PB85-214310
DIOXIN TRANSPORT FROM CONTAMINATED SITES TO EXPOSURE LOCATIONS:
A METHODOLOGY FOR CALCULATING CONVERSION FACTORS
Office of Hazardous Waste Management
Richland, WA
 Jun 85
                U.S. DEPARTMENT OF COMMERCE
             National Technical Information Service
                              NTIS

-------
                                                               EE85-21U310
                                                      (EPA/600/8-85/012
                                                      June 1985'
OIOXIN TRANSPORT FROM CONTAMINATED SITES TO EXPOSURE LOCATIONS
       A METHODOLOGY FOR CALCULATING CONVERSION FACTORS
                          Prepared by

                       Gaynor W. Dawson
                        Ji'11 M. Meuser
                        Mary C.  Lilga
             Battelle Project Management Division
             Office of Hazardous Waste Management
                      Richland,  MA  99352
                      EPA Project Officer
                          John Schaum
                    EPA Contract 68-01-6861
         OFFICE OF HEALTH AND ENVIRONMENTAL ASSESSMENT
              OFFICE OF RESEARCH AND DEVELOPMENT
             U.S. ENVIRONMENTAL PROTECTION AGENCY
                    WASHINGTON,.DC 20460

-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
1. REPORT NO.

  EPA/600/8-85/012
                             2.
                                     3. RECIPIENT'S ACCESSION NO.
                                           5   27. >o 10 /AS
4. TITLE AND SUBTITLE
   Dioxin Transport from Contaminated Sites To Exposure
   Locations:  A Methodology for  Calculating Conversion
               Factors
                                     5. REPORT DATE

                                         June 1985
                                     6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
   Gaynor W. Dawson
   Jill M.  Meuser
Mary C. Lilga
8. PERFORMING ORGANIZATION REPORT NO.


  B556-21101
9. PERFORMING ORGANIZATION NAME AND ADDRESS
   Office of Hazardous Waste Management
   Battelle Project Management  Division
   601 Williams Blvd.
   Richland, WA   99352
                                                           10. PROGRAM ELEMENT NO.
                                      1 1
                                              CT/G RANT NO.
                                       Subcontract EPA  38-9
                                       Unrk flcginnmonl"  11
12. SPONSORING AGENCY NAME AND ADDRESS
   Exposure Assessment Group
   US EPA, Office of Research  and  Development
   Washington, DC   20460  •
                                      13. TYPE OF RETORT AND PERIOD COVERED
                                       Final 7/1/84 -  1/15/85
                                      14. SPONSORING AGENCY CODE
                                                             EPA/600/21
16. SUPPLEMENTARY NOTES
16. ABSTRAC
         Procedures have been  developed by the US EPA for estimating  the risk
    associated with exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin  (dioxin).
    Cdncentrations of dioxin at the contaminant source are usually known, but
    exposure may occur  at  locations away from the source where  concentrations are
    usually unknown.  As a  result,  a need was identified for estimating  dioxin
    concentrations away from the source.
         This report discusses the  transport of dioxin from a source  and presents
    methods for estimating  dioxin concentrations at potential points  of  exposure
    away from a source.  The transport pathways that were considered  to  be important
    were volatilization, suspension and depositon of windblown  particles, overland
    sediment runoff, and in-stream  sediment transport.  Concentrations at locations
    away from a source  can  be  estimated using conversion factors  for  air, soil, and
    sediment.  Concentrations  in these media at potential points  of exposure can be
    estimated using the source concentration and factors that describe the physical
    characteristics of  the  source and the transport pathways.
         Because ingestion  of  contaminated foodstuffs will result in  exposure to
    dioxin, an example  is  provided  for estimating the amount of dioxin in beef.
    Missouri beef distribution patterns and a market dilution concept were used to
    estimate potential  chronic exposure to contaminated beef products within the
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                        b.lDENTIFIERS/OPEN ENDED TERMS  C. COS AT I Field/Group
  2,3,7,8-Tetrachlorodibenzo-p-dioxin  (Dioxii
  Contaminant Transport
  Conversion Factors
  Environmental Contamination
18. DISTRIBUTION STATEMENT
   Distribute to Public
                                              19. SECURITY CLASS /This Report/
                                                Unclassified
                                                   21. NO. OF PAGES
                                                         89
                                              20. SECURITY CLASS (This page 1

                                                Unclassified
                                                                        22. PRICE
EPA Form 2220-1 (R«». 4-77)   PREVIOUS EDITION is OBSOLETE

-------
                                  DISCLAIMER

     This  report  has been  reviewed  in  accordance  with  U.S.  Environmental
Protection  Agency  policy and  approved  for  publication.   Mention  of  trade
names   or   conmercial   products   does    not   constitute   endorsement   or
recommendation for use.
        Reproduced from
        best available copy.
                                      -ii-

-------
                              TABLE OF CONTENTS


                                                                         Page

Tables and Figures		      iv

Foreword	       v

Abstract	      vi

Acknowledgements 	     vii

1.0  Introduction	       1

2.0  Summary	       8

3.0  Atmospheric Concentrations	      13
     3.1  Introduction .	      13
     3.2  Particulate Exposure ..... 	      13
     3.3  Gaussian Dispersion	.  .      22
     3.4  Vapor Exposure	 . . '.	      24

4.0  Soil Concentrations .......	      27
     4.1  Introduction . . . .	      27
     4.2  Oioxin Behavior in Soil	      27
     4.3  Photodegradation	      28
     4.4  Volatilization	      30
     4.5  Approach	      30
     4.6  Average Soil Losses. . .	      33
     4.7  Soil Deposition	      41
     4.8  Wind Deposition	  .      46

5.0  Sediment Concentrations 	      49
     5.1  Introduction . . .	      49
     5.2  Dioxin Behavior in Water	      49
     5.3  Sediment Transport in Streams.	      50

6.0  Missouri Beef Distribution Patterns 	      56

References	      63

Appendix A - Example Site 1	"	      68

Appendix B - Example Site 2	      74
                                    -iii-

-------
                              TABLES AND FIGURES

Table 1.       Physical Properties  of 2,3,7,8-Tetrachlorodibenzo-
               p-dioxin	     2
Table 2.       Average K Values for Soils on Erosion
               Research Stations  	  	    37
Table 3.       General Magnitude of the Soil/Erodibility Factor,
               K, When Organic Content Data are Available	    38
Table 4.       Values of the Topographic Factor, LS, for Specific
               Combinations of Slope Length and Steepness	    39

Figure 1.      Dioxin Structure	     1
Figure 2.      Pathways- for Exposure from Contaminated Soil	     4
Figure 3.      Relationship Between Conversion Factors . .	     8
Figure 4.      The Product of ayaz  as a Function of
               Downwind Distance from the Source for each
               of the Six Stability Classes	    10
Figure 5.      Relative Resuspension Factors Under Various
               Site Conditions	    16
Figure 6.      Average Annual Values of the Rainfall
               Erosion Index 	    35
Figure 7.      Generalized Relationship Between Size of
               Drainage Basin Area and Sediment Delivery Ratio  ....    43
Figure 8.      Origin of Beef Consumed in Missouri	    60
Figure A-l.    Example Site 1	    69
Figure B-l.    Example Site 2	    75
                                     -TV-

-------
                                   FOREWORD


     The  Exposure  Assessment  Group  (EAG)  of  EPA's Office  of  Research  and
Development has  three  main functions: 1) to  conduct exposure assessments; 2)
to review assessments  and  related  documents;  and 3) to develop guidelines for
Agency  exposure  assessments.   The activities  under each of  these functions
are supported by and respond to the needs of the various EPA program offices.
In relation to the  third function, EAG  sponsors  projects  aimed at developing
or refining  techniques  used  in exposure assessments.   This study  is  one of
these  projects  and  was done  for  the   Office  of  Solid  Waste  and Emergency
Response.

     Dioxin  problems  first  surfaced  in the  U.S.   in  the  early  1970's with
Agent Orange and the Missouri  Horse Arenas.   Since then, dioxin contamination
has been  found  elsewhere  in  Missouri,  Arkansas, Michigan,  New York,  and New
Jersey.   EPA has become  increasingly involved  in  the  discovery, assessment,
and clean-up  of these sites.   The  purpose  of  this  document  is  to  provide
methods  to  use  in  conducting  exposure  and  risk  assessments  of   dioxin
contamination sites.

                                                James W. Falco, Director
                                               Exposure Assessment Group
                                     -v-

-------
                                   ABSTRACT

     Procedures  have  been developed  by  the U.S. EPA  for  estimating the risk
associated  with  exposure  to  2,3,7,8-tetrachlorodibenzo-p-dioxin   (dioxin).
Concentrations  of dioxin  at  the  contaminant source  are  usually  known,  but
exposure may occur at  locations  away from the source where concentrations are
usually  unknown.   In  response  to this  problem,, a  need  was  identified  for
estimating dioxin concentrations away from the source.
     This report  discusses  the  transport of dioxin  from a source and presents
methods  for  estimating dioxin concentrations at  potential  points  of exposure
away  from a source.    The  transport  pathways  that were  considered  to  be
important  were   volatilization;   suspension  and   deposition   of  windblown
particles;   overland   sediment   runoff;   and  in-stream  sediment  transport.
Concentrations  at  locations   away  from  a  source   can  be  estimated  using
conversion factors  for  air,   soil,  and  sediment.    Concentrations  in  these
media  at potential  points of  exposure  can be  estimated  using   the  source
concentration and factors  that  describe the physical characteristics  of the
source and the transport pathways.
     Because  ingestion  of contaminated foodstuffs  results  in exposure  to
dioxin,  the  report  includes  an  example  of how  to estimate  the  amount  of
dioxin  in  beef-  Missouri  beef distributic-  patterns and  a  market dilution
concept were used to estimate potential  chronic  exposure to contaminated beef
products within the state.
                                     -vi-

-------
                              ACKNOWLEDGEMENTS

     The authors would  like to thank  the  following individuals who  provided
valuable comments during the peer review process:
     Stuart M. Brown (CH2M HILL)
     Jerald L. Schnoor (University of Iowa)
     Louis J. Thibodeaux (Louisiana State University)
The guidance, aid, and encouragement of  the  Project Officer,  John  Schaum, are
gratefully acknowledged.
                                    -vn-

-------
       DIOXIN TRANSPORT FROM CONTAMINATED SITES TO EXPOSURE LOCATIONS:
              A METHODOLOGY  FOR CALCULATING CONVERSION FACTORS

1.0  INTRODUCTION
     A  great  deal   of  literature  has  been  published  recently  concerning
2,3,7,8-tetrachlorodibenzo-p-dioxin.   This compound  is  one extremely  toxic
member  of   a  class  of  compounds  containing  the  basic  dioxin  nucleus
(Figure 1).
                                                   O'
     Dioxin Nucleus               _2(3,7.8-Tetr»chlorod1benzo-p-dioxin

                         Figure 1.  Dioxin Structure

     There    are    75    possible   chlorinated    dioxins,     including    22
tetrachlorodibenzo-p-dioxins.  However,  the  2,3,7,8-tetrachloro isomer  is one
of  the  most  toxic  substances  known.    Throughout this  report,  the  term
"dioxin"  has  been used  to  refer  to  the   2,3,7,8-tetrachloro  isomer,  the
properties of  which are  given  in  Table 1.    Although  other  isomers may  be
transported  by- the mechanisms  described, in  this  report,  the other  isomers
have  different   physical   properties   which  may   render   inapplicable   the
concentration relationships derived for the 2,3,7,8-tetrachloro isomer.

-------
     Table  1.   Physical  Properties  of 2,3,7,8-Tetrachlorodibenzo-p-dioxin
                          Esposito et    Mabey  et     Freeman  and    Perkaw et
                           al.. 1980    a!.. 1981   Schroy. 1984   al.. 1980
Molecular Weight             322          322            —           322
Melting Point, °C            305            —            —         303-305
Vapor Pressure at 25°C,                      ,              o        c    7
 mm Hg                         -           10'6       1.5xlO"9     lO^-NT'
Water Solubility (ug/L)        0.2 .         0.2        0.317          0.2
Octanol-Water                                  .                           _
 Partition Coefficient         —         6.9x10°         --         1.38x10'
     In response  to  the  discovery of a growing  number  of dioxin-contaminated
sites, the  Exposure  Assessment Group within  the U.S.   Environmental  Protec-
tion  Agency Office  of  Research  and  Development has  drafted procedures  for
estimating the human  health  risk  associated with these  sites  (Schaum,  1984).
The procedural algorithm developed for .calculating exposure is of the form:
              Dioxin
           Concentration
lifetime    ln ^°^ at   Y Conversion   Contact   Exposure   Absorption
Average  =    Source     x   Factor   x  Rate       Rate   x  Fraction
Exosure               Bod   Weiht  x  70   r Lifetime
Average  =     ource          acor       ae
Exposure                Body Weight x 70 yr  Lifetime

The algorithm  contains a  conversion factor that  relates contamination  at  a
point  of  exposure  to  contaminant  levels  at  a  site  (e.g.,  conversion  of  a
dioxin concentration  in soil to  a  downwind  airborne dioxin  concentration).
All conversion factors  are based  on  dioxin  concentrations  in  soil   at  the
primary contaminant  source.   The  purpose "of this report  is to  describe  and
                                     -2-

-------
quantify  the  conversion  factor.    The  other  factors  in,  and  use of,  the
exposure algorithm are described by Schaum (1984).
     Depending on exposure mechanisms,  it  is  apparent that several conversion
factors must  be  considered.    Exposure  to contaminants  at  a point  away from
the source  may  occur as a result  of contaminant transport  to  receptors by a
number of  potential  routes,  as  illustrated  in  Figure  2  for  a  generic site
contaminant.   Due to  the intrinsic properties  of  dioxin,  there  is  limited
pdtential for transport  in the  dissolved  phase.   Thus, certain  pathways  in
Figure 2  may  be  ignored.   While dioxin may  be transported by  these  routes,
the concentrations at  any given  time would  be low  and would not  result  in
high exposure risks.
     The pathways of primary  interest are  those associated with the transport
of solid  particles containing adsorbed  dioxin.   The  following mechanisms are
   *•
considered to be important in the transport of dioxin from a site:
•    resuspension and deposition of windblown particles;
•    sediment runoff; and
t    sediment transport in streams.
In addition,  recent  literature  suggests  that  volatilization of  dioxin from
contaminated soils may occur,  despite the  very  low  vapor pressure of the pure
compound  (Thibodeaux,  1983;   Freeman  and  Schroy, 1984).   Consequently,  vapor
transport is  a  fourth  mechanism that must be considered  when  evaluating the
transport of dioxin from a source area.
     Based  on   the   considerations  discussed   above,   the  development  of
conversion  factors  for  use   in  the exposure  algorithm was  limited  to five
cases as follow, where [dioxin] refers to concentration  of dioxin:
                                     -3-

-------
                    SOURCE AREA


                  CONTAMINATED SOIL
                                               ATMOSPHERE
                                                     en
                                                     c

                                                     CT
                                                     c
                                                     Q>
                                                     u
in
O
Q.
                                               SURFACE WATER
Figure 2.  Pathways for Exposure from Contaminated  Soils

-------
        •      [dioxin] in Air at Point of Exposure  (uq/m^)
       air "  [dioxin] in Soil at Original Source (ug/Kg)
2\   £p  ., _ [dioxin] in Soil at Point of Inqestion  (ug/Kg)
       50       [dioxin] in Soil at Original Source  (ug/Kg)
3)   CF  -i -  [diPX"*"] in Soil at Point of Contact  (ug/Kg)                 ,.\
       5011    [dioxin] in Soil at Original Source (ug/Kg)                  (  '
4)   CF       _ [dioxin]  in Pasture Soils  (ug/Kg)                      /5v
       5011 "  [dioxin] in Soil at Original Source (ug/Kg)
5\   CF           [dioxin] in Sediment where Fish are Caught  (uq/Kg)       ,fiv
 '     sediment -      [d1oxin] 1n Soi1 at Original  Source  (ug/Kg)          {  '
     CF •   is  dependent  on  both  particulate  and  vapor emissions  from  the
source.   Section  3.0 describes  the conversion  factor  for  particulates  and
discusses recent research concerning possible vapor  emissions.
     The second,  third,  and  fourth  factors (CFsoi-i)  differ  only in the  mode
of exposure.   The  physical   processes  that transport  contaminated  soil  from
the  source  to the  point  of exposure  are  the  same  in  all  three  scenarios.
Thus, the CF   -j-,  for  all  three  is  the  same.   The  treatment  of the different
exposure routes is  discussed in  the predecessor paper on appropriate exposure
algorithms (Schaum, 1984).   The  derivation of CF$oil  is discussed  in  Section
4.0.
     The fifth factor,  CF  ^.   ^,  relates in-stream sediment  concentrations
to  source  strength.     It  is   dependent  on  runoff-derived  particles  that
accumulate  in  the  stream  bed.    Estimation  of  CF  ..    t  is  discussed  in
Section 5.0.   A summary  of the three factors  (CFair, CFsoil,  and CFsediment)
is provided in Section 2.0.
                                     -5-

-------
     Each  of  the  concentration  factors presented  in  this report  is derived
from  generalized  ranges of  environmental  observations.   The  relationship
between  anthropogenic,  topographic,   hydrologic,  climatic,  and  vegetative
influences at a  particular  site  will  be very complex.   However, consideration
of  these  influences  generally  requires  computer modeling,  and   for  some
influences the mathematical models do not exist.
     The concentration  factors  developed  in this  report  are  based  primarily
on average, gross  transport.   Airborne  contaminant transport  is assumed to be
downwind from a  source.   Areas of concern  from overland transport will  be
natural  drainageways and  topographic  breaks.    Finally,  catastrophic events
such as  hurricanes,  tornadoes,  and floods  are  not included in the derivation
of   conversion   factors,   although  such   events  may   effect   significant
environmental  transport of   contaminants.    For  example,   Collier  (1963)
reported that the  sediment  yield from a single-day storm exceeded  40% of the
yield for  that  year and exceeded the  annual sediment yield  for  the previous
three  years.    Despite  the potential  importance  of  catastrophic  events  in
transporting  contaminated  material, the factors  that  describe  the  events can
only be  considered for  specific  sites.   These  events are  not  easily reduced
to  generic description  due  to   their  intensity, variability, and   irregular
occurrence.   When estimation of acute  event  transport is  necessary,  ASCE
(1975) should be consulted.
     It  must  be emphasized that the conversion  factor  approach  is  a survey
method to provide  rapid  but approximate estimates of  dioxin transport and its
implications.   Ideally,  empirical monitoring data or  sophisticated  numerical
modeling approaches  would  be employed  for more quantitative estimates,  as
                                     -6-

-------
recommended  by  Schaum  (1984).    Conversion  factors  are  useful  to  screen
candidate  source sites  and  to  help  prioritize  those  sites  for which  more
quantitative estimates are needed.
     The  final  section  of   this  report   is  concerned  with  Missouri  beef
distribution patterns.   The  discussion  in  Section 6.0  is  intended  to provide
a  means for  determining  the extent  to  which  Missouri  inhabitants may  be
exposed to dioxin through  consumption  of contaminated beef.   The emphasis  is
placed  on  developing  a means- of estimating the fraction of  beef consumed  in
the state that is likely to  have  come from contaminated herds.  Schaum (1984)
has developed  a method  for  calculating dioxin concentrations  in beef  as a
function  of soil  levels  in  the pasture  area.    These  concentrations   are
determined  by  using  the conversion factors  (CF  .,)  derived in this  work.
Hence,  this  report   complements  Schaum's  work   by  providing  a  complete
methodology for  determining,  on  a qualitative  basis, possible dioxin exposure
at sites in Missouri.
                                     -7-

-------
2.0  SUMMARY
     Human  exposure  to dioxin may  occur through contact with  soil  particles
on  which  dioxin  is  adsorbed.   Contaminated  soil may  be transported  from  a
dioxin-contaminated  source   area  by  several  mechanisms:     suspension  and
deposition of windblown  particles;  overland  transport  and deposition of .soil;
and  in-stream  sediment  transport.    The  relationship  between  conversion
factors is illustrated in Figure 3.
              Figure  3.   Relationship  Between  Conversion  Factors
     The  conversion   factor  for   atmospheric   concentrations   of   dioxin,
depending on  wind  speed,   is  determined  by  both  the  particulate and  vapor
levels of  dioxin, as follows:
          CF .   =-= (l x 1(T7 [W2(W"- 9)]}
                  Soil
(7),
                                     -8-

-------
                                                         o
where C .   = dioxin concentration in air above soil (ug/m )
       a 11
         W = average wind velocity (m/sec)
     CSQ.-|  = dioxin concentration in soil at point of interest "(
                                                               P
      °"y°"z = product of the Gaussian dispersion coefficients (m ),
             from Figure 4
         F = fetch or downwind dimension of the source (m)
         x = distance from source boundary to the point of interest  (m)
If W < 10 m/sec, the expression reduces to:
          CFair -f^ • l x 10~7
                  usoil
(8)
These  expressions  can  be  used  to  estimate  atmospheric  concentrations  of
dioxin,  Ca^r>. at the  source  if  contaminated source conditions  are used  for
C  .,.   To  determine  atmospheric  concentrations  downwind  from  the  source,
soil   concentrations,  either calculated  from conversion  factors  or  measured,
are  used for  C$oil.   If  Csource  is  used  for C$oil,  the  degradation  and
dispersion factors become unity.
     Dioxin  concentration  in   the  surface  soil   in   the  vicinity   of   a
contaminated source  is a  function of the amount  of sediment- delivered to  the
point of exposure.  The  conversion factor for overland transport  is:

          PC       Csoil	   Lsource x Asource                            ,Q,
          CFsoil=p	=-r	—	                          .(9)
                    source     basin x  basin
where CS01--| = (fioxin concentration  in soil at point of interest (ug/Kg)
    C       = dioxin concentration  in soil at the source  (ug/Kg)
                                     -9-

-------


Reproduced  from   ^f
best available copy, ^m
              Figure  4.   The Product of  OyOz  as  a Function of Downwind
                          Distance from the  Source (from Turner,  1970)
                          for each of- the  Six  Stability Classes A-F
                                        -10-

-------
    LSQurce = estimated  soil  loss  from  the  , Universal   Soil  Loss  Equation
              (USLE) for the source (tons/acre/yr)
    Asource = source area (acres)
     Lbasin = estimated  soil  loss  from  the USLE  for the  watershed  upstream
              of the point of interest (tons/acre/yr)
     Abasin = watershed area upstream of the point of interest (acres)
The conversion factor for windblown contributions to soil contamination is:
          PF  -, .
          CFsoil =
                   -source
                                                               o
where ay0z = product of the Gaussian dispersion coefficients (m ),
             from Figure 4
         x = distance from source boundary to the point of interest (m)
         F = fetch or downwind dimension of the source (m)
If runoff  patterns and  wind  direction  are  coincident,   the  contributions to
soil  contamination  are  additive and the expression  for  the conversion factor
becomes :
                    Csoi1    Lsource x Asource  .
                                        -
                   ^source    Lbasin x Abasin      yzx+F
     In order  to  determine the dioxin concentration  on  sediment delivered to
a stream  in  the vicinity of  a  contaminated  source,  the  procedure is the same
as that used to determine  surface soil  concentrations.  The conversion factor
for  estimating stream  sediment  concentrations  at distances  downstream from
the source is:
                                     -11-

-------
                        f           I        v A
           rr            sediment    source    source
           u sediment = ~c        =  L   .   x A,
                          source      basin    basin

 where   Csediment = dioxin  concentration   in  sediment  at   the  point   of
                     interest (ug/Kg)
                   s estimated  USLE soil  loss for source (tons/acre/yr)
           Asource = source area
            h>asin s estimatecl s01^  l°ss ^rom tne U^LE for the watershed
                     upstream of point of interest { tons/acre/yr)
            \asin a watershed area  upstream of the point of interest
                     (acres)
      The  conversion  factors  developed  for  the three exposure  modes  have
 limitations.   Because  chronic,  long-term exposure  rates are of concern,  the
•methods  developed utilize average  or estimated values for parameters  such  as
 wind   velocity,   precipitation,   runoff,   soil   erodibility,    topography,
 vegetative    cover,   and    stream    characteristics.      Because   of   these
 generalizations,  the methods  must  be  applied with  caution, particularly  in
 areas such  as  the  western  United  States where  little experimental work  on
 sediment yields  has  been  conducted.
      Use  of conversion factors  is  further  limited  by the absence  of  actual,
 comprehensive site  data  with  which  to  test  the  methods developed by  this
 study.    Further  site  studies,  designed  to  collect  the necessary data  on
 dioxin  concentrations  and  environmental  characteristics  at  a  source  and  at
 potential  points  of exposure,  will   be necessary  in order  to  verify  these
 methods.
                                     -12-

-------
3.0  ATMOSPHERIC CONCENTRATIONS

3.1  Introduction
     Human exposure to dioxin may  result  from  inhalation  of contaminated soil
particles.suspended in the  atmosphere  or  vapors  resulting from volatilization
from contaminated  soil.    In  order to  quantify  exposure, it  is  necessary to
derive a  means  of  estimating  atmospheric dioxin  levels  in the environs  of a
contaminated source.

3.2  Particulate Exposure
     Particulate exposure   results  from  the  presence of dioxin-contaminated
soil in  the  atmosphere.   Suspension of  soil  particles in  the atmosphere may
result from  the erosive action  of wind or from  activities  which disturb the
soil, such as  plowing  or excavation.   The  concentration of  particles  in the
air column and  their residence  time are highly dependent on particle size and
atmospheric  conditions.   Ideally,  the conversion  factors  should  be derived
from studies in which'  the  soil  size fractions  were isolated,  and in which the
concentration  versus  size  fraction  in the  air  column  was known  at various
distances from  a source.    Unfortunately, no data were found on particle size
distribution of dioxin-contaminated  soils.   However,  data do exist for dioxin
concentrations  on  fly ash  and   larger  ash  particles  caught  in precipitators
(Fred C. Hart, 1984).  These  data  indicate  that  dioxin levels may increase by
a factor  of  2  to  12  on  smaller size.particles  compared  to larger particles.
                                     -13-

-------
If  this  relationship were  found  to be true for  soils,  it would provide some
guidance  for adjusting  airborne  particulate  level  estimates  based  on soil
conditions.   In  the"  absence of  data to confirm increased concentrations on
small  soil  particles,  selection  of  exposure  factors has  been based  on the
assumption  that  concentrations  of dioxin adsorbed on suspended particles will
be  the  same  as  dioxin  concentrations  in  the  bulk  soil   source.    This
assumption,  in   turn,  assumes  that  dioxin   will  be  adsorbed  on smaller-size
particles,  which  are  subject  to  resuspension  and  respiration,   at   levels
comparable  to  concentrations in  bulk soil.    The assumptions  conform to work
by  Thibodeaux  (1983),  who found  dioxin levels  on  dust  in  the  air  at the
Vertac site to be 1.1 ug/Kg compared  to soil levels of 1.3  ug/Kg.
     Conversion   factors   can  be  derived  by   deterministic  or  empirical
approaches.   Deterministic  approaches use  numerical models  to  simulate the
basic  phenomena  involved.   The models  mathematically describe  the physical
processes that effect transport.   Many models  have been developed to describe
resuspension  and  deposition;  however,  inadequate  data   are  available  to
validate  these  models  (Sehmel,  1980).'  For those models   that  do  exist, the
large  number of  input requirements can  be  severely  limiting.   Sehmel  (1980)
lists   over   40" factors    which    influence   resuspension,    although  the
relationships  between  these  factors  are  not thoroughly  understood.   This
degree   of   complexity   is  too   great  for   derivation  of   simple  site
characteristic guides  for  predicting atmospheric contaminant  or  particulate
levels.    Gillette   (1973)  employed  a  simplified  relationship to  describe
horizontal flux  (Fh)  of particulates:
                                     -14-

-------
          Fh = Ch W*(W*-Wt)                                             (13)
where W* = wind shear velocity (m/sec)
      Wt = a threshold velocity (m/sec)
      Ch = empirical constant
However, because  Ch is  empirically derived, site-specific  data  are required
to calibrate the algorithm.
     The empirical  approach  is  based  on  correlation  analysis  of  data  from
actual  sites,  with subsequent  selection  of  a  factor  that  best  matches  the
relationship between soil  contamination and atmospheric contamination levels.
As  noted  previously,   this  approach  is   required  to  calibrate  Gillette's
simplified model.   A  single datum has been found for  atmospheric  levels of
dioxin  at,  near,  or  downwind  from  contaminated sites.    Thibodeaux  (1983)
reviewed monitoring data  from the Vertac  site  and  found  atmospheric  dust
concentrations of  54  ug/m  ,  dioxin  concentrations on the  dust  of 1.1 ug/Kg,
and  soil  dioxin   levels   of  1.3   ug/Kg   over  a  37  day  period.    These
                                          o     i
concentrations yield a  CF,.   of 4.6 x 10    Kg/m .  Based  on only this datum,
                         air
an empirical  approach  specific to  dioxin  cannot  be  determined  at this time.
However, data  are  available  on the  relationship of  particulate contaminant
levels  and surface  contaminant  levels for  a range of wind conditions and for
a  variety  of  soil  disturbing  activities.    These values  are  summarized in
Figure  5.  The resuspension  factors  in Figure  5 were determined by comparing
atmospheric contamination  (volumetric)  to  surface soil  contamination  (areal)
for  particulate-based   contaminants.    The  areal  measurements  were possible
                                     -15-

-------
            REFERENCE
STEWART, 1967
STEWART. 1967
STEWART, 1967
LANGHAM, 1971
STEWART. 1967
STtWART. 1967
STEWART, 1967
CALC. FROM MILHAM et al.
CALC. FROM MIUIAM ot al.
CALC. FROM MILHAM et al.







. , 1976
. , 1976
. , 1976
STEWART,  1967
STEWART,  1967
STEWART,  1967
STEWART.  1967
CALC.  FROM BENNETT,  1976
CALC.  FROM MILHAM et  al.,  1976
CALC.  FROM IRANZO AND SALVADOR,
STEWART,  1967
CALC.  FROM MILHAM et  al.,  1976
MILHAM et al., 1976
CALC.  FROM MYERS et  al.,  1976
CALC.  FROM MILHAM et  a].,  1976
STEWART,  1967
ANSI'AUGH  Pt  al. , 1970
STEWAIIT,  1967
STEWART.  1967
STEWART,  1967
SEHMEL AND ORGII.L, 1973
CALC.  FROM IRANZO AND SALVADOR,
LANGHAM,  1971
STEWART,  1967
CALC.  FROM BENNETT,  1976
STEWART,  1967
SEHMEL AND LLOYI), 1976
CAIC.  BY  RENftETT, 1976
HAMILTON  CALC. BY BENNETT, 1976
KRF.Y et al., 1976
ANSPAUGH  et  al., 1975
	  	LOCATION	

       MARAL1NGA TRIALS
       MARALINGA TRIALS
       MARALINGA TRIALS
       NEVADA TEST SITE
       MARALINGA TRIALS
       MONTI HELLO ISLANDS
       MONTE BELLO ISLANDS
       FIELD
       FIELD
       FIELD
       MONTE BELLO ISLANDS
       C. D. TRIALS
       AUSTRALIAN DESERT
       MARALINGA TRIALS
       NEW YORK
       FIELD
1970   PALMARES, SPAIN
       AUSTRALIAN DESERT
       FIELD
       FIELD
       SLUDGE
       FIELD
       MARALINGA TRIALS
       NEVADA TEST SITE
       SANDY-GRASS
       SANDY-DEBRIS
       MONTE BELIO ISLANDS
       ROCKY FLATS
1970   PALMARES, SPAIN
       NEVADA TEST SITE
       PAVING STONES
       HEW YORK-FALLOUT
       SANDY-CLEARED
       HANFORD
       HEW YORK-U
       UNITED KINGDOM-U
       ROCKY FLATS
       NEVADA TEST SITE
                                                                STRESS
STIRRED  DUST
VEHICLE, 0.3 m
WALKING
VEHICULAR
CAB. LANDROVER
VEHICLE, 7TII DAY
VEHICLE, 4TH DAY
MOWING
PLANTING.  DISKING
SUBSOIL ING
VEHICLE, 7TH DAY
WORK, OPEN
VEHICLE  TAILBOARD
VEHICLE, 1-2 DAY
FALLOUT  CONCS.
TRACTOR. DOWNWIND
YEARLY FARMING
WALKING
TRACTOR CAB
TRACTOR
ROTOTILLING
FERTILIZING
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
WIND
                                                                                    1ECHANICAL RESUSPENSIOM  STRESSES
                                                                                                                   t-
MIND RESUSPENSIOH STRESSES
                                                      1  . . .1^  , , ,I..J  . . .I.-J  . . .1	I  .  . .1_J  . . ,l.fl  i I llalJ

                                                      10-10  ui-9   jo-8    10-7    i0-6   jo-b    i0-4   10'

                                                                             RESUSPEHSIOH FACTOR, m-1
                                                                                                                                             10
                                                                                                                                               -2
              Figure  5.   Relative  Resuspension Factors* Under Various  Site  Conditions  (After  Sehmel,  1980)

 *Resuspension  Factor  (RF)  = Ratio of Contaminant per  Volume  Air to Contaminant  Per Unit Area of Soil  (nr*  units)

-------
because radio-contaminant  levels can  be measured  in  situ without  regard to
sample  depth.    For  chemicals  such  as  dioxin,  concentrations  would  be
required.
     An alternate  method of calculating particulate  levels  in the atmosphere
is  to  compare  atmospheric  monitoring  data  to  average  soil  levels  for
conservative  contaminants  other   than  dioxin.     Data  were  collected  on
fluorides,  chromium,  copper,  manganese, nickel,  vanadium,  and  arsenic  for
this  approach (National  Research  Council,  1971;  1974a;  1973;  1975;  1974b;
Versar, 1976;  Sullivan, 1969;  and  Nriagu,  1979,  1980).   Atmospheric  values
from  rural  areas  were  used   to   minimize  the   influence  of  anthropogenic
materials emitted from  stacks.   Values for each  contaminant  were combined to
derive a conversion factor, CF  ,  which is  defined as the ratio of atmospheric
                              A
concentrations   (ug/m)   to   soil   levels  (ug/Kg).     Assuming  atmospheric
particulates  are derived from nearby  soils,  the predicted average contaminant
level  in  ambient air  at the  site can be obtained by multiplying  CF   by the
                                                                      A
average concentration  of the contaminant  in  soil.   Additionally,  CF   is an
                                                                      /\
estimate of particulate  levels  in air  (Kg/m  ).   The values for CFX determined
for each of.the seven contaminants are provided below:


Fluorides:  CFF =    °'02 to °'05 U9 F/m3 air                            (14)
              F   20,000 to 500,000 ug F/Kg soil

                = 3 x 10'6 to 4 x 10"8 Kg soil/m3  air
                 (Data from National Research Council, 1971)
                                     -17-

-------
Chromium:  CFr  =  °-01 U9 Cr/m  air                                     (15)
             O   37,000 ug Cr/Kg soil                                   v   '

                = 3 x 10"7 Kg soil/m3 air
                 (Data from National  Research Council,  1974a)
Copper:  CFr  • °-005 to °-05 US Cu/m  air                               (  .
  ™       Cu     20,000 ug Cu/Kg soil                                   •

              = 3 x 10'6 to 3 x 10'7 Kg soil/m3 air
                           (Data from Nriagu,  1979)

Manganese:  CFMn =   0.08 ug Mn/m3 air                                   (1 }
              Mn   800,000 ug Mn/Kg soil

                 = 1 x 10"7 Kg soil/m3 air
                 (Data from National Research Council, 1973)

Nickel:  CF,. = - 0.006 ug Ni/m3 air -                           (  }
           Nl   30,000 to 80,000 ug Ni/Kg soil                           v  '

              = 2 x 10'7 to 8 x 10'8 Kg soil/m3 air
         (Data from  National Research  Council   1975  and  Nriagu,  1980)

Vanadium:  CFV =   0.002 ug V/m3 air
             v   200,000 ug V/Kg soil                                    ^  '

               = 1 x 10"7 Kg soil/m3 air
                 (Data from  National  Research  Council, 1974b)
                                     -18-

-------
 Arsenic:   CF.C  =  °-001  U9  As/m  air                                       (   }
             As    5,000  ug  As/Kg soil


                =  2 x 10"7  Kg soil/m3  air

                  (Data  from Versar,  1976 and  Sullivan,  1969)
      In  general,  CFX values  are in the range of 1-3 x  10"7 Kg/m3.   This range

 is  comparable  to  100 ug/m  ,  a level  commonly  found  in  polluted air  and  in

 excess of  the Federal  Ambient Air Quality Standard of  75  ug/m3.   If  a typical

 surface  soil  density of  1600 Kg/m3  and soil  depth  of  1.0  cm  are  assumed,

 these CF¥  values  would convert to resuspension  factors  (RF) of  1  x 10   Kg/m
         A
 x 1.6 x 10"3 m3/Kg x 1  x 102 ttf1 =  2 x 10"8 nf1.   This value is consistent

 with  the work  from  which  the values of  RF  were derived;  in his  work,  Sehmel

 (1980)  determined that the  tracer had mixed to  a'depth of 1.0 cm.   The value
           O   1
-of  2  x  10    m   corresponds  to the RF  values  in Figure 5 for the  lowest wind

 or  activity stresses.  Hence,  the CFX  values derived  from metal concentration

 ratios  agree  with  empirical  data for particulates  in general.    Then,  as  a

 rule  of  thumb:     Cair   in  ug/m3  =  (1   x  10  ~7)  Csoji  in  ug/Kg.    This
                                                      Q     1
 relationship compares favorably with  the 4.6 x  10     Kg/m   value calculated

 from  the  Thibodeaux  (1983)  data  for  the  Vertac  site.   The  relationship

 represents  average  conditions over  a  year  rather   than  those  that  would

 prevail  during storm  events  or  in areas with  high soil disruption  activity

 levels.    The relationship  addresses  atmospheric levels  at  the  perimeter  of

 the source.   For downwind concentrations, the dispersion factor  developed  in

 Sections 3.3  can  be used,  or CFai-r  can be calculated  based on  soil  dioxin

 concentrations at the point  of exposure (measured or  calculated  using CF  .,

 from  Section 4.0).

                                     -19-

-------
     It  has  been  observed  that soil  erosion  is a function  of wind velocity
cubed   (Sehmel,   1980).     More  specifically,  erosion  is  proportional  to
W2(W -Wt)  (Gillette, 1973)  where  W  is  wind  velocity  and  Wt  is  a threshold
value of  6 - 13 m/sec.   For calculations  in  wind,  the median value  (9) for
the  threshold wind  velocity was  used and  the relationship  for determining
particulate concentration in air can  be described as follows:

           W < 10 m/sec:  C.   = 1 x  10"7 Cen<1                           (21)
                         CFair = I x 10'7 Kg/m3
W > 10 m/sec:  C    = (1 x 10"7) W2 (W - 9) C                  (22)
                          a1r                    -     soil
                        CFai> »  (1 x 10'7) W2  (W - 9) Kg/m3
where Cai-r = dioxin concentration in air above soil (ug/m  )
     C  .-, = dioxin concentration in soil at the point of  interest  (ug/Kg)
Because W is an  average  wind  speed  over time for chronic  exposure, the second
relationship will  not  be  required unless  acute  exposure  calculations  are
desired.
     Mechanical  disturbances  can increase dust  emissions  significantly.   For
dioxin levels that would  occur  during  episodes of mechanical disturbance, the
relevant  values  in  Figure   5   should  be   employed   to   increase  CF  .
                                                                           air
proportionally  to  the  ratio of the  relative  resuspension  factors  for  the
                                             Q    1
disturbed state  and the  calm  state  (RF = 10    m" ), respectively.  Hence, for
                                     -20-

-------
tractor  use  (RF  =  10'7   to  10"6  nf1),  CF .    would  be  1  to  2  orders  of
                                            <11 i
magnitude higher,  or CFair = 1 x 10"5 to 1 x 10"6 Kg/m3.
     This  approach   is  based   on   calculation   of  atmospheric   levels  as  a
function  of the  soil  at the  point of  interest  and  does  not  account  for
dispersion  downwind.   Thus,  for  downwind  areas,  C   •••  is  not  the original
concentration at  the source  but the  concentration  at the point  of exposure.
If a  measured value is  not available,  an  estimated  value  is  employed using
the  methods  described  in  Section  4.0.    This  approach  was  taken  because
atmospheric levels are believed  to arise  primarily  from resuspension of local
soils  (Sehmel,  1980).   This assumption further assumes that  activity levels
at the point  of  exposure.are greater than  or equal to  activity levels at the
source.   More complex  methods  for rapid assessment  of  particulate emissions
have been summarized by  MRI  (1984)  and  are  recommended if time and resources
allow.
     Sites  downwind  from the  original  source  will  reflect dioxin  levels in
soil  resulting  from  all transport mechanisms,  runoff and  wind.   Further,
atmospheric  levels  for chronic  exposure will  be  a  function of  average wind
conditions, i.e., wind speeds  of  about  5.5 m/sec  (12  mi/hr)  (Versar, 1983).
Most  wind-transported  dioxin  will have  arrived from discrete  storm events,
therefore, the conversion factor CFa.  for W >   10    m/sec   is    required   to
                                    air       ~
determine   downwind  soil  levels   as  described   in  Section   4.0.    These
atmospheric  deposition  contributions are  combined  with  runoff  input,  also
described  in  Section   4.0,   to yield   an   estimate  of  total   average  soil
concentration at  the downwind  point of  exposure.  The conversion  factor CF  .
                                                                           a i r
is then applied to calculate atmospheric levels, as follows:
                                     -21-

-------
          Cair ' ("air) ("=„„)
where C,,-,, = dioxin concentration  in air above soil (ug/m )
       a 1 1
     Csoj-| = dioxin concentration  in soil at the point of interest (ug/Kg)
3.3  Gaussian Dispersion
     The conversion  factor derivation  described in Section  3.2  applies only
to  the  air directly  above  a point  of  interest.   Atmospheric concentrations
downwind from  a source can  be  estimated without  knowing  soil concentrations
if  the  airborne particulate  plume  is  considered  'to  be subject  to Gaussian
dispersion.   In this case,  atmospheric concentrations, C  .  ,  at  point x can
be calculated using the following  formula (Turner, 1970):

                     Q2
      air(x,o,o) = TTO a u                                                ^
                     y zw
where Cai-r = concentration in air  (g/m3)
        Q2 =• emission rate (g/sec)
                                                                           r)
      ayaz = ProdlJCt   °^   tne    Gaussian   dispersion   coefficients   (m ),
             from Figure 4
         W = wind speed (m/sec)
Therefore,  the ratio of concentrations  at  two  points  can be determined by the
ratio:
          Cairl _ (gyqz)2                                                ....
          Cair2" l
                                     -22-

-------
If Ca-jr2  is  set  at  the  boundary of  the  source,  it can  be equated  to
Therefore,
           CF    (aya2
where x = distance from source boundary to the point of interest
      F = fetch or downwind dimension of the source

and       CX+F = CF     yz F                                             (27)
                     \y z'x+F
     The  approach  described  by  equations   24,   25,   26,   and  27  assumes
conservation of mass.
     Values for a a   downwind  from a  source  area  are  shown in  Figure  4 for
the following six stability classifications:
     A = Extremely Unstable
     B = Moderately Unstable
     C = Slightly Unstable
     D = Neutral,  considered  to  be  representative  of  average,  long-term
         conditions
     E = Slightly Stable
     F = Moderately Stable
          The  combined conversion  factors that  describe  particulate dioxin
concentrations downwind from a source are as follows:
                                     -23-

-------
If W < 10 m/sec:
CFair = CFparticulate = l x 10
If W > 10 m/sec:
                              -7
                                        'x+F
Kg/trT
(8)
                              "7
CFair ' ^articulate = l * 10"  w   
-------
1984;  Thibodeaiix,   1983),  with  ultimate photolytic  decomposition  (Nash  and
Beall, 1980).
     In  his  study  at  the  Vertac  site,  Thibodeaux  (1983)  calculated  that
vaporization  of  dioxin from the  soil  surface was  the  major route  of dioxin
loss from  the site.   Mass flux  calculations,  based on  estimated values for
pertinent environmental  and  chemical  properties,  predicted  that  vaporization
losses from the site  were much  greater than losses  from entrainment of soil
particles.
     Nash  and Beall  (1980)  reported  that  dioxin volatilized  from  soil  in
microecosystem  chambers and  from  field plots.    Significant   quantities  of
dioxin in  the air  from both experiments  appeared to be  dechlorinated.   The
researchers  concluded  that" atmospheric  photodegradation  was occurring.   The
rates  of   both  volatilization  and  degradation   depended   on  the  dioxin
application formulation and the temperature of the  systems.
     Freeman  and Schroy  (1984)  used vaporization processes to model dioxin
movement  in   a  soil  column.     However,  the   researchers   suggest  that
photodegradation at the  soil surface  will  dominate vaporization losses during
daylight hours.  Thus  dioxin losses to  the  atmosphere  should occur  primarily
at night, with rapid photodegradation the next morning.
     No  research has  been conducted  to date  on   atmospheric  degradation  of
dioxin.  Research  results  suggest that  dioxin  is  lost  from the soil, but the
loss mechanism  and  environmental fate  are  only poorly  understood.    Although
volatilization may  be  an  important  loss mechanism, potential photodegradation
may  reduce  any environmental   transport.    If  the  degradation  process  is
                                     -25-

-------
occurring,  as  postulated  by  Nash  and  Seal!  (1980) or  Freeman  and  Schroy
(1984), the potential off-site exposure to vapor will be very low.
     Due  to  the uncertainties  in volatilization and photodegradation,  it is
not   possible  to   derive  a   conversion   factor   for  dioxin   vapor   air
concentration.   As  more  research  is  conducted,  derivation  of  a conversion
factor for vapor (CFvapor) may  be possible.   This CFy   r  should be added to
the  CFa-r presented in  this  report  to  estimate  total   atmospheric  dioxin
concentrations at potential exposure points.
                                     -26-

-------
4.0  SOIL CONCENTRATIONS

4.1  Introduction
     Overland  transport  of contaminated  soil  via  runoff  is  an  important
mechanism which  contributes to  the potential  for  human exposure  to  dioxin.
Human activities  such as  farming,  gardening, excavating,  and  recreation can
result  in  dermal  absorption  of contaminants  or  ingestion of  contaminated
soil, particularly by.  children.    In  order  to estimate potential  exposures
downflow from  a  source,  an approximation of  the  soil  loss  from a  source and
the redeposit.ion of contaminated soil away from a source must be calculated.

4.2  Dioxin Behavior in Soil
     Soil at a source becomes  contaminated by  adsorption of dioxin.  No data
were  found  to  quantitatively  describe  dioxin  concentration  versus  soil
characteristics,  such  as  particle  size or organic  content.  However,  it is
assumed  that due to  its  high  KQW,  dioxin will  be adsorbed  primarily  on the
organic  fraction  of  the soil.    It  is this high affinity for organics in soil
and  low solubility  in  water that  are  believed to  account for  the vertical
immobility of  dioxin  (Kearney,  Woolson,  and Ellington, 1972;  Matsumura and
Benezet,  1973).    Because  small  particles  have  a higher surface-to-volume
ratio than  large particles, it  is  also assumed  that  the small  particle-size
fraction of  the  soil  would have  a higher contaminant concentration  than  a
bulk  soil  sample.   Walling  (1983) summarizes  the relationship  of particle
                                     -27-

-------
size  and  organic content  characteristics  of eroded  soil  to  those  of the
original   soil   in  five   test-plot   studies.    These   data  suggest  that
contaminants  such as nutrients or pesticides  may  be enriched up to 1.5  times
on clay-sized particles, and more  than  2 times on  the organic fraction.  Lack
of   quantitative  data   concerning   these   phenomena,   however,   precludes
incorporating them into the derivation  of conversion  factors.   It  has been
assumed,   therefore,  that   all   transported   soil  has   the   same  dioxin
concentration as  determined for bulk soil  samples  from the  site.

4.3  Photodegradation
     Photodegradation  is  another  process which  may  affect  the  amount  of
dioxin available  for transport from a  site and the amount to which humans may
be  exposed.   Ultraviolet  wavelengths  have been  shown to  be   effective  in
photodegrading  dioxins.   Photolysis apparently removes one  or  more  chlorine
atoms  from  the  dioxin  molecule,  thereby  making   it   less toxic  but not
destroying the  basic dioxin nucleus (Crosby et  al.,  1971).
     Esposito  et  al.,   (1980)  provide  a  comprehensive  review of  numerous
Photodegradation  studies and the  inconsistent  results.   Crosby et al.,  (1971)
applied  dioxin  to  several matrices.    Although  decomposition  was  rapid  in
alcohol  solution,  there  was negligible  loss  from aqueous  suspension  and wet
or dry  soil  after 96  hr of irradiation.   However,  the researchers suggested
that in  the natural  environment,  waxy  leaf cuticles, surface slicks on water,
and  spray oils  or solvents commonly  incorporated  in  pesticide formulations
may serve as the organic hydrogen donors necessary for Photodegradation.
     Dioxin applied  to  soil  and exposed to artificial  sunlight  (sunlamp) for
96  hr  showed   no  degradation,  as reported  by  Crosby,   Moilanen,  and  Wong
                                    -28-

-------
(1973).    In  other  studies   (Crosby   and  Wong,  1977),  dioxin-contaminated
Herbicide  Orange  was  applied  to   plant   leaves  and  soil  and  exposed  to
sunlight.  After  6 hr of exposure,  0-30% of  the  dioxin  remained on the plant
leaves,  with  30%  remaining  on  soil  which  had  received  the  lowest  applied
                           n                                                  O
concentration  (1.3   ng/cnr).     At   the   application  rate   of   10  mg/cm
approximately  90%  of  the  dioxin  remained  after  6  hr  of exposure.    The
researchers  believed  that  surface  soil   particles  shaded  the  underlying
particles, thereby  preventing  photodecomposition  at  depth.   It was concluded
from   the    1977   study   that    the   three    requirements    for   dioxin
decomposition/dechlorination are:   1)  dissolution  in light-transmitting film;
2) the presence  of an  organic hydrogen donor, such  as  solvent or pesticide;
and  3)  ultraviolet  light.   All  three  conditions  should  be present  in  the
application or accidental  loss of materials commonly contaminated with dioxin
                                             f
(2,4,5-T,  trichlorophenol,  PCB  road oils).   Crosby  and  Wong  conclude  that
dioxin is  not stable  as  a  contaminant in thin  herbicide films  exposed  to
outdoor light.
     In  response  to the work  by Crosby and  co-workers,  photodegradation  was
evaluated  as  a decontamination  technique   in  Seveso,  Italy  (Liberti  et  al.,
1978).   Exposure  of  dioxin-contaminated soils to  artificial ultraviolet light
and  natural   sunlight  in  the  presence of   a  hydrogen   donor  resulted  in
degradation at the surface and  to  a certain  extent,  degradation beneath the
soil  surface.   The degradation  rate in soil  from  natural  sunlight  would  be
affected    by   sunlight   intensity,  nature   of   the   contaminated   medium,
temperature,  and the amount of vegetative cover at a site.
                                     -29-

-------
      Although photodegradation may have  a  significant  effect on environmental
 dioxin  concentrations,   it   was   not  included  in  deriving  soil  conversion
 factors.   The  amount  of  degradation   appears   to  depend  on  site-specific
 factors  (sunlight  intensity,  temperature,  substrate,  cq-contaminants)  that
.are not  amenable  to  a  generic approach.   If photodegradation  is  occurring,
 the  method   presented   in   this  report  will   overestimate  the   dioxin
 concentrations at  the point  of exposure.
 4.4  Volatilization
      As  discussed   in  Chapter  3.0,  volatilization  from  a  site  may  be  a
 significant  loss  mechanism.   Nash  and  Beall  (1980)  and  fhibodeaux  (1983)
 report that volatilization may be a major  pathway.   Matsumura and Ward (1976)
 indicate that  the  water  content  of  soil  may mediate  the   evaporation  rate.
 The effects  of volatilization on  concentrations  at a  site and  at points  of
 exposure have  not  been considered  in deriving  CFsoii,  so  overestimation  of
 soil  exposure concentrations may result when using this method.
 4.5  Approach
      Empirical and deterministic  methods were evaluated for  applicability  in
 deriving conversion  factors.   The   empirical  approach  involved  collecting.
 monitoring   data  from  specific  sites  and  trying   to  correlate  observed
 distributions with characteristic  site  parameters.   Because it  was  expected
 that  few comprehensive sets  of dioxin  data would  be available,  monitoring
 data  for  other  persistent   contaminants  such  as  polychlorinated  biphenyls
 (PCBs), polybrominated biphenyls  (PBBs), heavy metals,  and  radionuclides  were
 also  sought.
                                     -30-

-------
     Only two sites  were  found where sampling had  been  conducted at and away
from a site.   Roberts, Cherry,  and  Schwartz (1982) studied  the distribution
and  surface  translocation   of  a   serious  PCB  spill  at   a  transformer
manufacturing plant  in Regina,  Saskatchewan, Canada.   The  researchers  found
that  particle  transport   in  runoff from   eroding  areas  was  an . important
migration   mechanism.      However,   PCB   distributions   were   "extremely
heterogeneous," with "no definable trends in concentrations."
     Dioxin  contamination  in  Seveso,  Italy,  was  caused by  wind-influenced
atmospheric  deposition  from  the  1976  explosion  at  the  Givaudan-LaRouche
ICMESA plant.   Sampling was  conducted  within 110  hectares  southeast  of the
site for  over 3  years (DiOomenico  et  al.,  1980).   Dioxin  concentrations at
locations 100 m  apart varied by  as  much  as  a factor of  100,  and this highly
irregular  distribution  changed  very  little  during  the  three-year  study
period.   Although  the mechanism  by which  dioxin  was  initially distributed
differs from that characteristic  of uncontrolled disposal  sites,  the method
of transport from  the  originally contaminated area is similar  to the problem
addressed in this report.  Of particular interest are the following:
•    Areas of high  contamination showed little  dioxin contamination reduction
     over three years.
•    Slightly contaminated or  uncontaminated areas  downwind  and within runoff
     routes   showed   no   statistically-significant   increase   in   dioxin
     concentrations over three years.
     Assessment  activities  at  dioxin-contaminated  sites  in  Missouri  and
Arkansas  did not  include  systematic  sampling/analysis   at  the  sites  and at
intervals  away  from  the  sites.   Sampling  was   not  conducted   at  known  high
                                     -31-

-------
concentration  sources  scheduled for remedial action  at  the Vertac site  (JRB,
Inc., 1983).   Sampling was conducted at  numerous  sites  in Missouri, but only
to  locate  areas of  high  concentration.  Few samples were acquired from each
site  and no  descriptions of  site  or  pathway characteristics  were provided
(U.S. EPA, 1982a; U.S.  EPA, 1982b).
     Based on  available site  data, it was not possible to  derive correlations
between  dioxin concentrations  at  sources  and  at  points  of  exposure.   The
Regina  and  Seveso  data  provide  qualitative   indications that  contaminant
concentration  distributions will  probably be irregular  and thus difficult to
predict.   Belli et  al.,  (1983) report that the  statistical  analysis of data
from regions of low  contamination  at Seveso was most strongly affected by the
sensitivity and precision of analytical instrumentation.
     Deterministic  approaches   involve  mathematical  modeling of  the physical
transport process from a  site  to  a selected  exposure point.  Onishi, Whelan,
and Skaggs  (1982)  present a  review of overland  soil and  sediment transport
models and divide  them into three  groups based on their degree of complexity
and the extent to which they represent  physical processes.
     The simplest models  require  the  least amount of site-specific  data and
use an  empirical  formula  to  estimate  average  soil   losses  from an  area.   A
modified version  of  the  Wischmeier  and  Smith  (1978)  USLE  requires limited
data on watershed characteristics.
     The second group  of models   requires  considerable amounts  of detailed
hydrologic,  meteorologic,  and  site-specific   physical  characteristics  to
simulate soil  erosion  and transport.    If  the required data  are available,
these models are generally more accurate  than  the simplest, empirical models.
                                     -32-

-------
Both the first and second model  groups  account for chemical distributions via
loading factors.
     The  final  group  of models simulates  environmental   chemical  behavior,
such as  adsorption-desorption  and decay, as well  as  runoff and  erosion.   In
addition  to  the  data  required  for  Group 2 models,  the most  complex models
require chemical  characteristics and distributions on the land surface.
     The  last  two groups of models  are useful only to  those  who have access
to a digital  computer,  and  are  therefore of  no interest to those  wishing to
calculate  simple  conversion   factors.    In  addition,  these  models  require
detailed  site and  chemical data that  are generally  not  available  without
extensive field investigations.
     The model used  to derive  conversion factors from  average  soil  loss and
deposition is consistent with  the first group of models  and does not require
    •>
a  computer or  cumbersome  amounts  of  site-specific  data.    Because  chronic
exposure  is of primary interest, average loss  and  deposition  are appropriate
for assessing lifetime exposure, rather than  for cyclic or acute events.  The
                                                           »
approach utilizes the USLE.

4.6  Average Soil Losses
     Average  annual  soil  losses from a  contaminated  site  can  be approximated
by using the  USLE, .an  empirical formula  which  was  developed for agricultural
land using data  from  numerous  field test  soil  plots  (Wischmeier  and Smith,
1978).    The  equation  input factors  have been  modified  slightly for  use in
areas  other  thjan cropland.  The USLE  provides  an  approximation of sheet and
rill erosion  losses, in tons per acre per year,  due to the interaction of six
physical  factors  which can be  expressed numerically  as site characteristics.
                                     -33-

-------
Tables  and  maps  are  provided for  use in  selecting  site-specific values for
these factors.
     The USLE defines loss as:

          L=RxKxLSxCxP                                          (28)

where L = computed average annual soil loss  (tons/acre/yr)
      R = rainfall erosion index  (yr"1)
      K = soil credibility factor (tons/acre per unit of rainfall  factor,
       .   R)  .       .
     LS = topographic factor  (dimensionless)
      C = cover and management factor  (dimensionless)
      P = support practice factor (dimensionless)
Useful procedures for the  estimation  of  USLE parameters for both  agricultural
and non-agricultural conditions can be found in Mills et al., (1982).
     The average  annual soil  loss  per  unit area,  L-,  represents  an average
annual value  and  is obtained  by  multiplying the rainfall  erosion  index (which
provides estimated  soil losses due  to rainfall  and  runoff for  a geographic
area)  by  a  series  of  ratios.    These ratios  represent  the  relationship of
actual  parameters   to   those  observed   in  test   soils   and   standardized
agricultural plots.
     The rainfall  erosion  index, R,  expresses erosion potential  for average
annual rainfall at a location.  A map  of  average  R values  for the U.S., based
on  over 30  yrs  of  measurements,  is provided  in  Figure  6.   Interpolation
between contour lines is necessary for many areas of the country.
                                     -34-

-------
e. 5
 °
O 3
  I
  CO
                       W.H. WlICIWMiw. SEA. 1976
                                                                                                ,t>
                         Figure 6.  Average Annual Values  of the •Rainfall  Erosion
                                    Index  (yr~l) (from Wischmeier and Smith,  1978)

-------
     Values  for  K,  the  soil  erodibility  factor,  have  been experimentally
determined  for a number  of  benchmark  soils  at erosion  research  stations in
the  U.S.   Average values of  K,  based on a range  of  soil types,  are provided
in  Tables  2   and  3.    The  soil  erodibility  for  a  particular  site  can be
approximated by  using the K  value corresponding to the predominant soil type.
Average values for basic soil types  are provided in Table  2.   Assuming  soil
organic content  is  known or  can  be  estimated,  more specific values for K are
available in Table  3.
     The  topographic  factor, LS,  combines  the  effect  of  slope  length  and
steepness.  Values  for the  area  under consideration can be determined using
the  average percent slope and slope  length, measured  in  ft.  A.listing of LS
values for  slopes of  varying gradients and lengths  is  provided   in  Table 4.
Interpolation  between  listed  values may be necessary.
     The cover and  management factor, C, is  most significant for agricultural
land  where  it  is  a" function   of  vegetative  cover,   crop  sequence,   crop
rotation,  and  tilling   practices.    Wischmeier  and  Smith  (1978)  provide
guidelines  for determining  C values for construction  sites,  pasture,  range,
idle land,  and forested  areas.   In order to  simplify  site characterization,
two  C  values  have  been  selected.  A  C value  of  1.0  represents  a worst-case
scenario and should be used  when vegetation  is completely  absent.   Examples
of this  type   of  site would  be  horse arenas,  unpaved  roads,  and unvegetated
landfills.   A C  value of 0.5 should  be  used  for any  other type  of site.
Because 0.5 represents  a  high value for permanent pasture, range, wooded, and
idle land,  a  worst-case  scenario for vegetated  land has been assumed.   For
wooded areas with highly  erodible soil  and  no  surface vegetative  cover, the C
                                     -36-

-------
Table 2.  Average Values for the Soil  Erodibility Factor,
          K, for Soils on Erosion Research Stations
          (After Wischmeier and Smith,  1978)
               Average          K Value
              Soil Type       (tons/acre)

           Silt Loam
           Loam                   0.4
           Sandy Clay Loam

           Silty Clay Loam
           Clay                   0.3
           Clay Loam

           Fine Sandy Loam        0.2

           Loamy Sand
           Flaggy Silt Loam       0.1

           Gravelly Loam         <0.1
                        -37r

-------
        Table 3.  General Magnitude of the Soil/Erodibility Factor,
                  K*, when Organic Content Data are Available
                  (Carsel et al., 1984)
                                    Organic Matter Content
Texture Class
Sand
Fine Sand
Very Fine Sand
Loamy Sand
Loamy Fine Sand
Loamy Very Fine Sand
Sandy Loam
Fine Sandy Loam
Very Fine Sandy Loam
Loam
Silt Loam
Silt .
Sandy Clay Loam
Clay Loam
Silty Clay Loam
Sandy Clay
Silty Clay
Clay
<0.5%
0.05
.16
.42
.12
.24
.44
.27
.35
.47
.38
.48
.60
.27
.28
.37
.14
.25

2%
0.03
.14
.36
.10
.20
.38
.24
.30
.41
.34
.42
.52
.25
.25
.32
.13
.23
0.13-0.29
4%
0.02
.10
.28
' .08
.16
.30
.19
.24
.33
.29
.33
.42
.21
.21
.26
.12
.19

*The values" shown are estimated averages of broad ranges of specifiw-soil
 values.  When a texture is near the borderline of two texture classes, use
 the average of the two K values.  For specific soils, Soil Conservation
 Service K-value tables will provide much greater accuracy.
                                 -38-

-------
  Table  4.   Values  of  the  Topographic Factor, LS,  for  Specific
               Combinations of Slope  Length  and Steepness (From
               Wischmeier and Smith,  1978)
Nrctn'
t-apt
0.2 ...
• 0,5
0.8
2
3 ....
4
3 	
6
8 ...
10
12 	
14
16 	
18
20
Slop* lingth ;f»»t:
23
. . 0.060
.073
.086
	 133
	 190
230
.268
. .336
496
685
	 903
.... 1.15
. ... 1.42
... 1 .72
. . . . 2.04
50
0.069
.083
.098
.163
.233
.303
.379
.476
.701
.968
1.28
I.i2
2.01
2.43
2 38
73
0.075
.090
.107
.185
.264
.357
.464
.583
859
119
1.56
1.99
2.46
2.97
353
100
0.380
.096.
.113
^201
.287
.400
.536
.673
.992
1.37
1.80
2.30
2.84
343
4.08
130
0.086
104
.123
.227
.325
'.471
.656
.824
1.21
1 68
2.21
2.81
3*8
4.21
5.00
200
0.092
.110
.130
.248
.354
.528
.758
.952
1.41
1.94
2.S5
3.25
4.01
3.86
5.77
300
0.099
.119
.141
.280
.400
.621
.928
1.17
1.72
2.37
3.13
3.98
4.92
5.95
7.07
400
0.105
.126
.149
.305
.437
.697
1.07
1.35
1.98
2.74
3.61
4.59
5.68
687
8.16
300
0.110
.132
.156
.326
.466
.762
1.20
1 50
2.22
3.06
4.04
5.13
6.35
7.68
9.12
600
0.114
.137
.162
.344
.492
.820
1.31
1.65
2.43
3.36
4.42
5.62
6.95
8.41
10.0
MO
O.!21
.143
.171
.376
.536
.920
1.52
1.90
2.81
3.37
5.11
6.49
8.03
9.71
11.5
1.000
0.126
.152
179
.402
.573
1 .01
1.69
2.13
3.14
4.33
5.71
7.26
8.98
10.9
12.9
  : LS = ;V 72.6 ™ '65.41 »in: 6 4- 4.56 tin 8 - 0.065) whtr*  X = slop* l«ngrh in f**t: m = 02 far

 grod.tnn < 1  percent, 0.3 (or 1 to 3 p«rc«nt ilopn, 0.4 for 3.5 to 4.5 p*r:tnl tlopci, 0.5 for 5 pxrcint

 ilopti and it»«p«r, and 6 = ongl* of llop«. (For oth«r combinationi of length and gradient, inttrpolatt

 b«fw««n adjactnt valu« or M< fig. 4.1
Reproduced  from    jjpyt
best  available copy.
                                   -39-

-------
value  would  approach  1.0.   Estimation  of  soil  losses  using the  C  values
discussed  here  would probably result  in  higher rates  than  would  be  actually
observed in the environment, but are deemed to  be  acceptable  approximations.
     The  support  practice  factor,  P,  is  also  dependent   on   agricultural
techniques  and  is a function  of such  practices  as  contouring and terracing.
Because there is  no counterpart to  P  on natural  land  or construction  sites,
the value of  P has been set at 1.
     Users  of the USLE  must be  aware of  its  limitations (Wischmeier,  1976;
Walters, 1983).  Soil losses from  a  source  or area can be determined  by  using
the USLE;  however, the USLE  provides  only an  estimate of the amount of  soil
eroded  from  a  specific area  and does  not  indicate  the amount  of  sediment
actually delivered to  streams.    The  sediment yield  is  the  total  amount of
soil loss from the area less the amount of deposition which occurs.
     R  values  obtained  from  Figure   6  (Wischmeier   and Smith,  1978)  are
applicable  only for  long-term erosion  averages.    Values for K,  the  soil-
erodibility  factor {Tables  2  and  3),  are  averages for  soil  types,  but the
actual amount of soil loss for any  soil  type  can  vary  widely as a function of
antecedent  soil   moisture   conditions.    The  amount   of   runoff  will  be
significantly different for saturated  and unsaturated soils.
     The USLE was developed  primarily from  data  obtained east of the  Rocky
Mountains,   so its applicability to  the arid  western  states  may  be  somewhat
limited.    Use  of  the  USLE  may  result  in   significant errors  due  to the
predominance  of high intensity,  short  duration rainfall  in  the West, and the
greater effect of other physical conditions such as wind, humidity, and heat.
                                     -40-

-------
     Because  the  USLE  was developed  using data  from small field  plots, it
predicts  sediment yields  of  particles  of  1-mm  diameter  (coarse  sand)  and
finer  sediments.   The  USLE  is  not applicable  to coarser  sand,  gravel,  and
larger particles.

4.7  Soil Deposition
     A  Modified  Universal   Soil  Loss  Equation  (MUSLE)   (Simons,  Li,  and
Associates,  1982;  Walters, 1983)  has  been developed  for  determining single-
.storm  event  sediment  yields  from drainage  basins.    The  substitution  of  a
runoff factor for  the rainfall  factor, R,  in the  USLE makes the MUSLE better
suited  for  use  in areas  west of  the Rocky  Mountains.    Use  of  the  MUSLE,
however,  requires  calculating site-specific coefficients,  which  preclude its
general  use  for  determining  sediment yields.    The  MUSLE has  been further
modified  for computing  annual  sediment yield,  but  this calculation is  also
site-specific based on weighted storm  yields for selected return periods.
     To determine  sediment yield  to a stream,  the sediment delivery ratio is
used (Piest and Miller,  1975):

          D = Y/L                                                        (29)

where     D = sediment delivery ratio, the change  per  unit  area of sediment
              delivery downstream  (dimensionless)
          Y = sediment yield  at measuring  point  (tons/acre/yr)
          L = total amount of sediment eroded from drainage area .-"stream
              of measuring point,  estimated using  the  USLE   (tons/acre/yr).
                                     -41-

-------
     To  date  there  have  been  no  comprehensive  studies  defining sediment
delivery relationships for  the U.S.  on  a regional  basis.   It is impossible to
define relationships  which  would  hold true for all geographic areas.  Walling
(1983) states  that the  processes  of sediment  delivery are  very  complex and
are  dependent  on  a  variety  of  factors,  including topography  of  the  source
area,  stream channel  characteristics,   drainage  patterns,  vegetative  cover,
land  use,   soil   properties,   and  the  distribution  of  sediment  sources.
Interrelationships  between  these  factors are difficult to  define,  and errors
can  be  introduced  because measured sediment  yields  are  compared  to  total
erosion from a source estimated with  a  generalized  soil loss  equation.
     As summarized by  Walling (1983),  there is  evidence  that only a small
percentage of the  drainage  basin  area provides storm runoff  in  humid regions,
and  the  actual  runoff  area for the  same delivery  location  varies  in  extent
and  location depending  on  antecedent   moisture  conditions.    This evidence
suggests  that  the  sediment  delivery  ratio  is  dependent   on   only  the
characteristics Of that  portion  of the  drainage  basin which  produces  storm
runoff and would change with time as the area changes.
     Some relationships  have  been  characterized sufficiently to show general
trends  between  the  size  of  the  drainage  basin   and  the  sediment delivery
ratio.   Piest  and  Miller  (1975)  present  a  summary  of this  relationship in
Figure 7.   Walling (1983)  provides a  curve  showing  the  relationship between
the  sediment delivery  ratio  and  drainage  basin   area  for  the central  and
eastern  U.S.,   as  developed   by   the  U.S.   Department  of   Agriculture  Soil
Conservation Service.   He also  gives  a summary  of  10  relationships  from
selected drainage basins  in the U.S.  and other countries  as well as a summary
                                     -42-

-------
£E
Cr .
r
r M
T
  , HEOmLLS PM»S*OWWHC ARtA TElAb AWO OKLAHOMA
  • WSSOvOi 8ASM LOESS r«.L5 • IO«A AX) FCBAASKA
  » eLACKLANOMAIRCS. TEXAS

  ^ SANL- CtAV NLLS -MSSi^^Pi3!

  "•>OUTH":.AI "•""•""i-        " r'"'v "

  i SMMOFiLDPLAn.LuNOlSINOT USED KOt TEHMK*TON OF O«v£)-

      I   I
 Figure  7.   Generalized  Relationship Between  Size  of
               Drainage Basin  and  Sediment Delivery  Ratio
               (from  Piest  and Miller, 1975)
Reproduced  from
best  available  copy.
                          -43-

-------
of  13 sediment  yield/drainage  basin area  relationships which  are directly
analogous  to  the delivery  ratio/drainage  basin area  relationship  (Piest and
Miller,  1975;  Walling,  1983).    Some  of  these  relationships  show  marked
similarities to the relationship in Figure 7.
     The  relationship  shown  in  Figure  7  is  generally  applicable  to  the
central and eastern  U.S.   No  comparable  data  could be found  for the  western
U.S.   The  sediment  delivery ratios shown  in Figure 7 vary widely for  a given
drainage basin  area.   The values for  basins often vary by  a  factor  of 2, and
sometimes by an order of magnitude.
     It is assumed  that sediment delivery ratio  is  inversely proportional to
drainage  basin  size  because  of  greater  redeposition  that  will   occur as
sediment  travels  over  greater  distances  before  reaching  the  point  of
interest.  It  can be  concluded  by analogy that  redeposition  of contaminated
soil will  become smaller  as  locations of interest  are more  distant from the
source site.
     The concentration of dioxin in soil at a point of interest,  x,  is:
                  mass of dioxin delivered to x
                   mass of soil delivered to x
        (mass of dioxin lost from source)(fraction delivered to x)       /,,N
                        mass of soil delivered to x
     The  fraction  delivered  from  the  source  to  point  x  is  the  sediment
delivery  ratio  for  an  assumed  watershed  which  begins  at  the   source  and
encompasses the  natural  drainage  area between  the  source  and  the  point of
interest.   The mass of soil  lost  from the source can  be  estimated using the
                                     -44-

-------
USLE and the source  area.   The total  mass of soil delivered to point x can be
estimated using, the  USLE, the  watershed  area,  and the sediment delivery ratio
for the  entire  watershed area.   Frequently, USLE estimates will  be based on
the average  slope and  slope  length  for the watershed.   The  expression  for
soil concentration, thus, becomes:

          r        source x Lsource x Asource x Psource                  ,,0x
          Soil  =	[	T~A	75	                  (  '
                         Lbasin x Abasin x ubasin
where   Ccnny,ra = dioxin concentration in soil at the source (ug/Kg)
         SOUiCc
        Lsource 3 estimated   soil   loss   from   the   USLE  for   the  source'
                  (tons/acre/yr)
        Asource = source area  (acres)
        "source = sediment  delivery  ratio  for  the   area  between  the  source
                  and the point of interest  (dimensionless)
         Lbasin = estimated  soil   1°ss   ^rom   tne  USLE  for  the  watershed
                  upstream of  the point  of interest (tons/acre/yr)
         Abasin = watershed area upstream of the point of  interest  (acres)
         Dbasin = sec'''ment delivery ratio  for  the watershed  area upstream of
                  the point of interest  (dimensionless)
     From  Figure  7,  it appears  that  the  sediment  delivery  ratio  for   a
specific  basin  area  can range  from  about  0.1  or 0.2  to  1.0,  depending on
factors other than  basin area.   On the other  hand,  the  delivery  ratio  as  a
function of basin  size only varies over about  the same range.  Therefore, in
general, the sediment  delivery ratio  for the source  and watershed basin areas
are  not  considered  to  be  significantly different.    If  sediment delivery
                                     -45-

-------
ratios  at  the  point of  interest are  not known,  C  •-,  can  nevertheless  be
estimated using a simplified form of equation 45:
          P     - P
           soil "  source
                          Lsource x  source
                            basin x Abasin
(9)
     In areas  where  the downgradient  points  of exposure  lie  in  the downwind
direction  from the  site,  it  will  be necessary  to  consider the  effect of
atmospheric  concentrations, as  well  as  soil   concentrations.    Atmospheric
concentrations  and  the  corresponding conversion  factor  were   discussed in
Chapter 3.0.

4.8  Wind Deposition
     The constructs  for calculating  Cair  presented in Chapter 3.0 incorporate
the assumption  that the  soil  concentration of  dioxin (Csoil)  is  known at the
location  of  interest.     If  observed  data  are  not  available,   a  means of
calculating  downwind   soil  concentrations  arising  from  deposition  during
previous storm events will be necessary.
     For the purposes  of this  report,  it  is  assumed that dioxin-contaminated
soils  in  downwind areas arise  from windblown  particulates.   Although   some
vapor   transport   may   occur,   this  process   was   neglected   due   to   the
uncertainties  discussed  in  Section  3.4.    If   the  transported   particles of
interest  are  in  the  range £  20 micron,  they  are subject  to   dilution as
predicted  by  a  Gaussian  distribution  for  the  plume.  Because  respirable
particles  are  £ 10 micron,  this assumption incorporates  all  particulates of
interest.    Larger particles containing dioxin  will  settle more rapidly and
therefore reduce  atmospheric dioxin  levels.  Hence,  this  assumption may  lead
                                     -46-

-------
to overprediction.   In this  case,  atmospheric  concentrations,  Cai>,  at point
x can be calculated using the following  formula  (Turner, 1970):

                          Qz
          Cair(x,o,o) *~^-^                                          (24)
                          yV
where     Qz = emission  rate  (gm/sec)
        o" a  = Product of the Gaussian dispersion coefficients  (m ),
               from Figure 4
           W = wind speed (m/sec)
Therefore, the ratio of concentrations at two points can be determined by the
ratio:

          ^M = lVil2            •                                   (25)
          Lair2   lCTyVl
     If the  phenomena that relate atmospheric  levels  to  soil concentrations
at a  given  location are essentially the same for  all  points  in the downwind
direction (i.e., CFair = Ca1r/Cso1l),  it holds that

          Csoill   (avaz)2
          c	= /   V                                              (33)
          Lsoil2   (yz)i

Values  for  a az  downwind  from a  source  are given  in  Figure 4 for  the  six
stability classifications.   Assuming  that most  particulate  transport arises
from  major  storm  events,  the  a az  values  for Stability  Class  A  are  most
appropriate for  predicting downwind  soil  concentrations.  Therefore, downwind
soil  dioxin  levels at  distance  x  can  be calculated  using  the  following
relationship:
                                    -47-

-------
          CF
               ..
            5011   Source   «VzW
where     F = fetch or downwind dimension of the source (m)
          x = distance from source boundary to the point of interest (m)
                                                                p
       ova, = product of the Gaussian dispersion coefficients (m ),
        A • £                   •                    .
              from Figure 4.
     When runoff  patterns  and prevailing wind direction  are  coincident, soil
concentrations should be based  on the summation of  the  two contributions,  as
follows:
                    Csoil  _ Lsource x Asource
            5011 " Csource "  Lbasin x Abasin
     This surface  concentration value relates  to the top  centimeter  of soil
and should  be  used for subsequent  calculation  of downwind dioxin-particulate
levels  in  the  atmosphere,  assuming  that  no  soil  mixing  occurs  prior  to
resuspension.
                                     -48-

-------
5.0  SEDIMENT CONCENTRATIONS

5.1  Introduction
     Human  health  risks  may  arise  when  individuals  are  exposed  to  dioxin
through the consumption  of dioxin-contaminated fish.  Fish  accumulate  dioxin
from an  aquatic environment  in  two ways  (Isensee  and Jones,  1975).   Bottom
feeding species,  in  particular,  may  ingest  contaminated sediment  along with
their food.   Any fish species can accumulate  dioxin directly from the water.
In these  cases, dioxin  is desorbed  from  contaminated sediments  or absorbed
from stream water.   To quantify human exposure,  it  is necessary to develop a
means of  approximating the  concentrations of  dioxin  in stream  sediments  in
the vicinity of  a contaminated site.   The  conversion factor described in this
report can  be  used  in  the algorithm developed by  Schaum  (1984)  to estimate
the  bioconcentration of  dioxin  in  various  fish species  and  the subsequent
human exposure.

5.2  Dioxin Behavior in Water
     Some of the processes which can affect dioxin  when  it is exposed to air
(as discussed  in Chapters 3.0 and 4.0)  are  expected  to  have minimal effects
on  dioxin  in  an  aquatic  environment.    Crosby et  al.  (1971)  report that
volatilization  does  not  appear  to be of  major  importance  in water.   Other
researchers report  that evaporation  from  or  with water may  be a major cause
of  the  disappearance  of dioxin  in. a model aquatic  ecosystem  (Ward  and
                                     -49-

-------
Matsumura,  1978),  but the experimental  results and their  application to the
natural  environment  are  as yet  inconclusive.   Photodegradation is thought to
occur  so slowly  as  to  be negligible in  water (Crosby, Moilanen,  and Wong,
1973;  Crosby  et  al.,  1971;   Isensee  and  Jones,  1975;  Matsumura  and Ward,
1976).   Some evidence of  microbial  degradation under  experimental conditions
has been  documented;  however,  dioxin in water  is  generally thought to resist
microbial effects  (Ward  and  Matsumura, 1978; Matsumura  and Ward, 1976).  The
half-life of dioxin  was found  to  be on  the  order of 600  days in  a model
aquatic ecosystem (Ward  and Matsumura, 1978).
     When present at  very  low  concentrations on sediment, dioxin  is generally
not expected to desorb  due to its low solubility.  At  concentrations as low
as 0.1 ppb,  however, dioxin can  desorb.   Isensee and Jones (1975) report that
under  experimental   conditions  the   concentrations  of  dioxin  in  water  and
sediment  reached  equilibrium  in 4 to  15  days.   The  temporal  variation was
attributed to the difference in  adsorption capacities of the two  soils  used.

5.3  Sediment Transport  in Streams
     Contaminants can be  transported in  streams  by  three processes:  1)  as
dissolved compounds  in  stream  water, 2) as  compounds  adsorbed onto sediments
and  transported  as  suspended  load,  and  3.)  as  compounds  adsorbed  onto
sediments and  transported  as bed load.   The  low solubility and high  affinity
of dioxin for  soils,  particularly soils high  in organic content  (Isensee and
Jones,  1975;  Kearney,  Woolson,  and  Ellington,  1972),  suggest  that dioxin
would be  transported  primarily  in ;? adsorbed  phase on  stream sediments.  No
data could  be  found  on  preferential  adsorption of  dioxin to  any particular
sediment  particle size,  so  it was  assumed  that  dioxin  would  be  adsorbed
                                     -50-

-------
equally on  all  available particle  size  fractions.   Pritchard  (1984) reports
that  for  polynuclear   aromatic  hydrocarbons,   partition  coefficients  and
natural  transport   of  sediments  adequately   accounted   for   the  observed
distribution of the contaminants in an aquatic environment.
     The  source  of  stream  sediment contamination   in   the  vicinity  of  a
contaminated  site  is  surface  soil   on   which   dioxin  is  adsorbed.    Such
contaminated soil may  reach  a  stream  by  direct stream erosion of the soil or,
more commonly,  by  overland  sediment  transport.   The  latter  process  is known
as  sheet  and rill  erosion  and occurs during  runoff  of precipitation.   This
type of sediment transport  was discussed  in Chapter 4.0.   Of  the soil eroded
and  transported by  overland  processes,  some  can  be expected to  reach both
major and minor streams within a drainage system.
     Finer soil  particles,  such as  clay,  silt,  and fine  sand,  that reach the
stream  are  usually  transported  as  part  of  the  suspended  sediment  load.
Coarser particles,  such  as  coarse sand and gravel, are usually transported as
bed sediment load or are deposited  in the stream  bed.
     Most  streams  normally  flow  at  less than  their capacity.    This normal
flow is called  the  mean  annual discharge  and corresponds  to  a  water depth of
only about one-third  of  the stream capacity or  bank-full  depth.   Mean annual
discharge  is  equaled  or  exceeded on  an  average  of 25% of  the  time  (Leopold,
Wolman, and Miller, 1964).
     For   streams   in  general,   the   amount  of  suspended  sediment  varies
logarithmically  with  respect  to  stream   discharge.    As  discharge increases,
streams can  also transport  larger  sediments.    Correspondingly,  the majority
of  sediment  transport in a  given  stream occurs during high  flow conditions
                                     -51-

-------
and  seasonal  floods,  rather  than  during  very  low  or  even  normal  flow
conditions.   During  the higher flow  events,  particles previously transported
as  bed  sediment load  and  sediments deposited in the  stream  bed  may be added
to  the  suspended  sediment  load.    Extremely high  flow events  may transport
very  large  quantities of sediments,  but are so  infrequent and  of  such short
duration  that  their  effect  on  the  average  sediment  discharge  is  minimal
(Longwell, Flint,  and  Sanders, 1969).
     Studies  of  PCB  concentrations  and  transport in  the Hudson  River in New
York  State  (Turk,  1980)   show  a  constant  transport  rate  of  PCBs  during
moderate  non-flood  discharges  and  increased  transport  during  floods.   PCB
concentrations were  found to  increase as discharge  fell below an intermediate
value.   At  low  discharge,  resuspension  of bottom material  was  minimal,  but
less  dilution of  contaminants occurred.   At  higher  discharges,  increased
concentrations  were   due  to  resuspension  of  contaminated  bottom  sediments.
During  intermediate   discharges,    PCB  concentrations  were  found  to  be  a
function of  both sediment resuspension and  dilution.   However,  the net effect
of these opposing  influences  produces concentrations  less than  those achieved
during either low or high stream discharges.
     Deterministic   and   empirical   approaches   were  examined   for   their
usefulness in deriving exposure factors  for stream  transported  sediments.   In
general,  deterministic approaches  involve  the   use  of  numerical  models  to
approximate  natural  physical  processes.    Onishi,  Whelan,  and  Skaggs  (1982)
provide a comprehensive  review of  a  number  of models which could  be  used to
simulate sediment  transport by streams.  These models are  divided  into three
groups, as  described in Chapter  4.O..  The  least complex models require  the
                                     -52-

-------
least amount  of site-specific  data.   Stream  transport models  in  this  group
involve  dilution  of   contaminant  concentrations  with  increases  in  stream
discharge and  distance downstream from  the  source.   No allowance  is  made in
these models for adsorption/desorption factors.
     The two remaining groups of models  require  much  more  site-specific data,
including  detailed  stream  channel   and flow  characteristics,  as  well  as
adsorption/desorption  and  contaminant degradation factors.   These  models are
best applied when such specific information  is available  and are unsuited for
characterization  of  transport  processes  operating  over   a  wide  range  of
geographic areas where site generalization must be made.
     The unsuitability of  the last- two  groups of models  for  deriving simple
conversion factors  stems from the extensive,  site-specific  data requirements
and  the  complexity of the computations.   However,   even  the first  group of
simple models  requires flow  rates  and  sediment size  distributions  for each
stream-transport scenario  being considered.   The sediment  transport  rate is
derived  using  stream-specific  characteristics   and  empirically   derived
constants that must be estimated for  each sediment size range.  Thus, none of
the models described  by.Onishi,  Whelan,  and Skaggs (1982)  are applicable for
estimating non-stream-specific sediment transport.
     Procedures  utilizing  USLE  losses  and   sediment  delivery  ratios  to
estimate the sediment  yield to  streams from a source  area have been discussed
in  Chapter  4.0.    In order  to  determine  the  concentration  of  dioxin  on
sediments delivered to a  stream,  or  at any  point  downstream of  the source
area,  it  is necessary to  estimate what "fraction of the  total  sediments at
that point were  derived  from the source  area.  Consequently,  the size of the
                                     -53-

-------
 source area and the concentration of dioxin  on  the  soil  at the source must  be
 known.   The  size of  the watershed  must  also be  known or  estimated  from
 topographic maps.   Mechanisms of  sediment  supply,  transport,  and  deposition
 within the drainage basin are assumed to be  in equilibrium.
      The sediment  delivery ratio for  a given  drainage  basin  decreases  with
 distance  downstream  of  a  contaminated  source ;  therefore,  with  distance
 downstream, a  decreasing  portion  of  the total  sediment yield  reaches the
 streams  in the   drainage  area.    The  concentration  of  dioxin  in  stream
 sediments at the  point of exposure  is  a function  of the  downstream  decrease
 in sediment yield, due to deposition of  contaminated  sediments  along  the  path
 of sediment transport between the source  area and the point  of exposure.  The
 relationship between sediment delivery ratio and drainage  area  can  and should
 be regionalized when  applied to a  given site  due  to the effect of  features
 such  as dams.   The required  data for  regionalization are available  for  some
.watersheds.
      Because the  processes described  in Section 4.7 are the  same  as  those
 affecting  sediment  transport,   the  dioxin   concentration  relationship  for
 sediments in a  drainage system is:

           rp            sediment    ^source x   source                     no,
           ur sediment  = ~7~       = T     71                            (U)
                          source       basin    basin
 where   C       -  dioxin concentration  in soil  at the  source  (ug/Kg)
       ^sediment =  dioxin  concentration   in   sediment  at   point  of   interest
                   (ug/Kg)
                                     -54-

-------
^source = estimatecl   soil   1°ss   ^rom  the   USI-E   for   tne   source
          (tons/acre/yr)
Asource = source area (acres)
 Lbasin = estiniated   soil   loss   from  the  USLE  for  the   watershed
          upstream of the point of interest  (tons/acres/yr)
 Abasin = watershed area upstream of  the point of  interest  (acres)
                             -55-

-------
6.0  MISSOURI BEEF DISTRIBUTION PATTERNS
     Human  exposure  to  dioxin  through the  consumption  of  beef  products may
result  if  the   livestock  ingested  and  accumulated  dioxin  as  a  result  of
contact  with contaminated  soils.   Although dioxin  exposure may  occur from
consumption  of  dairy  products  from contaminated cattle,  only meat consumption
is considered  in this discussion.   In areas where  beef consumption involves
locally grown and  fed cattle,  this pathway  can  be  additive  to those stemming
from consumption of  local  fish,   inhalation  of dusts and vapor,  and contact
with  (and/or  ingestion   of)   soil.    In  order  to   quantify this  potential
pathway,  an  understanding  of   the  pattern  of  beef  production  and  meat
processing  in  the area  of  interest is necessary.   The area  of  interest for
this report  is the State of Missouri.
     The  beef  industry  in  Missouri focuses largely on  cow-calf production,
i.e.,  grazing   herds  which  are   utilized  to produce calves  to  a  point  of
weaning.  Backgrounding  (preparing calves for feedlots)  occurs in Missouri to
a lesser extent, and  feeding comprises  only  a small  segment  of the state beef
industry.    The  small feeding segment  is  due,  in  part,   to  the  fact that
Missouri is  a grain-deficit state and  does not  produce  a sufficient excess of
grain to economically support  feeding  operations.  As  a consequence, a large
percentage  of  the Missouri calf  crop  is  shipped to  Nebraska and  Kansas for
backgrounding and feeding until it reaches" marketable size.
                                     -56-

-------
     The Missouri  Crop  Reporting Service reported that the January,  1984  herd
consisted of  2,376,000  head  of cows (Sestak, 1984).  Approximately 25% of  the
herd is  turned  over through  replacement, i.e., heifer calves  are  held back  to
replace  death  losses,  dry   and  barren cows,  or  older  animals which   are
slaughtered  for  low  grade  meat.    The  low  grade  meat  is  later  distributed
nationally for  hamburger and could  be shipped  anywhere in the United States.
Of  the 25% of  the herd turned  over  through replacement, roughly 60%, or  15%
of  the total  herd,  represent cows  sent to. slaughter for  a variety of reasons.
.Total  national   input  to. this  pool  averages 7,000,000  head  annually  (USDA,
1984),  so,  the  594,000  replacement  figure   in Missouri  constitutes 5% of  the
national   inventory  of   cows   destined  for  slaughter  [(0.60)(594,000)
7,000,000].
     At  the  replacement  rate  of  25%,  1,782,000  head  of Missouri  cows  are
available for transport to feeding.  Of that number,  roughly 150,000 are  fed
in  state for  commercial slaughter  (Sestak,  1984).   A second  group of cows  is
held for home slaughter.  Because there are 107,000 cattle ranches  and  beef-
raising  farms in Missouri  (Grimes, 1984), with an  estimated household size of
3.8 people,  and  assuming  an  average   annual  beef  consumption  level   77.4
Ibs/capita  (Berglund,  1984), home slaughter could  account for 107,000 x  3.8 x
77.4  =  31,470,840 Ibs/yr.   This  estimate   is  conservative   because not  all
farms  slaughter  their   own  beef  for personal  consumption.    If  the average
yield  per head  for  home feeding  is 550   Ibs,  31,470,840 Ibs/yr  equate  to
57,219   head.    Hence,  roughly 200,000  head  of  cattle  are   raised,   fed,
slaughtered,  and  consumed  in  Missouri  each  year.   The remaining  1,582,000
head  of calves   are  shipped  out of  state for feeding,  the  bulk  of  which  are
                                     -57-

-------
sent  to  Nebraska  and  Kansas.   In  1983,  Nebraska marketed  4,580,000 head of
fed cattle  and  Kansas  marketed 3,410,000 head, for  a  total  of 7,990,000 head
(Gustafson,  1984).   If  it  is  assumed that  the  1,582,000  head  of Missouri-
raised  calves  are uniformly  mixed  into  this  pool,  fed,   slaughtered,  and
distributed throughout the  area  as  retail beef,  approximately 20% of the beef
imported into Missouri would have  been calved in Missouri.   This estimate is
conservative  because  some beef  may  also  be  imported  from Illinois and other
neighboring states.
     The fraction  of  cattle that leave Missouri  and then return as wholesale
and  retail  beef  will  have been  subjected  to   "clean" feed   during  their
confinement.   The  feed period often  lasts  as long  as  6  months,  or 24 weeks.
Agricultural researchers  have  determined  that the half-life  of dioxin in beef
is  16.5  +  1.4  weeks  (Jensen  et a!.,  1981).  Therefore,  cattle  fed  out of
state will have had a period of  up  to 1.5 half-lives to eliminate dioxin from
their bodies.  This will lead to an overall reduction of
                 -(In 2)t
          C = CQe   \                                                   (34)
                 -(In 2)24
       -  C = CQe    16
          C - C^'-
       C/C0 =0.35
where C = dioxin concentration at time of consumption
     C  = initial dioxin concentration.
                                     -58-

-------
Hence, the beef  produced  from those Missouri-raised calves  that were sent to
out-of-state feed operations and then returned  to  the  state  will carry 35% of
their  original   dioxin  levels.    At  the  same  time,  the  animals  will  have
doubled in  size; thus,  35%  of the original mass  of dioxin  in  these animals
will  be  distributed  over twice  the  total  volume  of  beef,  resulting  in  a
dioxin concentration  equal  to 18% of  the  original contamination  level  CQ in
the beef (i.e.,  C/CQ = 0.18).
     Assuming an average  annual consumption  of 77.4 Ibs/capita,  the 5,000,000
people residing  in  Missouri  will  consume  390,000,000  Ibs of  beef each year,
or  the  equivalent  of 710,000 head.   As  noted  above,  50,000 head  will  have
been  home  slaughtered  and   150,000  head  will  have  been raised  and  fed  in
Missouri.    Therefore, 510,000 head,  or  72% of all  beef consumed in Missouri
(77%  of the  beef consumed  by non-cattle-raising  inhabitants) will  have been
imported from adjoining states.
     Actually,  none  of  the  Missouri   herd  is   known   to  be  contaminated.
However,  if  contamination were detected,  a factor  of  H/HQ»  where H  is the
size of the contaminated  Missouri  herd  and H0  is the total Missouri herd size
of  2,376,000, could be used  to  calculate potential market  dilution effects.
H  should  be estimated on  the basis of  animals on contaminated  pasture.   If
the  actual  herd size  is  not  known,  it can be estimated  based  on  acres  of
contaminated  pasture  and  cattle  density  for  the   state   (i.e.,  cows/acre
pasture normal use).
     Based   on   the  considerations   discussed   previously,   the  following
conclusions can  be drawn  with  respect  to dioxin exposure in Missouri from the
consumption of beef:
                                     -59-

-------
No  natural   degradation  or  market  dilution  will  occur  for  the  home-
slaughter situation.   Thus,  any people consuming  home-slaughtered beef
where  contamination  is  found would  be in a  high risk  population with
respect to beef consumption as a dioxin-exposure pathway.
The remaining Missouri  inhabitants  who  purchase wholesale or retail beef
will  consume beef consisting of  23% raised  and  fed  in  Missouri;  15%
calved in Missouri but fed  out of  state,  and  62% calved, raised, and fed
out of state (Figure 8).
  raised  and  fed
  out of state
raised and fed
 in Missouri

 raised in Missouri
 and fed out of state
          Figure 8.  Origin of Beef Consumed in Missouri
                                -60-

-------
     On the  average,  dioxin  levels  (C)  in  beef  consumed by  the  non-cattle-
     raising inhabitants will  be:

          C = [(0.23)CQ + (0.15)(0.18)CQ  + (0.62)(0)cJ H/HQ             (35)
            = (0.23 + 0.027 + 0)(C0)(H/HQ)
            = 0.26(CQ)(H/H0)
     where  C  =  predicted   level  of  dioxin  for  beef  raised  entirely  in  a
                dioxin-contaminated area
            H = size of the contaminated herd
           H  = total size of the state herd.
     This relationship assumes that all out-of-state cattle are dioxin free.
          The total herd size (HQ) in January,  1984 was 2,376,000, therefore:
          C = (0.26)(CQ)(H/2,376,000)  .
            = 1.1 x 10'7 CQH                    •    '    _
     Approximately  5%  of   the   U.S.   slaughter   cow   inventory  comes   from
     Missouri.   Thus,  subsequent meat products such as  hamburger may contain
     dioxin contamination at
          C = 0.05 (H/HQ)C0                                              (36)
     and for January, 1984 data
          C = 2.0 x 10"8 HCQ
     It should  be  noted that if an  entire  beef or half a  beef  is purchased,
or if  a  large  amount of retail  cuts are purchased  at  a single  time,  all  of
                                     -61-

-------
the meat will  be contaminated or contamination  free  depending on the source.
When viewed  in this manner,  the market dilution  concept  appears inaccurate.
However, when  chronic  exposure is considered, the market  dilution concept is
analogous to  the purchase of  a  small percentage  of  dioxin-contaminated beef
within a larger volume of total beef purchased.
                                     -62-

-------
REFERENCES

ASCE.    1975.    Sedimentation  Engineering.     American   Society  of  Civil
Engineers,  Manuals   and  Reports   on   Engineering  Practice,  No.  54,  V.  A.
Vanoni, Editor, New  York, NY.

Belli,  G.,  G.  Bressi,  S.  Cerlesi  and  S.  P. Ratti.    1983.    "The  Chemical
Accident  at  Seveso  (Italy):     Statistical  Analysis   in  Regions   of  Low
Contamination."  Chemosphere,  Vol.  12, No. 415, pp.  517-521,  Pergamon Press,
LTD, Great Britain.

Berglund, R.   1984.  Personal  Communication.  September 11,  1984.   National
Cattleman's Association, Denver, CO.

Carsel,  et  al.   1984.   User's  Manual  for  the Pesticide Root  Zone  Model.
Release  I.    Athens  EnvironmentalResearchLaboratory,U.S.Environmental
Protection Agency,  Athens, GA.

Collier, C. R.  1963.   "Sediment  Characteristics of  Small   Streams in Southern
Wisconsin, 1954 - 1959."  Water-Supply Paper, 1669-B.  U.S.  Geological Survey,
Washington, DC.

Crosby,  D.  G.,  K.  W.  Moilanen  and. A. S.  Wong.     1973.    "Environmental
Generation    and    Degradation   of   Dibenzodioxins    and   Dibenzofurans."
Environmental  Health Perspectives. Vol.  5, pp. 259-265.

Crosby, D. G., and  A. S. Wong.  1977.   "Environmental  Degradation of 2,3,7,8-
Tetrachlorodibenzo-p-dioxin (TCDO)."  Science. Vol.  195, pp. 1337-1338.

Crosby,  D.  G.,  A.  S.  Wong,  J.   R.   Plimmer   and-   E.  A.  Woolson.    1971.
"Photodecomposition   of  Chlorinated  Dibenzo-p-dioxins."    Science, Vol.  173,
pp. 748-749.

DiDomenico, A.,  V.   Silano,  G. Viviano  and  G.  Zapponi.    1980.   "Accidental
Release of  2,3,7,8-Tetrachlorodibenzo-p-dioxin  (TCDD)  at Seveso,  Italy:   IV.
Vertical  Distribution   of  TCDD  in Soil."    Ecotoxicology and  Environmental
Safety. Vol. 4, pp.  327-338.

Esposito, M.  P., H.  M.  Drake, J.  A.  Smith and T. W.  Owens.   1980.   Pi oxins:
Volume  I.   Sources. Exposure.  Transport, and  Control.  PEDCo.-Environmental,
Inc., Cincinnati,  OH, for U.S. Environmental Protection Agency.

Esposito, M.  P., T.  0.  Tiernan and F.  E. Dryden.  1980.  Dioxins.  EPA-600/2-
80-197.   Industrial Environmental  Research  Laboratory,  U.S.   Environmental
Protection Agency,  Cincinnati, OH.


                                    -63-

-------
Fred C.  Hart  Associates, Inc.   1984.   Assessment of Potential Health  Impacts
Associated with  Predicted  Emissions of PCDD and  PCDF  from Brooklyn Navy Yard
Resource Recovery FacilityINew York, NY.

Freeman,  R.  A.,  and  J.   M.  Schroy.    1984.    "Environmental  Mobility  of
Dioxins."     8th  ASTM  Aquatic  Toxicology   Symposium,   Fort  Mitchell,  KY
(unpublished).

Gillette,  D.  A.   1973.    "On  the  Production  of Soil Wind  Erosion Aerosols
Having the Potential  for Long Range Transport."   Special  Issue of  Journal de
Recherches Atmospherique on  the  Nice Symposium  on  the Chemistry  of  Sea-Air
Particulate Exchange Processes, Nice, France.

Grimes,  G.  1984.   Personal  Communication.    August 30,  1984.   Agricultural
Economics Department, University of Missouri, Columbia, MO.

Gustafson,  R.  1984.    Personal Communciation.    September  12,  1984.    USDA
Livestock Economist, Washington, D.C.

Isensee,  A.  R.,  and  G.  E.  Jones.    1975.    "Distribution  of  2,3,7,8-
Tetrachlorodibenzo-p-dioxin     (TCDD)    in    Aquatic    Model    Ecosystem."
Envir. Sci. Tech.. Vol. 9,  No. 7, pp. 668-672.

Jensen, D. J.,  R.  A. Hummel,  N. H. Mahle, C.  W. Kocher,   and  H.  S. Higgins.
1981.   "Residue Study on Beef Cattle Consuming  2,3,7,8-Tetrachlorodibenzo-p-
dioxin," J. Agric. Food Chem.. Vol. 29, No. 2, pp. 265T266.

JRB,  Inc.    1983.    Draft  Case Study;  Vertac Chemical Corporation.    McLean,
VA.

Kearney, P. C.,  E.  A. Woolson  and  C.  P.  Ellington, Jr.   1972.  "Persistence
and Metabolism  of  Chlorodioxins in Soils."   Envir.  Sci.   Tech.,  Vol.  6, No.
12, pp. 1017-1019.

Leopold, L. 8.,  M. G. Wolman and  J. P. Miller.   1964,  . Fluvial Processes in
Geomorphology.  W. H. Freeman and Co., San Franc.isco, CA.

Liberti, A.,  D.  Brocco,  I.  Allegrini,  A.  Cencinato and M.   Possanzini.   1978.
"Solar  and  UV  Photodecomposition   of 2,3,7,8-Tetrachlorodibenzo-p-dioxin  in
the  Environment."   The  Science of  the Total Environment.   Vol.   10, pp. 97-
104, Elsevier Scientific Publishing Company, Amsterdam.

Longwell, C.  R., R.  F.  Flint  and  J. E.  Sanders.  1969.   Physical Geology.
John Wiley and Sons, Inc.,  New York, NY.

Mabey, W.  R.  et al.   1981.   Aquatic  Fate Process  Data for  Organic Priority
Pollutants.  EPA-440/4-81-014.
                                     -64-

-------
Matsumura, F.,  and H.  J.  Benezet.   1973.   "Studies on  the Bio-accumulation
and    Microbial    Degradation    of    2,3,7,8-Tetrachlorodibenzo-p-dioxin."
Environmental Health Perspectives.  Vol. 5, pp. 253-258.

Matsumura, F., and C. T. Ward.  1976.   Studies  on  the Degradation of 2,3,7.8-
Tetrachlorodibenzo-p-dioxin  (TCDD)  in  Lake  Water  and  Sediment.
Wisconsin University,  Madison,  for Office  of  Water  Research  and Technology,
Washington, DC.

Mills, et  al.  1982.   Water Quality  Assessment:   A Screening  Procedure for
Toxic and Convential Pollutants - Part 1.  EPA-600/6-82-004a.

Midwest  Research   Institute  (MRI).    1984.   Rapid Assessment of Exposure  to
Particulate Emissions from Surface  Contamination Sites.   Anderson-Nichols and
Company,  Inc.

Nash, R.  6., and  M.  L. Beall.   1980.   "Distribution  of Silvex,  2,4-D, and
TCDD Applied  to Turf in Chambers and Field  Plots."    Journal  of Agr.   Food
Chem.. Vol. 28, pp. 614-623.

National   Research  Council.     1971.    Biological  Effects  of  Atmospheric
Pollutants:  Fluorides.  National Academy of Sciences, Washington, DC.

National   Research  Council.     1973.    Biological  Effects  of  Atmospheric
Pollutants: Manganese.  National Academy of Sciences, Washington, DC.

National   Research  Council.     1974a.    Biological  Effects  of  Atmsopheric .
Pollutants:  Chromium.  National Academy of Sciences, Washington, DC.

National   Research  Council.     1974b.    Biological  Effects  of  Atmospheric
Pollutants:  Vanadium.  National Academy of Sciences, Washington, DC.

National   Research  Council.     1975.    Biological  Effects  of  Atmospheric
Pollutants:  Nickel.  National Academy of Sciences, Washington, DC.

Nriagu, J. 0.   1979.    "The Global  Copper  Cycle."   Copper in the Environment.
Part 1.  J. 0. Nriagu (editor), Wiley  Interscience, New York, NY.

Nriagu,  J.   0.    1980.    "Global   Cycle  and  Properties  of  Nickel."   Nickel
in the Environment.  Part 1.    J.  0. Nriagu (editor), Wiley  Interscience, New
York, NY.

Onishi, Y.,  G.  Whelan  and R.  L.  Skaggs.   1982.   Development of a Multimedia
Radionuclide  Exposure  Assessment  for  Low-Level  Waste Management.Battelle,
Pacific Northwest Laboratories, Richland, WA.'

Perkaw, J. A.,-A.  Eschenroeder, M.  Gayer,. J.  Stevens and A.  V'^chsler.   1980.
An  Exposure  and  Risk  Assessment  for  2,3,7,8-Tetrachlorodibenzo-p-dioxin.
Office of WaterRegulationsandStandards,   U.S.EnvironmentalProtection
Agency, Washington, DC.
                                     -65-

-------
Piest,  R.  F.,  and  C.  R.  Miller.   1975.    "Sediment  Sources  and  Sediment
Yields."   Sedimentation Engineering, V.  A. Vanoni  (Editor),  Am.  Soc.   Civ.
Eng. Manual and Reports  on Engineering Practice No. 54, New York, NY.

Pritchard,  P.  H.   1984.   "Fate  of Environmental  Pollutants."   Journal of
Water Pollution Control  Federation,  Vol. 56, No. 6, pp. 718-724.

Roberts,  J. R., J.  A.  Cherry and F.  W.  Schwartz.   1982.   "A  Case  Study of a
Chemical  Spill:   Polychlorinated Biphenyls (PCBs); History, Distribution, and
Surface Trans location."  Water Resources Research.   Vol.  17,  No. 3, pp.  525-
534.

Schaum,  J.   1984.   Risk Analysis  of TCDD Contaminated  Soil.   EPA-600/8-84-
031.    Office  of  Health  and  Environmental   Assessment,   U.S.  Environmental
Protection Agency, Washington, DC.

Sehmel,  G.  A.    1980.   "Particle  Resuspension;   A  Review."   Environment
International, Vol.  4,  pp. 107-127, Pergamon Press.

Sestak,  J.   1984.   Personal  Communication.   August 8, 1984.   Missouri Crop
Reporting Service, Kansas City, MO.

Simons, Li, and Associates.   1982.   Engineering  Analysis of Fluvial Systems.
Ft. Collins, CO.

Sullivan, R.  J.   1969.   Preliminary Air Pollution Survey of Arsenic and its
Compounds.  U.S. Department of Health, Education, and Welfare, Raleigh,  NC. ,

Thibodeaux, L. J.   1983.  "Offsite Transport of 2,3,7,8-Tetrachlorodibenzo-p-
dioxin  from  a  Production   Disposal   Facility."    Chlorinated Dioxins  and
Dibenzofurans in  the  Total  Environment.   G.  Choudhary,  L. H.  Keith,  and C.
Rappe (editors), Butterworth Publishers, Boston, MA..

Turk,  J.  T.    1980.    "Applications  of  Hudson  River  Basin  PCB-Transport
Studies."      Contaminants  and  Sediments,  Vol.  1,   Ann   Arbor   Science
Publishers, Ann Arbor, MI.

Turner,  D.  B.    1970.   Workbook of Atmospheric Dispersion Estimates.    U.S.
Environmental  Protection Agency, Research Triangle Park, NC.

U.S.  Department  of  Agriculture  (USDA).    1984.    "Livestock  and  Poultry
Situation."  p.  4, August.

U.S. Environmental  Protection Agency.   1982a.   Briefing Memorandum  on Horse
Arenas Investigation.  May 3.

U.S.  Environmeatal  Protection  Agency.    1982b.   Summary  of  TCDD  Data *rom
Horse Arenas and Secondary Sites.  August 27.
                                     -66-

-------
Versar, Inc.  1976.   "Technical  and  Microeconomic Analysis."  Arsenic and its
Compounds."  NTIS PB-253 980, Springfield, VA.

Versar,   Inc.  .   1983.     Exposure   Assessment   for   Incidentally  Produced
Polychlorinated Biphenyls (PCBs).EPA-OTS.

Walling,  D.  E.    1983.     "The  Sediment  Delivery  Problem."    Journal  of
Hydrology, Vol. 65, pp. 209-237.

Walters,  W.   H.     1983.     Overland  Erosion   of   Uranium  Mill  Tailings
Impoundments:    Physical  Processes   and  Computational  Methods.Battelle,
Pacific Northwest Laboratory, Richland, WA.

Ward, C. T., and  F. Matsumura.   1978.  "Fate of 2,3,7,8-Tetrachlorodibenzo-p-
dioxin  (TCDD)   in  a  Model  Aquatic  Environment."    Arch.  Envir.  Contam.
Toxicol.   Vol. 7, No. 3, pp.  349-357.

Wischmeier,  W.   H.     1976.    "Use  and  Misuse  of  the  Universal   Soil  Loss
Equation."  Jour. Soil Water Conserv., Vol. 31.

Wischmeier,  W.   H.,  and  D.  D.  Smith.    1978.    Predicting Rainfall Erosion
Losses — A Guide to Conservation Planning.   U.S.  Department  of  Agriculture
Handbook, No. 537, Washington, DC.
                                     -67-

-------
                                  APPENDIX A
                                EXAMPLE SITE 1

A.I  SITE DESCRIPTION
     Consider the  case  of a property  where  dioxin-contaminated  soil  was used
as fill.  The filled area (the source) is  approximately  100 ft  (30.5 m) long
and  100  ft (30.5  m)  wide (0.23  acre).   Sampling  indicated  that  the average
dioxin concentration  in the source  was 150  ug/Kg.   The  source  is currently
without vegetative cover.
     The  'property  is  located  in  a  valley  through  which  a  creek  flows
(Figure A-l).  The source is about 50  ft  (15  m)  from the  creek.   The slope of
the property is  1%.   Soil  type in the area is primarily clay.  Average annual.
wind speed  is  5 mph  (2.24  m/sec), with  the  predominant  wind  direction down
the valley.

A.2  PROBLEM
     Sampling was  conducted  only in the source  area where human health risks
were considered to be highest.   However,  potential  exposure to dioxin is also
of  interest for  areas downwind,  downs lope,  and  downstream  of  the  site.
Concentrations of  dioxin at such  points  of  exposure can  be  predicted  using
the  appropriate  conversion  factors  for  the  various modes  of  environmental
transport.
                                     -68-

-------
Figure A-l.  Example Site 1
          -69-

-------
     In  this  example,   there  is  concern  for  dioxin  concentrations  at  the
following locations:
1)   atmospheric concentrations  at  the  source;
2)   soil and  atmospheric concentrations  near the creek;  and
3)   sediment  concentrations  in  the creek  .1,0.00 ft  (300 m)  down  the  creek
     from the  source.
     The  drainage  basin area  at  Point  2  is  0.05  sq  mi  (32  acres).    The
average  slope in  the area  is  1%,  with  an  average slope  length  of  200 ft
(61 m).
     The  drainage  basin  area  at Point  3 is 0.7  sq mi  (448  acres),  with an
average slope of 2% and  an average  slope  length of 500 ft.

A.3  ATMOSPHERIC CONCENTRATION AT THE SOURCE
     Because  the  average wind  speed is  less  than  10  m/sec,  Equation  21 is
used to calculate air concentrations.
                  C  •
          CFai> « -iHL = i x 10"7 Kg/m3                                  (21)
                  Csoil
           Caif - (1 x 10'7 Kg/m3) Csoi] ug/Kg
                = (1 x 10"7 Kg/m3)(150 ug/Kg)
                - 0.000015 ug/m3
                = 15 pg/m3
A.4  SOIL CONCENTRATION NEAR THE CREEK
     The volume of soil lost from the site source is:
           L=RxKxLSxCxP tons/acre/yr = 12 tons/acre/yr
                                     -70-

-------
           R = 215 yr'1 (interpolated from Figure 6)
           K = 0.3 tons/acre (from Table 2)
          LS = 0.186 (interpolated from Table 4)
           C = 1
           P = 1
Because the  source  area is 0.23 acre, the mass  of  soil  from the site  is  2.76
ton/yr (2,500 Kg/yr).  At  a soil dioxin  concentration  of 150 ug/Kg, 376 mg of
dioxin are transported from the source annually.
     The volume of soil lost from the drainage area above Point  1  is:
           L=RxKxLSxCxP= 7.35 tons/acre/yr
where      R = 215 yr"1
           K = 0.3 tons/acre
          LS = 0.228
           C = 0.5
           p = 1
With a drainage  basin  area of 32 acres,  the  mass  of  soil  lost from the basin
is 235 tons/yr (213,000 Kg/yr).
     The dioxin concentration in soil at Point 2 is:
          C     - r
           soil ~  source
 L       x A
  source    source
. Lbasin x Abasin .
(9)
                = (150 ug/Kg)I"12 tons/acre/yr x 0.23 acre
                             [_7.35 tons/acre/yr x 32 acre
                                     -71-

-------
                   150 ug/Kg  ||Zi
                =1.8 ug/Kg
Because  the  predominant  wind  direction is down the valley (at right angles to
overland  flow  direction),  atmospheric deposition need  not  be considered when
calculating  soil concentrations at Point  2.

A.5  ATMOSPHERIC CONCENTRATION NEAR THE CREEK
     Because  the   predominant  wind  direction  is  down  the  valley  (at right
angles to overland flow  direction), dioxin  in  the soil  at Point 2 will be the
primary  source of atmospheric contaminants.   At an  average wind velocity of
2.24 m/sec,  atmospheric  concentrations can  be calculated using  Equation  21,
as follows:
          Cai> =(1 x 10"7 Kg/m3)Csoil ug/Kg                              (21)
            .   = (1 x 10"7 Kg/m3)(1.8 ug/Kg)
               = 0.00000018 ug/m3
               = 0.18 pg/m3

A.6  SEDIMENT CONCENTRATION DOWNSTREAM
     Soil  losses  from  the  source  are  the   same   as  those  calculated  in
Section A.4.  Soil losses from the basin upstream of Point 3 are:
           L=RxKxLSxCxP=10.5 tons/acre/yr
                      1
where      R = 215 yr
           K = 0.3 tons/acre
                                     -72-

-------
          LS = 0.326

           C = 0.5

           P = 1

From Equation 12, the dioxin concentration in sediment at Point 3 is:
          C         - r
           sediment   ^source
source x  source
 basin x Abasin
(12)
                    = 150 ug/Kg
                    = 150 ug/Kg
 12 tons/acre/yr x 0.23 acre
 10.5 tons/acre/yr x 448 acres
[2.76'
[4700
                    = 0.088 ug/Kg


Alternately, the basin size and  estimated  dioxin  concentration at Point 2 can

be  used  to  estimate  the  dioxin  concentration  in  sediment  at  Point  3,  as

follows:
           sediment " CPoint 2
LPoint 2 x APoint 2.
  Lbasin x Abasin
                     = 1.8 ug/Kg
                     =1.8 ug/Kg
   7.35 tons/acre/yr x 32 acres
   10.5 tons/acre/yr x 448 acres
   235
  4,700
                     = 0.09 ug/Kg
                                     -73-

-------
                                  APPENDIX B
                                EXAMPLE SITE 2

B.I  SITE DESCRIPTION
     Consider the case  of  a private,  rural lane (the source) that was sprayed
several times with dioxin-contaminated  oil  to control  dust.   The lane is 0.25
mi  (400 m)  long and 15 ft  (4.6 m)  wide (0.45 acre).  Sampling indicated that
the average dioxin concentration in the source was 90 ug/Kg.
     The  source is  located on  a hill  with a  reported  average slope  of 3%
(Figure B-l).   A creek flows  through the  valley  at the bottom  of  the hill.
Soil type in  the region is  primarily silt loam.  Average annual wind speed is
12 mph (5.4 m/sec), with the predominant wind direction down the valley.

B.2  PROBLEM
     Sampling was conducted only in the  source  area where  human health risks
were considered to be highest.   However,  potential  exposure  to dioxin is also
of  interest  for areas downwind,  downs lope,  and  downstream  of  the  site.
Concentrations  of  dioxin  at such  points of  exposure  can be  predicted using
the  appropriate conversion  factors  for  the  various  modes  of environmental
transport.
     In this  example,  there are  concerns about dioxin concentrations  at  the
following locations:
1)   atmospheric concentrations at the source;
                                     -74-

-------
    Wind Direction
Overland Flow
        Figure B-l.  Example Site 2
                  -75-

-------
2)   atmospheric  concentrations near  the  residence, which  is 100  ft (30 m)
     beyond the end of the  lane  in the downwind direction;
3)   soil and  atmospheric  concentrations  adjacent to the creek that is 500 ft
     (150 m) downslope from the  source; and
4)   sediment  concentrations  in the  creek  at   a  point  4,000  ft  (1,200 m)
     downstream of Point 3.
     The  drainage basin at  Point 3  has  an  average slope of  3%,  an average
slope length of 1,000 ft (300 m), and an area of  0.25 sq mi (160 acres).
     The  drainage basin at  Point 4  has  an  average slope of  3%,  an average
slope length of 1,000 ft (300 m), and an area of 1 sq mi (640  acres).
     Soil type and average  annual  wind speed  at the points  of  interest are
the same as those at the source.

B.3  ATMOSPHERIC CONCENTRATION AT THE SOURCE
     Because the  average wind  speed is  less than  10  m/sec,  Equation  21 is
used to calculate air concentrations.
                  C
          CFa.r =-^= 1 x ID'7 Kg/m3                                  (21)
                  Soil
Cair = (1 x 10~  K9/IT|) Csoil ug/Kg
                    m
                    3
           Cai> = (1 x 10'7 Kg/m3) (90 ug/Kg)
           Cai> = 0.000009 ug/m
                = 9 pg/m3
                                     -76-

-------
B.4  ATMOSPHERIC CONCENTRATION AT THE RESIDENCE
     Equation 21 is modified  to  include  the  dispersion factor for calculating
air concentrations downwind of the source.
          CF
            air
                   soil
• 1 x lO
                                '7
           (ayazW
where     ^y^F = 46° m'
                  = 540m
                         2
                             -1
               k_ - 0.001 sec
                x = 30 m
                F = 400 m
                W = 5.4 m/sec
Substituting these values into Equation 8:
                             460
                             540
          'soil
(8)
           Cair = 0.00000009 ug/m3
                = 0.09 pg/m3

B.5  SOIL CONCENTRATION NEAR THE CREEK
     The soil lost from the source is
           L=RxKxLSxCxP tons/acre/yr = 20 tons/acre/yr
where      R = 215 yr"1
           K = 0.4 tons/acre
          LS = 0.233
                                     -77-

-------
           C = 1
           p * -1
Because  the  source  area  is  0.45  acre,  the mass  of soil  from the  site  is
9 tons/yr  (8,200 Kg/yr).   At a soil dioxin  concentration  of  90 ug/Kg, 738  mg
of dioxin  are transported  from the  source annually.
     The soil lost from the  drainage basin above Point 3  is:
           L=RxKxLSxCxP tons/acre/yr = 25 tons/acre/yr
where      R = 215 yr"1
           K = 0.4 tons/acre
           LS = 0.573
           C = 0.5
           P = 1
With a drainage  basin of  160  acres,  the  mass of soil lost  from .the basin  is
4,000 tons/yr (3,629,000 Kg/yr).
     From  Equation 9, the  dioxin concentration in soil at'Point  3 is:
          C     - r
           soil "  source
                = 90 ug/Kg
source x  source
Lbasin x Abasin
                            20 x 0.45
(9)
                            25 x 160
                =0.2 ug/Kg
     Because  the  predominant  wind  direction  is  down  the  valley  (at right
angles  to  overland  flow  direction),   atmospheric  deposition  need   not  be
considered in calculating soil concentrations at Point 3.
                                     -78-

-------
B.6  ATMOSPHERIC CONCENTRATION NEAR THE CREEK
     Because  the predominant  wind direction  is  down the  valley  (at  right
angles to overland flow  direction), dioxin  in  the soil at Point 3 will be the
primary source  of  atmospheric  contaminants.   At an average  wind  velocity of
5.4 m/sec,  atmospheric  concentrations will be due  to  vaporization, according
to Equation 21:
          Ca.r -(1 x 10'7 Kg/m3)Csoil ug/Kg-                               (21)
               « (1 x 10'7 Kg/m3)(0.2 ug/Kg)
               = 0.00000002 ug/m3
               = 0.02 pg/m3

B.7  SEDIMENT CONCENTRATION DOWNSTREAM
     Soil  losses  from  the  source  are  the  same   as   those  calculated   in
Section B.5.  Soil losses from the basin upstream of  Point 4 are:
           L=RxKxLSxCxP=25 tons/acre/yr
where      R = 215 yr"
           K = 0.4 tons/acre
          LS = 0.573
           C = 0.5
           P = 1
From Equation 12, the dioxin concentration at Point 4 is:
                               L       x A"
          C           r         source     source
           sediment ~  source   i       Y /\
                                 basin. * "basin
                                     -79-

-------
                = 90 ug/Kg
20 x 0.45
25 x 640
                = 0.05 ug/Kg


Alternately, the basin size  and  estimated dioxin concentration at Point 3 can

be  used  to  estimate the  dioxin  concentration  in  sediment  at  Point  4,  as

follows:
          CSediment = cPoint 3
                    = 0.2 ug/Kg
                    - 0.2 ug/Kg
                      0.05 ug/Kg
    LPoint 3 x APoint 3
      Lbasin x Abasin
     25 tons/acre/yr x 160 acres
     25 tons/acre/yr x 640 acres
                                  4,000
                                 16,000
                                     -80-

-------