. B. Watson are with Anderson-Nichols & Co.,
       Inc., Palo A/to, CA 94303.
    L. A. Mulkey is the EPA Project Officer (see below).
    The complete report, entitled "Testing and Evaluation of TOX-SCREEN Model,"
       (Order No. PB 85-160 356/AS; Cost: $14.50.  subject to change) will be
       available only from:
            National Technical Information Service
            5285 Port Royal Road
            Springfield, VA 22161
            Telephone: 703-487-4650
    The EPA Project Officer can be contacted at:
            Environmental Research Laboratory
            U.S. Environmental Protection Agency
            College Station Road
            Athens, GA 30613
                                      •t, U.S. GOVERNMENT PRINTING OFFICE: 1985-559-016/27004
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
                                 OCOC329   PS
                                 U  S  ENVIR  PROTECTION  *<5ENCY

-------