UNITED STATES £N V 1 PQNMENT AL
                      "WASHING 10% DC
 July  31,  1989                             EPA-SAB-&ETFC-89-027

The Honorable William Reilly
Administrator
U.S. Environmental Protection Agency
401 M. Street, S.W.
Washington, D.C.   20460

Dear Mr. Reilly:

     The Sediment Criteria Subcommittee of  the  Science Advisory
Board has completed its review of the Apparent Effects Threshold
(AET)  Approach to setting sediment criteria.  This approach was
developed by EFA's Region 10, Office of Puget Sound.  The review
was conducted  at  the request of Region  10's Administrator,  Mr.
Robie Russell, and was conducted on October  27 and  28 in Seattle,
Washington.

     The AET approach is designed to  identify adverse effects .due
to  chemical  contamination  in sediments by  determining specific
chemical concentrations above which  adverse effects will always
occur.    The method  has major strengths  in  its  ability  to
determine biological  effects and assess interactive  chemical
effects.  The method is considered by the  Subcommittee to contain
sufficient scientific merit that, with appropriate  validation, it
could  be  used to .estimate sediment  quality at specific sites.

     In the  Subcommittee's opinion,  the  AET approach should not
be used to develop  general,  broadly applicable  sediment quality
criteria.   Some major limitations  drive  this opinion, including
the site  specific  nature of the  approach, its inability to
describe cause and effect relationships,  its lack  of independent
validation,   and  its inability to  describe  differences  in
bioavailability of chemicals  on  different  sediments.

     The Subcommittee  recognizes the Apparent Effects Threshold
Approach as a credible step towards development  of a technically
defensible and publicly acceptable tool for  managing contaminated
sediments.   The  approach  provides a constructive  beginning
towards assessing  the impact of mixtures of chemicals  as they
occur in actual situations.   Such innovative empirical  approaches
that  assess  actual  contamination  and concomitant  effects are
encouraged and applauded by the  Subcommittee.   However, the
Subcommittee also recommends that  neither  the AET or any  other
existing methodology  be used as a  stand-alone  decision tool to
provide absolute pass/fail criteria  to dictate regulatory action.

-------
     The Subcommittee has  several  suggestions  for strengthening
the AET  approach.    These  include utilizing replicate  sediment
samples  for assessments,  devising  criteria  for selection  of
reference sites,  including considerations of  physical  factors,
and developing measures  of variance.    In addition,  the use  of
both  carbon  normalization  and  benthic  infaunal  assays  are
strongly  supported by the  Subcommittee.

     The Subcommittee appreciates  the opportunity to conduct this
scientific review.   We request that the Agency formally respond
to the scientific advice  transmitted in the attached report.
                                  Sincerely
ENC
cci Rebecca
    Martha Frothro
    James M. Conlon
    Robie Russell
    Chris Zarba
    Catherine Krueger
    Donald Barnes
                                  Dr '. 'Raymond Loehr,  chairman
                                  Executive Committee
                                  Science Advisory Board
                                  Dr. Kenneth Dickson, Chairman
                                  Environmental Effects,
                                     Transport and Fate
                                     Committee
w./
                                                           L/
                                   Dr. Robert^luggfet:
                                   Sediment briteri
               chairman
              ubconunittee

-------
    United States     Office of the AdnMstrator SAB-EETFC-89-027
    Environmental Protection Science Advisory Board July 1989
    Agency       Washington, D. C - 20460
4>EPA  Report of the Sediment
         Criteria Subcommittee
         Evaluation of the Apparent
                            »
         Effects Threshold (AET)
         Approach for Assessing
         Sediment Quality

-------
             UN|TED STATES ENVIRONMENTAL PROTECTS AGENCY
                         I»V A ft H i ,\ r, t Q fj. o C  20460

 July 31,  1989
                                             EPA-SAE-EETFG-89-027
The Honorable William Reilly
Administrator
U.S. Environmental protection Agency                   ,L,t ^;,.:"' -
401 M. Street,  s.w.
Washington, D.C.  20460

Dear Mr, Reilly:

     The Sediment  Criteria  Subcommittee  of  the Science Advisory
Board has completed its review of the Apparent Effects Threshold
(AET) Approach to  setting  sediment  criteria.   This approach was
developed by EPA's Region 10, Office of Puget Sound.  The review
was conducted  at the request of  Region 10'a Administrator, Mr.
Robie Russell,  and was conducted on October 27  and  28  in Seattle,
Washington.

     The AST approach is designed to identify adverse  effects due
to  chemical  contamination in sediments  by  determining specific
chemical concentrations above which  adverse  effects will always
occur.   The method  has  major  strengths  in its ability  to
determine  biological  effects and assess interactive  chemical
effects.  The method is considered by the Subcommittee to contain
sufficient scientific merit that,  with appropriate  validation, it
could  be  used to estimate sediment quality  at specific sites.

     In the  subcommittee's  opinion,  the AIT approach should not
be used to  develop general, broadly applicable sediment quality
criteria.   Some  major limitations drive this opinion, including
the  site  specific nature  of the  approach,  its  inability to
describe cause and effect relationships, its lack  of  independent
validation,  and  its inability  to describe differences  in
bioavailability of chemicals on different sediments*

     The Subcommittee  recognizes  the Apparent Effects Threshold
Approach as a credible step towards development of  a  technically
defensible and publicly acceptable tool  for managing contaminated
sediments.   The approach provides a  constructive  beginning
towards assessing  the impact of  mixtures of  chemicals  as they
occur in actual situations.   Such innovative  empirical approaches
that  assess  actual  contamination and  concomitant  effects are
encouraged and  applauded  by the  Subcommittee.   However,  from  a
scientific standpoint, the  Subcommittee  recommends  that multiple
approaches  be used  to  estimate  sediment  quality,  develop
criteria,  and  guide  regulatory  action,  since the AET approach
alone provides  insufficient certainty  for  broad-scale decision
making.

-------
     The Subcommittee has  several  suggestions for strengthening
the AET  approach.    These include utilizing  replicate sediment
samples  for assessments, devising  criteria for selection of
reference sites,  including considerations of physical factors,
and developing  measures  of variance.   in addition, the  use of
both  carbon  normalization  and  benthic  infaunal  assays  are
strongly  supported by the Subcommittee.

     The Subcommittee appreciates the opportunity to conduct  this
scientific review.   We request that the  Agency formally respond
to the scientific advice transmitted  in  the attached report.
                                   Sincerely
ENC
cc: Eebecca Manner
    Martha Prothro
    James M. Conlon
    Mobie Russell
    Chris Zarba
    Catherine Krueger
    Donald Barnes
                                   Dr. 'Raymond Loehr,   Chairman
                                   Executive Committee
                                   Science Advisory Board
                                             L
                                   Dr.  Kenneth Dickson,  Chairman
                                   Environmental Effects,
                                      Transport and Fate
                                      Committee
                                   Dr.  .Robertyipigget,
                                   Sediment criteri
 Chairman
ubcowmittee

-------
                           ABSTRACT


     This report presents  the  conclusions and recommendations of
the U.S,  Environmental Protection Agency's Science Advisory Board
summarizing a review of EPA's Apparent Effects Threshold Approach
for  setting  sediment quality  criteria.    The  AET  approach
integrates data from  bulk  sediment  chemistry, sediment bioassays
and  infaunal  species measurements  to  provide  estimates  of
sediment   chemical   concentrations   above   which   adverse
environmental effects  will  occur.  An  objective  of the  AET
methodology is to identify adverse  effects  due to  chemicals
occurring in mixtures  in  sediments by determining  specific
chemical  concentrations  above  which adverse effects  will always
be  found.     The  method has major  strengths in  its ability to
determine biological effects  and  assess interactive chemical
effects.   The method  is considered by the Subcommittee  to contain
sufficient scientific  merit  that, with appropriate validation of
the AET values, it could  be used to  establish sediment quality
values for use at  specific sites.  In the Subcommittee's opinion,
the AIT approach should not be used to develop general, broadly
applicable sediment  quality  criteria.   Some  major limitations
drive this  opinion,  including  the  site specific nature  of the
approach,   its   inability   to  describe   cause   and  effect
relationships,  its  lack  of  independent  validation,  and its
inability to describe  differences in bioavailability  of chemicals
on different sediments.  The Subcommittee has several suggestions
for  strengthening  the AiT approach including:  building in
replicate sediment samples to  assessments,  devising  criteria for
selection of  reference  sites,  including  considerations  of
physical  factors,  and  developing measures of variance.

Kay words;  Sediment•,  AIT? Apparent Effects  Threshold

-------
              U.S. ENVIRONMENTAL PROTECTION AGENCY

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

-------
              U.S. ENVIRONMENTAL PROTECTION AGENCY
                     SCIENCE ADVISORY BOARD
       ENVIRONMENTAL EFFECTS, TRANSPORT AND FATE COMMITTEE
                 SEDIMENT CRITERIA SUBCOMMITTEE

                             ROSTER
CHAIRMAN
     Dr. Robert Huggett
          Professor of Marine Science
          Virginia Institute of Marine Science
          School of Marine Sciences
          College of William and Mary
          Gloucester Point, Virginia  23062

VICE CHAIRMAN
     Dr. Rolf Hartung
          Professor of Environmental Toxicology
          University of Michigan
          3125 Fernwood Avenue
          Ann Arbor, Michigan  48108-1955

MEMBERS
     Dr. William J. Adams
          Monsanto Company (U4G)
          800 N. Lindbergh Blvd.
          St. Louis, Missouri  63167

***  Dr. Peter M. chapman
          Partner
          E.V.S. Consultants
          195 pemberton Avenue
          North Vancouver, B.C.
Canada V7P 2R4
     Dr. Kenneth L. Dickson
          University of North Texas
          Institute of Applied Sciences
          P.O. Box 13078
          Denton, Texas  76203

     Dr. Benjamin C. Dysart III
          Professor of Environmental and Water
            Resources Engineering
          401 Rhodes Engineering Research Center
          Clentson University
          Clemson, South Carolina  29634-0919

     Dr. Eugene Kenaga
          Consultant (Ret*/Dow)
          1584 E* Pine River Road
          Midland, Michigan  48640
*** Dr. Chapman served on the Subcommittee until December  6,  1988,

                               iii

-------
     Dr. Frederic K. Pfaender
          Department of Environmental Sciences
            and Engineering
          University of North Carolina
          Chapel Hill, ^North Carolina  27599-7400

     Dr. Y. Peter Sheng
          Professor, Department of coastal and
          Qceanographic Engineering
          336 Weil Hall
          University of Florida
          Gainesville, Florida  32611

SUBCOMMITTEE LIAISONS

     Dr. Robert M. Engler
          (CEWESEP-D)
          Waterways Experiment Station
          U.S. Army Corps of Engineers
          P.O. Box 631
          vicksburg, Mississippi  39180

     Dr. Chris Ingersoll
          National Fisheries contaminant
            Research Center
          U.S. Fish and Wildlife Service
          Route 2
          Columbia, Missouri  65201

     Dr. H. Suzanne Bo1ton
          NGAA-DOG
          Office of Legislative Affairs
          (LAX-2)
          Herbert c. Hoover Building
          loom 5222
          Washington, D.C*  20230
SCIENCE ADVISORY BOARD STAFF

     Ms. Janis C. Kurtz
          Environmental Scientist & Executive Secretary
          U.S. Environmental Protection Agency
          Science Advisory Board
          401 M Street, s.W.
          Washington, D.C.  20460
                              iv

-------
                        TABLE OF CONTENTS



l.o  Executive Summary-   ................    l

2,0  introduction ....................    2

     2.1  Request for Science Advisory Board Review . . ,    2

          2,1.1 Charge to the subcommittee	    2

     2.2  Subcommittee Review Procedures  ........    2
     2.3  Expected Future Activities	* . .    3

3.0  Evaluation of the Apparent Effects Threshold Approach   4

     3,1  Quality of Data for AET Determination	    4
     3.2  Determining spatial Extent of Contamination . .    5
     3.3  Field Applicability and Site Specificity  ...    6
     3.4  Cause-and-Effect Relationships  ........    6
     3.5  Endpoint Considerations .,, 	 ....    8
     3.6  Complex Mixtures  .......... 	   10
     3.7  Physical Factors  . . 	 ........   11
     3.8  Carbon Normalization  .............   11
     3.9  Uncertainties in the AIT	   12

4.0  sumraary of Subcommittee Recommendations   ..*..,   15
Appendix A - Briefing Report to the EPA Science Advisory Board:
               The Apparent Effects Threshold Approach

-------
1.0  EXECUTIVE SUMMARY


     The Apparent Effects Threshold (AET)  approach for deriving
sediment  quality values integrates  data  from bulk  sediment
chemistry, sediment bioassays and  infaunal  species measurements
to  provide  estimates  of sediment cheaical   concentrations.
Concentrations above these  estimates  are thought  to  result in
adverse  environmental  effects.    An  objective  of  the  AET
methodology  is to identify adverse effects due  to  chemicals
occurring in mixtures in  sediments by  determining  specific
chemical concentrations above which adverse effects will always
be found.

     The major strengths  of  the  approach  are  a)  the  fact  that
both infaunal analyses  and laboratory bioassays  are incorporated
to determine  biological effects and b)  the fact  that the approach
has the ability to incorporate interactive cheaical  effects,  such
as  synergism  and antagonism under  the  specific conditions
encountered in the environment where  the AET  is  applied.   The
method is  considered by the Subcomnittee to contain sufficient
scientific merit  that,  with appropriate validation of  the AET
values, it could be used to establish sediment quality values for
use at specific sites.

     The major limitations of  the AET method  are  a)  its  site-
specific nature   b)  its inability to  describe  cause and effect
relationships for specific chemicals c)  its  lack of independent
validation of the AIT  values and  4)  its inability to describe
differences   in   bioavailability  of chemicals  on  different
sediments.   These factors restrict the applicability  of the
specific AEt  values  to the locality and conditions under  which
they were developed.  In the Subcommittee's opinion,  the AET
approach  should not  be  used  to develop general,  broadly
applicable sediment quality criteria.

     The validity of  the AET estimates can be improved by  formal-
izing  criteria,  designating reference  sites, and  developing
measures   of  uncertainty  for   specific   AET  values.    The
Subcommittee  recommends that  AET values be  derived  based  on more
than one sediment bioassay, and that benthic infaunal analyses be
included in AET  development.   in  addition,  the accuracy of AET
values can toe assessed  by comparison with  endpoints and  through
other  experimental  approaches that  permit  examination  of the
major convergences and  divergences in the values.

-------
2.0  INTRODUCTION

2.1  Request for Science  Advisory Board Review
     At the request of the Regional Administrator of EPA's Region
10,  the Science  Advisory Board (SAB)  agreed  to  conduct  a
scientific review of  the Apparent Effects Threshold  (AET)
approach.   The* SAB's Environmental  Effects,  Transport and  Fate
Committee authorized the formation of a Subcommittee to perform a
series  of  tasks related to  the  technical  aspects of  sediment
quality assessment  and criteria development, with the approval of
the Board's Executive Committee.

2.1.1  Charge to the Subcommittee

     This review is the first  to be completed in a series of SAB
reviews related to sediment  quality  criteria.   Preliminary
reviews focus  on  approaches  to examining  non-ionic  organic
contaminants.   Specifically,  the  charge  accepted  by  the
Subcommittee is to:

     a)  Evaluate  individually  the scientific  and technical
foundations for the various methodologies available to the Agency
to estimate  sediment toxicity  and biological impact  of  contam-
inated sediments in situ.

     b)  Evaluate  the feasibility of utilizing  each methodology
to determine extent of contamination  and risk posed to the envi-
ronment and human health,

     c)  Identify research  directions that would provide informa-
tion to strengthen  each methodology.

     d)  Conduct a technical review of documents provided by the
Agency  that  compare and contrast the scientific  and technical
strengths  and  weaknesses of  the methodologies available  to the
Agency  to  estimate  sediment toxicity  and biological  impact of
contaminated sediments insitu.
2.2

     The Sediment  Criteria  Subcommittee  met on August  8  and 9,
1988,  in Denver,  Colorado,  to  assess  the  Agency's  activities
regarding  contaminated sediment and to explore avenues  for
providing oversight.   A second meeting was held on October 27 and
28, 1988, in Seattle, Washington.  Informative and well-prepared
briefings were provided on the objectives, historical perspective
and technical components of the AET.  The  briefings were supported
by extensive documentation provided to Subcommittee members prior
to the meeting.   Both the  briefings and the documentation were
considered by the Subcommittee to be very well done and relevant
to the  issues under review.   This preparation and  support  was
provided by Region  10 staff and associated contractors (PTI

-------
Environmental Service,  Inc.). The Environmental protection Agency
is divided  into  10 Regional Offices, and  Region 10 serves  the
states of Alaska, Washington,  Oregon,  and  Idaho  from  its  base in
Seattle,  Washington.

     The contamination present  in  Puget  Sound was characterized.
Biological  assays used  in AET  development were  described.
Additional  briefings  highlighted  the  statistical  procedures
utilized  in  the  method,   and  procedures  used  to  treat
uncertainties in  both biological and chemical  data.   Comments
were heard  from  interested  members  of industry  and  the  public,
and  intended  application of  the nethod  was  described  by  the
Washington state Department  of  Ecology,


2,3  Expected Future  Activities

     The Sediment  Criteria Subcommittee plans to review  and re-
port on  the  technical  appropriateness of  the Equilibrium Parti-
tioning  (IP) method  in the  near future.    In addition,  bioassay
procedures that are used as an integral part of  the  AET,  EP and
other  approaches  will be subjected to review  for scientific
adequacy.  Other  methods, including methods  for assessing metal
contaminants,    are  expected  to  be developed  and existing
approaches refined, and as this occurs,  they  will be transmitted
to the Subcommittee  for review.   The Contaminated Sediment
Technical Committee,  established by  EPA's  Office  of Water,  is in
the  process of preparing a manual which describes currently
available methods  that may  be  applied to  establishing  sediment
criteria.  The Subcommittee will  comment  on this  manual  in the
future.

     During the course of these critical evaluation processes, it
is likely that areas for additional or future  research  will be
targeted.  To  facilitate  the  incorporation of these recommenda-
tions into EPA research planning,  the Subcommittee may conduct a
review of the office  of Research and Development's proposed Sedi-
ment Initiative.   The  time  sequence of these proposed events is
contingent on their completion  by  Agency staff.

-------
3.0  EVALUATION OF THE APPARENT EPFBCT8 THRESHOLD APPROACH


     The AET approach provides useful information for developing
the  weight  of evidence  needed to make  decision regarding
contaminants in sediment.  The  AET is a  statistically-based
empirical  approach  which attempts  to establish quantitative
relationships  between  sediment  contaminants   and resulting
biological  effects.  It has been developed using synoptic data on
chemical contaminants  and biological  effects  (as   assessed  by
benthic infaunal  analyses  and sediment bioassays) at suspected
contaminated sites and reference sites.   The AET represents  a
credible step towards development  of  a technically defensible and
publicly acceptable tool for  managing contaminated sediments.

     The use of quantified data  for  the derivation  of relation-
ships  between  exposure and  effects  is the  only scientifically
justifiable basis  for describing  such relationships quantitative-
ly.   Thus,  the  technical  acceptability of  the  development  of
sediment quality  values depends  upon  the  fidelity  with  which
measures of exposures and effects on the environment can reflect
"true" environmental  conditions,  and upon the characteristics of
the model that seeks to summarize these complex relationships in
the form of a descriptor of sediment  quality.

     The AET  approach provides a constructive  beginning toward
assessing the  impact  of mixtures  of chemicals as they  occur in
actual situations, as  opposed  to solely assessing the influence
of single  chemicals  under  laboratory conditions.   Its  major
strength  is  its   ability to  assess  impacts  of contaminated
sediments on aquatic life.    Innovative empirical approaches such
as the AET that  assess actual  contamination and  concomitant
effects are encouraged and  applauded  by  the Subcommittee.


3.1 Quality of Data for AET Determination

     The subcommittee  is concerned  that several  limitations of
the data sets used to develop the AET values have influenced the
interpretation  of  relationships  between  sediment contamination
levels and resulting biological effects.   This concern steins in
part from the lack of true field replication  of the chemistry and
bioassay data used in the development of the  AET.  The preponder-
ance of the chemical  and bioassay data  is based on single samples
from each site.   While a sample may be divided into sub-samples
and chemical  analyses  and/or bioassays  performed on  sub-samples
(thus  measuring the  precision  of the respective methods),  there
is  no estimate of  the within-site  variability   of  chemical
concentrations  and sediment  toxicity  (as  measured  by  sediment
bioassays}.  Although in some  cases  {e.g.,  assessments  of the
suitability of material for  dredging) within-site variability is
less  important  than  in other  cases  (e.g.,  determination of the
extent and significance of problem areas),  some estimate of this
variability is considered necessary.

-------
     On the other hand,  the data base used to develop an AET does
contain an estimate of  infaunal variability at  each  site,  since
replicate field samples  are  independently  analyzed.    The
appropriate level of  infaunal community  structure  analysis must
be carefully evaluated.and  applied.  Since there  is  no estimate
of the variance for any of the  chemical concentrations,  an
undefined uncertainty exists  around  the  AET,    The su*™™"***^*"*
recommends that  future  application of  the  AET include replicate
sediment  samples  for  determination of  chemical  contaminantsand
for assessment of toxicitv.   By so doing,  the uncertainties can
be better defined and *  stronger relationship between contaminant
levels and  biological effects established.   In  addition,  with
field replication data available on chemical contaminants, it may
be  feasible  to  statistically estimate  the variance of  the
chemical data used in the development  of the  AET,  a  feature not
presently incorporated in this or most  other field approaches*
3.2

     One of the proposed uses of AET is to quantitatively define
the  spatial  extent  of  biological  impacts associated  with
contaminated sediments.   Because the  method  depends  on  a
statistical comparison of  test sites to reference  sites,  it is
critical that valid reference sites be  determined.   A crucial
factor in determining the location of the reference  sites is that
they be physically  similar to  the test sites.   Although several
sediment  characteristics  were  considered  in  the  selection  of
reference sites  (e.g.  season,  water depth, grain size, or organic
carbon content)  and decisions  were made  using best professional
judgment, it does not  appear that a formalized  set of decision
criteria have been  established. If  inappropriate reference sites
are used, an assessment of whether or not a test site is impacted
is  difficult.   The Subcommittee  recommends	that  criteria  for
selecting   reference   sites   be   formalized.	   The
aeleetion/re-i action criteria need  to be  clearly  defined and the
rationale for their choice explained.

     Furthermore,  it is  important that the  AET approach clearly
recognize and incorporate systematic temporal changes  that may
impact both the  materials present in the sediment and the effects
they may have.   In reference areas there will  be temporal as well
as  spatial  changes in the  composition  and  functioning  of the
resident faunal  community.  In impacted areas there will also be
temporal changes in the  concentrations of chemicals as a result
of physical (burial, re-suspension, etc.), chemical  (degradation,
sorption,  etc.),  and biological  (biodegradation) processes.
Biologically, the community that is present at any site  exists in
response to the matrix of  chemical,  biological  and  physical
factors that are present.  Therefore, the presence  of pollutants
represents a selective influence that may  lead over  time  to a
community adapted  to  the  conditions present, which may be
different than the  reference area but not necessarily  "impacted"
in a functional  sense.

-------
     In  its development,  the  AET  methodology  has  not clearly
addressed the influences of temporal factors on the  distribution
of  benthic biota  and  the  bioavailability  of  sedimentary
contaminants.   It would be  useful  to  conduct repeated sampling
over  some  time period- at a  few stations  to establish whether
changes  occur  that would significantly alter the generated AET
value.   Some Knowledge of how  temporal changes impact the AST is
neededto characterize tie uncertainty thatmay be added by  this
variable.
3,3  Fieia Applicability and Site Specificity

     The AET values produced from the Puget Sound data appear to
work well in Puget Sound*  The Subcommittee recognizes the merits
of the AET for identifying potential problem areas and potential
problem chemicals, since AET are currently being proposed for use
as part  of a process that  involves site-specific  biological
testing  in Puget  Sound,  as  opposed to broader,  more generic
application,  this  application  seems  to be  consistent  with the
Subcommittees recommendations.

     However,  application of Puget Sound AET to  another location
or different physical  setting  must  be done with extreme caution
because the effects of physical, chemical  and  biological factors
on AET  are not  well  understood.   For  example, differences in
current and wave conditions from one site to another may lead to
very different bottom sediment compositions, even within the same
body  of  water.    The  density,  size  distribution,  salinity and
degree of  flocculation  of  sediments  may be  quite different at
different sites.   Moreover, the AET's  inability to address the
causality of biological  response by a single  chemical, although
not unique to the AET approach, must be  taken into  consideration.
Even  though  one  can argue  that  an AET may  be  generically
protective, there is presently no evidence that AITs developed in
one place are protective in another environment.

     High levels of chemical contamination may result in toxicity
to laboratory species,  but may cause no alteration of the real
world benthos.  This  situation may indicate a potential problem
that  is  not  yet  realized  due  to  such  factors as  adaptation,
hormesis  (i.e.,  stimulation of growth   and/or fecundity  in the
presence of low levels of contamination  or stress-induced vigor),
etc.   such information is valuable and could be misinterpreted by
the AET as presently  applied.   Tha Subcommittee r«
-------
method does  not provide a  way of  isolating  biological  effects
that are caused by  one  specific  chemical  when the same chemical
is present in a mixture of known and unknown chemicals sorbed to
sediment.   Therefore the scientific defensibility  of chenical-
specific AET values is .unknown.   A  logical conclusion is that the
AET method  is not  capable  of demonstrating  specific  cause and
effect relationships for any one  specific  chemical.  Furthermore,
the report  provided to the  Board  ("Briefing Report to  the EPA
Science  Advisory  Board" ,  page  15,  see  excerpt in Appendix A)
states  that  "...   it  cannot  distinguish  and  quantify  the
contributions of  interactive  effects,  unmeasured  chemicals or
matrix effects in environmental samples. M

     The AET  method can generate  values  for a  wide  variety of
chemicals.   However, to assess th*  general applicability of AETs.
the  implied  relationship  between   the   chemical  sediment
concentration and  biological affect  must be validated.   The
following points are presented to support  this contention;

     1. The existing data set does not contain verification  data
from experiments with spiked sediments  or  from toxicity identifi-
cation evaluations which would assess  the accuracy of the AETs.
The work of Swartz with fluoranthene and amphipods is an example
of  the  type  of  study  that  could  be done  to  assess  AET
accuracy (Dr* Richard Swartz, EPA/QRD,  Newport, Oregon; personal
communication) .      Additionally,   toxicity   fractionation-
identification for  specific  chemicals  from  elutriate  or liquid-
phase sediments at selected  sites could be used.

     2. The AET values  for  phthalates  do  not reflect the physi-
cal^chemical properties of this group of chemicals.  it would be
expected that the AET values would increase  for the  more water
soluble,  less toxic phthalates such as dimethyl and  diethyl
phthalate.   This is not the  case.  This may  reflect an  inadequate
data set for  the phthalate class,  which would  improve  with
additional  sampling, but it also indicates a  fundamental lack of
consistency with  the  existing  literature  on  this class of
compounds.  Belated comments  can be  made for  some other  chemicals.
It JLa y*«!ggHiiift^ttded that tha_-_data M|t__be ray4.«*a4_JEoy all tlia  ABTs
in relation to their phsical-chemical   roperties (e.g., Koc. Lo
P  and  water  solubility)  and  feftair  existing
properties  (e.g.,  acut* versus chronic or  genetic effects)  for
tha wirot   o^  •ftfc-l Jailin   the sensitivit   of  3UBT  values  to
parameters that  are known to  affect toxic! tv.   These types  of
data can be used  as supportive evidence of cause-and-ef feet
relationships.

     3, Several of the AETs are below existing acute and  chronic
toxicity values for aquatic organisms based on aqueous exposure.
In water,  most chemicals  are  assumed to  be  bioavailable, with
some exceptions.   Bioavailability  is generally considered to  be
reduced when a chemical is sorbed to sediment.  Therefore  it does
not seem  plausible that AET values for protecting aquatic life
from chemicals sorbed to  sediments can  be lower than  chronic
water exposure values.  It— is  recommended  tftat tha 2US*E.. values

-------
compared to  existing  toxioity values	for aquatic  organisms  for
the  purpose  of  identifying  those   AET  values  vhieft  arc
inconsistent vith the existingbody of toxicoloaical data.

     4. The  sediment Jaioassay and  field  infaunal biological
effects measurements do not show a strong dose-response relation-
ship above the AET values*   A basic premise of toxicology is that
effects intensify with increasing dose.  Inability to demonstrate
this relationship  argues  against  a  causal  relationship  between
exposure and effect for these chemicals near  the AET concentra-
tion.   It therefore  diminishes  the weight  of evidence  of  the
accuracy of the AET values  as they  relate  to specific chemicals
and  indicates  that the toxieity  may be  due  to  more than  one
chemical  or  that  bioavailability  or  population  resistance  is
changing from station to station.   Both factors detract from the
method's  ability  to  relate observed effects to  a specific
chemical  or chemical  concentration.   it  la recommended that
additional researchbe conducted to evaluate AETs relative to the
applicable,  available  data  for  dose-response relationships.
Additionally,  most of  the existing AETs  have  only  a  limited
number of stations that fall above the  AET value.  This suggests
that additional sampling and analysis could be beneficial to the
AET method if  the samples  were  collected  from sites  with high
chemical concentrations.

     5, Several of the AET values are for chemicals which do not
show acute toxicity in laboratory aqueous  toxicity tests.   This
is generally due to the limited solubility of the test chemical.
Since  the AET  values are derived primarily  from acute toxicity
tests,  one would expect a priori that these same chemicals would
not be acutely  toxic  when sorbed to sediments.   since the bio-
availability of  chemicals sorbed to sediments is  usually less
than that in the aqueous phase, the trend is  for chemicals to
become  less  toxic  on  sediments,  not  more toxic.    These data
suggest that the  effects  are due,  in  part,  to  the  presence of
other  chemicals,  the  physical  properties  of  the  sediment,
or  some other  unknown factor(s).   Inconsistency with  known
toxicological principles raises questions about  the accuracy of
AET values and, therefore, the scientific validity  of the method.

     The  AET  has clear  utility  for  assessing contaminated
sediments.   Points such  as  those  raised above.,  once addressed
through  additional  research   or  comparison  with  existing
literature,  can go a long  way towards improving confidence  in the
method  and  in  demonstrating its  consistency  with established
scientific principles.


3.5 Endpoint Considerations

     The ability  of the  AET to  effectively  assess  impacts at
contaminated sites  is related to  the  choice  of  the  biological
effect  endpoint(s)  used.   The application  of the  AIT to Puget
Sound  utilizes  data  on  benthic infaunal  assessments,  amphipod
sediment bioassays, oyster larvae bioassays and Microtox«-RJ bio-


                           8

-------
assays.   The AET  can theoretically  incorporate  any biological
effect(s) endpoint.  In the  generation of  the  AET,  a battery of
biological effects endpoints  should be used,  in the selection of
biological effects endpoints  one should considers

     1.  The range of organismal responses  to the contaminated
         sediments.
     2.  The methods used  for assessment  of acute and chronic
         toxicity.

     The AET  incorporates the  first  consideration  by requiring
multiple biological effects enuSpoints.   In the Puget Sound data
base, the latter concern was also addressed by including benthic
infaunal assessments.   However,  if  infaunal  data  are  not
available,  the  AET fails  to consider  any estimate  of  chronic
toxicity.  This  conclusion  is  based on  recognizing that  the
infaunal  analyses  are  integrated measurements  of  chronic  and
acute toxieities while the bioassays chosen and available to date
reflect acute toxicities only.

     It  ia  the conclusion of  the subcommittea  that; properly
designed and applied infaunal analyses ar» extremely valuable to
the development and validation of the  AET and that they should be
Includedin  future activities.  Research should be  initiated to
develop  improved toxicitv  assay methods  that can  be  used to
assess long-term impacts of sediment tOKJcity.

     The AET approach factors in both in situ  biological effects
through  the benthic  infaunal  analysis   and  laboratory toxicity
tests through bioassays.  The former measures alteration, but is
prone to all the problems  inherent in  field data.  These  include
natural variability from biotic  (predation, competition)  as well
as  abiotic  factors  (grain-size,  salinity).    As  the  method is
presently  being used,   few ecological  endpoints  are  being
measured*   A  50%  reduction  in major  taxa  is  a  gross change
indicative of major adverse effects. The  use of major taxa rather
than species changes may  be  correct,  but it  is heavily weighted
to  serious  acute  effects.   It  may incorporate chronic and more
subtle changes, but these may be masked by the overriding lethal
effects,  even though chronic  effects  can  also  result  in
lethality.

     Regarding bioassays,  both acute and chronic tests should be
used.  The  later  is defined  as  a test that encompasses at least
one full life-cycle of  an organism.   The more tests the better,
given the value of  a  preponderance of evidence approach, but it
is  recognized that  there  are always  going to  be limitations to
resources.  It  ia  recommended that not only different organisms
be used but also  that different oxpoaure  routes and  life-stages
bo  used.    It is  important that  those using AET  clearly and
precisely specify  what  each biological  parameter  measured is
assumed  to  mean.   For  instance,  it   is  very  difficult from the
data presented to  determine what MicrotoxW  is  supposed to
indicate.   If the test is  intended  to  indicate  impacts on the
procaryotic level of the  community, then its  use in sediment is

-------
 probably  inappropriate since it  is based  on  photoluminescence  by
 aerobic organisms that are not normally inhabitants of sedinents,
 which  are usually  anaerobic.  If  it  is  used  as an  empirically
 derived  indicator  of  toxicity,  then  the  choice  is  more
 reasonable.   The need . for clearly defining the assumptions  is
 most important  for the potential users  of both  the  AET  value and
 the methodology.

     The  biological indicators used  in Puget Sound for the
 development  of  AET  values are some  of  the  best tools  presently
 available,  but  this  is  a  rapidly evolving field.   At  present
 there  is  a  need for  better sediment  bioassays,   in  particular
 chronic  bioassays.    We  are  now using  first  generation  (water
 bioassays adapted   for  sediment, usually  through the  use  of
 elutriates) and  second generation  bioassays  (acute bioassays
 specifically  adapted   for use  with  whole  sediments),  The
 developers  of  the  AET approach  in  Puget  Sound  are  to  be
 complimented for  incorporating  a range of biological  evaluation
 teata_. in their approach.   The  fact that  there   may  be  other
 methods or even better ones does not detract from the effort.   As
 stated above, there  is  a definite need to incorporate  more than
 one assay and the concordance of the  results provides reassuring
 evidence  for effects.
 3.6  ComplexMixtures

     In  some  circumstances,  the AET method has the  potential  to
 assess  the impact  of  complex chemical  mixtures on  indigenous
 benthic  infauna.  This  is  because  organisms utilized in the bio-
 assays and organisms enumerated in the environment are exposed to
 the array of  chemicals  incorporated  in the sediment  of question.
 If  the composition of  the complex  mixture  responsible for  the
 measured biological effects  remains  relatively  constant in sedi-
 ments  throughout  the  area of  interest  (i.e.,  that portion of  a
 water  body to  which  an AET  value is to be  applied),  then  the
•predictive potential of the  AET is increased.   It is also  neces-
 sary  that  the  chemicals measured be those  responsible for  the
 biological effects or else that they act as surrogates for those
 that are the  causative  agents.  In some  situations,  such  as  im-
 pacts  from PAH, these  criteria may be  met,   but in situations
 where  the  complex  mixture  changes  in  composition  rather  than
 concentration or where  the chemicals measured do not vary propor-
 tionally with the concentrations  of the  substances responsible
 for  the effects, AETs  will be  faulty  predictors of biological
 impacts of complex nixtures.

     Cause/effect relationships,  as pointed  out  in  section  3.4
 and  in the briefing document  provided to the SAB,  are not well
 defined  by the AET  approach.    This  is because  there  is  no
 certainty  that the chemical  substances  measured are,  in fact,
 those  responsible for the  noted biological effect.  A chemical(s)
 which  appears to be the causative  agent may only co-exist and" co-
 vary with the toxicants rather than be  involved in the toxicity.
                           10

-------
In  such cases,  one may regulate  surrogates  rather  than the
toxicants.

     Even with these  limitations,  there are definite advantages
associated with  this  or similar iwiltivariate methodologies  over
univariate methodologies  which consider  only  single chemicals.
These advantages  are  that additive,  synergistic or  antagonistic
effects can be taken into account, as long as the  composition  of
the contaminants remains constant.
3.7
             Factors
     Attention should be given to the effects of physical factors
on biological response (s) to  chemicals  since many areas  of  con-
cern are in  environments  that are highly dynamic.  It should  be
noted, however, that many other areas are relatively placid  and
physical  factors  are  less   important.  The  currents ,  turbulent
mixing, and dispersion of sediments  generally vary  significantly
with time  and space according  to the  tides,  winds,  waves,  and
freshwater discharges.   All  of these factors can have significant
impacts on biological response (s)  to  chemicals.   For  example,
variation in currents,  salinity, turbulence level, sediment char-
acteristics  (mineralogical  composition, size distribution) ,  and
dissolved oxygen levels can all affect the biota.

     The Puget Sound study concentrated on the  chemical  and  bio-
logical data and used little  or  no physical  data (currents,  sal-
inity, turbulence, and sediment  characteristics)  in the  develop-
ment of  AIT.   Until  the effect of physical  factors  on  AET  is
adequately studied,  the present  AET  values  could contain signif-
icant  errors and  the AET  cannot  be  applied  generically  with
confidence .

     H comprehensive study  ghould be conducted  to  determine  the
of feet of  physical  factors  on biological response  to  chemicals.
It ia  yjtftammflnneg that hydyegynMii e »|id. sedinenfcary  information
(via monitoring or modeling  effort) be collected and   used  to
  lide
	i« select ion of &BT ^^mplincf stations and reference sites.
 Such effort will not only provide  a  scientifically  more defens-
ible basis for  site  selection,  but may also lead to a reduction
in the number of required sampling sites*
3.8  Carbon normalization

     A large amount  of data exists in the  literature,  from both
laboratory and field studies, demonstrating the utility of carbon
normalization for relating the bioavailability  of non-ionic
organic chemicals  sorbed to  sediments.  These  data  indicate that
the  free  form of  the  chemical which  is  available  for organism
uptake, whether  by ingestion or by transport  across respiratory
and  external membranes,  can best be approximated by  carbon
normalization of the measured sediment chemical concentrations.
This also tends to reduce variability in the data.
                           11

-------
     The information presented on the AET approach indicates  that
no improvement in the  sensitivity  of the method was achieved,  by
carbon normalization of the data for the individual stations  with
subsequent recalculation of the AET values.  Several reasons may
exist for this lack of change in sensitivity*

     a)  Most  of the sediment chemical concentration  data  have
organic  carbon values in the range of 1-3%.   This is not a  wide
range,  and therefore one would not expect major changes  in  the
normalized values.
     b)  Chemical  concentrations on sediments do not always cor-
relate well with organic carbon in areas impacted by massive dis-
charges of chemicals above their water solubilities, i.e.,  due to
spills.   In such situations it is possible to find high  concen-
trations of chemicals on sediments which are very low in carbon.

     The  Subcommittee ree**"""«>"*3  that  carbon normalization  be
used  to develop  the  proposed AET  values.   The use  of  this
approach is consistent with theory and it provides an AET which
is based  on  mass of chemical  per  mass of carbon.   In  practice,
converting a  carbon-normalized AST for a  specific chemical into
non-normalized concentrations results in a range  of (AET)  values
depending  upon  the  value  of  carbon  that  is used  to  make  the
conversion.   For  example,  the carbon-normalised  Microtox^J  AET
for hexachlorobenzene  is listed  as 2.3 mg/Kg  C  (USEPA,  Briefing
Report  to the  EPA  Science Advisory  Board,  page  35,  Table  4,
September, 1988)   To  interpret environmental sediment  chemical
concentrations in relation  to  this normalized  AET (2,3  mg/Kg C),
the non-normalized sediment  chemical  concentration would  have to
be divided by its respective sediment total  organic carbon (TOC)
content.      This   means    that   a   non-normalized  sediment
concentration, which is what is  usually determined analytically,
would  exceed  the normalized  AET value (2.3  mg/Kg  C)   if  it
contained  0.02  mg/Kg hexachlorobenzene and  a TOC  of 1.0% (0.20
mg/Kg/1% TOC =20 mg/Kg C).  However, if the same sediment sample
had a TOC  of  10%,  it would  not exceed  the carbon normalized AET
(0.20 mg/Kg/10% =2.0 mg/Kg C).

     Carbon normalization does not assume  that all sediments are
equal.   The  same sediment  chemical concentration  detected in a
range  of  different  sediments  may or  may  not   exceed  the  AET
depending on the TOC of the sediment.  The use of TOC-normalized
values  in  the above manner should eliminate  the criticism that
the AET approach is  insensitive to differences in sediment types
and differences in bioavailability.


3*9 Uncertainties in the RET

     The AET approach relies on the use of a single AST value for
each chemical and each biological indicator/endpoint*  The  choice
of such single value(s) is  necessarily somewhat  subjective.  Any


                            12

-------
uncertainty and variability of the actual data are not adequately
incorporated into single-valued AET.

     Uncertainties in the data used for  the AET process are based
in  part  upon the variabilities  inherent  in  sampling,  naturally
occurring patchiness in  the distribution of  organisms,  and
similar patchiness in the spatial distribution of chemicals.   It
may be possible  to  display these kinds  of uncertainties through
the use of appropriate sampling strategies.

     Uncertainties that nay be due to inadvertent  biases or due
to  uncontrolled  variables are  much more difficult  to uncover.
The  inclusion of such  uncertainty  may lead  to artificially
elevated AET values.  This factor  is of concern since AET values
(by definition)  can  only  be  elevated  as more data is generated.
Examples of  factors that may  give rise to biased relationships
between the exposure and response variables are;

     a.  Measurement of total  metals, rather than of  indicators
          of bio-available metals.

     b.  Failure  to  consider  the  strength  of sorption  of con-
          taminants to the sediment matrix.

     c.  Estimation of effects based on responses of higher taxa,-
          changes  in lower  taxa would  be  expected  to  be
          more sensitive.

     d.  Bioassays  conducted with homogenized   sediments or with
          supernatants derived from agitated  sediments as opposed
          to undisturbed sediments,

     e.  Use of acute bioassays and consequent lack of considera-
          tion for chronic and genotoxic response.

     Environmental significant parameters  that have  not been
measured may influence uncertainties and biases in unpredictable
ways.  The authors  of the AET methodology clearly were aware of
some of the major categories of important variables that have not
been measured.  Among such neglected variables that may influence
effects are:

     a.  Interactions  between chemicals in  the production of
          effects,

     b.  Matrix effects that influence bioavailability.

     c.  Physical  parameters  that influence  the  distribution of
          chemicals and benthic organisms.

     The AET method correctly assumes  that,  as the exposure vari-
able is  increased,  the response  or effects variable will even-
tually become resolvable.  In the  method, the exposure variables
are maintained  throughout as  individual chemical concentration
measurements*  Although the list of chemicals tested  for  is rela-


                           13

-------
tively small, the list contains those substances that are  of  cur-
rent regulatory concern, and  it  can be increased to  a  broader
list of chemicals in the  future.

     The effects variables  are treated  in a  more simplistic
fashion.   Indicators of effects are  limited to an in situ assess-
ment of four major benthic  taxa, which should provide information
on chronic  and  interactive effects.  In  addition,  the  sediment
samples  are subjected to  one to  three  short-term laboratory
bioassays.    The  observed deviations from  "reference"  conditions
are  then  converted  into quantal  or  affected/non-affected
categories.   This  process  collapses the  continuous data of the
effects  variable  into  two  categories,  losing  most  of  the
information related to the  variability of  effects in the process.
A  further  problem  associated  with the categorization  into
affected vs. non-affected  (non-impacted)  conditions is  based  on
the   selection   of   "reference"   conditions.     The   exact
categorization criteria are not rigorous or self-evident, but are
based  upon  best  professional  judgment.    Use  of  the  term
"inconclusive" rather than  non-affected or  non-impacted may
better indicate the subjective nature of this determination.

     The categorization  process  into  affected/non-affected has
the advantage of providing a  starting point for the development
of single-value delimiters  for the AIT Methodology.  Such single-
value delimiters  have obvious attractiveness  as administrative
tools for setting criteria, standards, or  other  limits.  However,
these values are dependent on the  available data,  their central
tendency, and more  importantly  on   the range and variability of
the available data.  This makes it very important to investigate
the variabilities associated with the development of the exposure
and effect  functions,  to examine the  relationships critically,
and to  incorporate indicators  or measures of  the degree  of
variability as part of any  apparent  effects threshold.

     If no  categorization  step  is   undertaken  initially, and if
the effects function is expressed in terms of the exposure  func-
tion,  then the relationship and its  associated variability can be
described directly,  leaving the need for professional judgment as
the final step.   It should  also be possible to substitute a  form-
alized decision process  for a portion  of  the professional  judg-
ment required in the last step.   Another alternative would be to
display the variability inherent  in  the process  as outlined  above
as an adjunct to the development of the AST along lines similar
to those suggested  in the  present  methodology. The subcommittee
•j-oaiimm^nagtthat a measure of variance  for  AST values ba developed
in addition to tha ainala-valued  statistical mean AET.
                           14

-------
                   committee
a. The Subcommittee recommends  that  future application of the A£T
include replicate sediment samples for determination of chemical
contaminants and for assessment of toxic ity*

b.  The Subcommittee recommends that  criteria  for  selecting
reference sites be formalized.    The selection/rejection criteria
need to  be  clearly defined  and the rationale  for  their choice
explained.

c.  Some  knowledge  of  how  temporal changes  impact  the AET  is
needed to characterize the uncertainty that may be added by this
variable.

d. The Subcommittee  recommends that multiple approaches be used
to  estimate sediment quality,  determine criteria and  guide
regulatory  actions since  the  AET  approach alone  provides
insufficient  certainty  for broad-scale  decision  making.

e. It  is  recommended that the data set be  reviewed  for all the
AETs in relation to their physical-chemical  properties  (e.g. KQC,
Log  P  and  water solubility)  and  their  existing toxicological
properties  (e.g.,  acute  versus chronic or  genetic  effects)  for
the  purpose of  establishing  the sensitivity  of  AET  values  to
parameters that are known to affect  toxicity.

f .  It is recommended that the AET values be compared to existing
toxicity values for  aquatic organisms   for the  purpose  of
identifying those  AET values  which  are  inconsistent  with  the
existing body of toxicological data.

g,   It is recommended that  additional research be  conducted to
evaluate  AETs  relative  to the  applicable,  available  data  for
dose-response relationships*

h.  it  is the conclusion  of  the  Subcommittee  that  properly
designed and applied infaunal analyses are extremely valuable to
the development and validation of the  AET and that they should be
included  in future activities. Research should be  initiated to
develop  improved toxicity  assay methods  that can  be  used  to
assess long-term impacts of sediment toxicity.

i. It  is recommended that not  only different organisms be used
but also that different exposure routes and life-stages be  used.

j . A comprehensive study  should be conducted  to determine the
effect of physical  factors on biological response to  chemicals.
It is  recommended  that hydrodynamic and sedimentary information
(via monitoring  or  modeling effort)  be  collected  and used to
guide the selection of AET sampling  stations and reference  sites.

k. The Subcommittee recommends that carbon  normalization be used
to develop the proposed AIT values.
                           15

-------
1. The Subcommittee recoawnends that a measure of variance for AET
values be developed in  addition  to the single-valued statistical
mean AET.
                            16

-------
                   APPENDIX
Briefing Report.to the EPA Science Advisory Board;
     The Apparent Effects Threshold
                   17

-------
BRIEFING REPORT TO THE
EPA SCIENCE ADVISORY BOARD:
THE APPARENT EFFECTS THRESHOLD APPROACH
Submitted by

Office of Puget Sound
Puget Sound Estuary Program
U.S. Environment!! Protection Agency, Region 10
1200 6th Avenue
Seattle, Washington 9SIOI
Prepared by

PTI Environmental Services
3625 132nd Avenue SE
Suite 301
Beilevue, Washington  98006
under Battelle Columbus Division
EPA Contract No. 68-03-3534
PTI Contract No, C714-01
September 1988

-------
                             .2. THE CONCEPT OF AET


     An  AET  is defined  as  the sediment concentration  of a given chemical  above  which
statistically  significant  (P
-------
           otherwise the AET is  only a  preliminary minimum estimate (or may  not
           exist).

     5,    Repeat Steps I -4 for «ach biological indicator.

     A  pictorial   representation  of   the  AET approach  for  two  example  chemicals  is
presented  in  Figure 2  based on  results  for  a toxicity bioassay.   Two  subpopulations of
all  sediments  analyzed  for chemistry  and  subjected  to a bioassay  are  represented  by
bars in the figure and include:

     •     Sediments  that   did  not  exhibit  statistically  significant  (P>0.05)  toxicity
           relative  to reference conditions ("nonimpacted" stations)

     •     Sediments   that   exhibited  statistically   significant  (P
-------
                                 LEAD
                NO SEDIMENT TOXIC1TY
                      SEDIMENT  TQXICITY  QBSEBVE0
        S(M5
                                                          6300 pprn
         r
        to
                           too
T
                                             tooo
                                       APPARENT
          CONCENTRATION (mg/Kg DW)  TOX!CrfY
                                       THRESHOLD
         MAXIMUM
         OBSERVED
         LEVELATA
         BIOLOGICAL
         STATION
J
                                                               10,000
                    4-METHYL  PHENOL
              NO SEDIMENT TOXICITY
                                                    OH
                        -SEDIMENT
      ./:/• •/.. /////////
     U10
                                     APPARENT
        CONCENTRATION (pi/Kg DW)  TQXOTY
U * undetected at detection limit shown
                                                                   W.OOO
                 MAXIMUM
                 OBSERVED
                 LEVEL AT A
                 BtOtOGICAL
                 STATION
Figure 2.   The AET approach  applied to  sediments tested for lead and 4-methyl phenol
          concentrations and toxicity response (Jnriflg tuoassays.

-------
other sites  without associated  biological effects).   Based  on the  results for  lead  and 4-
raethylphenol, effects  at  4  of the 28 impacted sites shown in the  figures may  be  associated
with  elevated  concentrations of 4-methylphenol,  and effects  at  7  other  sites  may be
associated with elevated concentrations of lead (or similarly distributed contaminants).

     These  results illustrate  that   the  occurrence of  biologically  impacted  stations  at
concentrations  below  the AET  of a single chemical  does not  imply  that AET in general
are  not  protective against biological effects, only that  single  chemicals may  not  account
for  all  stations  with  biological  effects.   By  developing  AET  for multiple  chemicals,  a
high percentage  of all   stations  with  biological effects  are accounted for  with  the  AET
approach (reliability results are presented in  Section  3 of this briefing document).

     AET can  be expected  to be  more predictive when developed from a large, diverse
database with  wide ranges of  chemical concentrations and a  wide diversity of measured
chemicals.    Data  sets   that  have  large  concentration gaps between  stations and/or do
not  cover  a wide range Of concentrations must  be  scrutinized carefully (e.g., to discern
whether   chemical  concentrations   in  the   data  set  exceed   reference   concentrations)
before generation of AET is appropriate.
Dose-Response Relationships and AET

      The  AET  concept  is consistent  with  empirical  observations  in  the  laboratory  of
dose-response  relationships  between  increasing concentrations of  individual toxic chemicals
and  increasing biological  effects.  A simple  hypothetical  example  of such single-chemical
relationships is shown  for  chemicals X  and Y in Figure  3.   In  the example,  data rare
shown for  laboratory exposures  of a test organism to sediment  containing only increasing
concentrations  of  chemical  X,  and  independently, for exposures  to  sediment containing
only  increasing concentrations  of chemical  Y.   The  magnitude of  toxic  response in  the
example  differs  for  the  two  chemicals  and   occurs  over  two  different   concentration
ranges.   It is  assumed  that at some level  of response, for example >25 percent,  the two
different  responses   can   be   distinguished  from  reference  conditions  (le.,   responses
resulting from  exposure  to sediments containing very  low or  undetectable  concentrations
of any toxic chemicals).

      These single-chemical  relationships  cannot  be proven  in  the  field  because organisms
are  exposed to  complex mixtures of chemicals  ia  environmental  samples.    In  addition,
unrelated  discharges  from   different sources can result  in  uncorrelated  distributions of
chemicals   in  environmental  samples.    To  demonstrate  the   potential  effects  of  these
distributions,  response data  are  shown  in Figure 4 for a  random  association of  chemicals
X and Y using  the  same concentration  data as  in Figure  3.   The data  have been plotted
according   to  increasing  concentrations  of  chemical  X,  and  the  same   dose-response
relationship observed  independently  for the  two chemicals  in  the  laboratory  has  been
assumed.   The  contributions  of  chemicals X and  Y to  the toxic  response  shown  for
these  simple mixtures  is  intended  only for  illustration purposes to enable direct comparison
to the relationships  shown  in  Figure 3, but are analogous to  an additive toxic response,
Other interactive effects are not considered in this example.

      In  Figure  4,  a  significant response  relative  to reference  conditions  would  result
whenever  elevated concentrations  of either chemical  X  or  chemical  Y  occurred  in  a
sample.   Because of  the  random association of Y with X  in  these  samples,  the significant
responses   would   appear   to  occur  randomly  over   the   lower  concentration  range  of

                                              12

-------
Bioassay Response
 100
                         Chemical X
                         Chemical Y
    0
         Significant
Toxiclty     / *~~~'T p
             Increasing X or  Y	>
 Figure 3.  Hypothetical example of dose-response relationship resuliing  from laboratory
        exposure to sin&te chemicals X and Y,

-------
Bioassay  Response
  100
      0
                                             Chemical X Q


                                             Chemical Y


          Significant Toxicity    -f-    +O"
                                      i
                                     4
                                      +   +
                 Increasing
Figure ^f.   Hypothetical example of toxic response resulting from exposure lo environmental

        samples of sedimeni contaminntetl with chemicals X ami Y.

-------
chemical X.   The  classification  of the responses  shown  in  Figure  4 into significant and
nonsignificant groups (i.e., >25 percent response  for  either chemical) results in generation
of Figure 5.                 .

     Figure  5   represents  the  appearance  of  the  environmental  results  when   ranked
according 'to concentration of chemical X using these data.   Below  the AET for  chemical
X,  significant  toxicity  is  produced  by  elevated  concentrations  of  chemical  Y, which  is
randomly associated  with the  distribution of  chemical  X.   Above  the  AET for  chemical
X,  significant   toxicity  is  always produced  by  elevated concentrations  of  chemical  X,
although in  some  samples,  elevated  concentrations of  chemical  Y  also contribute  to the
overall  toxicity.   The  AET  for chemical  X  corresponds  conceptually,  in  this  simple
example, to  the concentration in  Figure  3  at which a significant  difference  in  response
was observed in the laboratory for chemical X,

     In  environmental  samples that  contain complex  mixtures  of chemicals, a mono tonic
dose -response relationship  such  as  in  this  simple two-chemical example  may  not  always
apply.    For example, a consistently increasing  biological response   may  not  always occur
at  increasing concentrations  of  a chemical  above  Its  AET,    Such  observations could
indicate  that the  AET is  coincidental  (i.e.,  that the  observed toxicky  in some  or  all
samples  above   the  AET is  unrelated  to  the presence of  that chemical),  or that  changing
environmental  factors in samples exceeding an  AET  obscure  a monotonic  dose-response
relationship.  Such  factors are discussed  in the following section.


Influence of Environmental Factors on AET Interpretation
                                 " '
      Although  the  AET concept  is  simple,  the  generation  of  AET  values  based  on
environmental   data  incorporates   many  complex   biological-chemical  interrelationships.
For  example,  the  AET  approach  incorporates the  net  effects  of  the  following  factors
that may be important in field-collected sediments:

      •    Interactive effects of chemicals (e.g., synergism, antagonism, and additivity)

      ii    Unmeasured chemicals and other unmeasured, potentially adverse variables

      •    Matrix   effects   and   bioavailability   [Le.,  phase   associations   between
           contaminants and sediments that affect bioavailability of the contaminants,
           such  as the  incorporation of  polycydic  aromatic hydrocarbons (PAH)  in
           soot particles].

      The  AET  approach cannot distinguish and  quantify  the contributions  of  interactive
effects,  unmeasured  chemicals, or  matrix  effects   in  environmental  samples,  but  AET
values  may  be  influenced  by  these  factors.   To   the  extent that the  samples   used  to
generate  AET  are representative of samples  for  which AET  are  used to predict effects,
the  above environmental factors  may  not detract from the  predictive  reliability of  AET.
Alternatively,  the  infrequent  occurrence  of the above  environmental factors  in a  data
set  used   to  generate  AET could  detract  from  the predictive  reliability  of those AET
values.   If  confounding environmental  factors render the  AET approach unreliable, this
should be evident  from  validation  tests  in which  biological  effects  are  predicted  in
environmental  samples.   Tests of  AET  values  generated  from  Puget  Sound  data (see
Section  3)  indicate   that   the  approach  is  relatively  reliable  in  predicting  biological
effects despite the potential uncertainties of confounding environmental factors.

                                             15

-------
Bioassay Response
100
  0
        Increasing X or Y
 Significant Toxicity 1 **
      No Toxicity
                Bioassay Response
                100
                  0
                      Increasing X
                       AET
                                               >
                   *  *     *  ****tt
I
oo  o  ooo  o  o
                              Increasing X
 Figure 5.  Hypothetical example of AET calculation for chemical X bused on classified ion
          of significant and nonsignificant  responses for  environmental samples con-
          taminated  with both chemicals X and  Y.  [Previous Figures 3 ami  A are also
          shown for comparison; dashed line indicates level of significant toxiehyj

-------
     Although  the above  environmental  factors  can influence the  generation of  field-
based  sediment quality values such  as  AET, they  also may influence the  application  of
all  sediment  quality  value  approaches  for  the prediction  of adverse  biological  effects.
For  example, sediment quality  values based on  laboratory sediment bioassays spiked  with
single  chemicals  would  not  be susceptible  to   the  effects  of  the environmental  factors
listed above.   However, in applying such  values to  field-collected samples, predictions  of
biological effects  could be less  successful  to the extent that  interactive effects,  unmeasured
chemicals, and matrix effects  occur in the environment.

     The nature  of the relationships between  AET values and confounding  environmental
factors is discussed in the remainder of this section.
      Interactive Effects  and  AET—AET uncertainty  is  increased  by  the  possibility  of
interactive  effects;  the increase in  uncertainty  5s expected  to  be  less pronounced  when
large data  sets collected  from  diverse  areas  are  used  to generate  AET,   Additivity and
synergism  can produce a  comparatively low AET for a given chemical by causing impacis
at  concentrations  that would  not  cause  impacts  in  the  absence  of   these  interactive
effects.   This would effectively reduce  the pool  of  nonirnpacted  stations  used to generate
AET.   This  effect  should  be  reduced if a  diverse  database is  used such that chemicals
occur  over  a  wide range  of  concentrations  at  stations  where  additivity  and  synergism
are  not  operative.   For  chemicals  that covary  regularly  in  the  environment (e,g.,  fluor-
anthene  and  pyrene),  even  a  large,  diverse database  will not  reduce the  effects  of
additivity  and/or  synergism  on  AET  generation.   The  resulting  AET  values  for such
chemicals   may  be  reliable  in  predicting   biological   effects  in  environmental  samples
although not representative of the toxidties of the chemicals acting independently.

      Antagonism will  produce comparatively  high  AET values if (and only if) the AET is
established  at a station  where antagonism occurs,  A  large, diverse database  could  not
rectify  this  elevation  of AET  if  the  station at  which  antagonism  occurred  was   the
nonimpacted  station  with  the  highest  concentration  (i.e.,  the   station setting  an   AET),
An  AET set  by a  station  at which antagonism  occurred  would not be  representative of
the  toxieity of  the  chemical acting  independently.  Hence,  if  antagonism did  not occur
widely, such antagonistic effects  would  cause  the AET to be  less sensitive  in  predicting
adverse effects in the environment.

      Empirical approaches  such  as  the  AET do  not  provide a  means  for  characterizing
interactive  effects.    Only   laboratory-spiked  sediment bioassays offer  a systematic  anci
reliable  method for  identifying  and quantifying additivity, synergism,  and antagonism.
A  great  deal  of  research   effort  would  be  required  to  test  the  range  of  chemicals
potentially  occurring   in  the  environment   (both  individually  and  in   combination),  a
sufficiently  wide range of  organisms, and a  wide range of  sediment matrices  to  establish
criteria.     In  addition,   the  applicability of  bioassays  conducted  with  laboratory-spiked
sediments to environmentally-contaminated sediments requires further testing.


      Unmeasured  Chemicals  and AET—Another source of uncertainty for AET and other
field-based approaches is the possibility  of effects being  caused by unmeasured,  covarying
chemicals.   Such  chemicals  would  not  be  expected  to  substantially  decrease  the  ability
of  AET  to  predict biologically  impacted stations (excluding interactive  effects  discussed
above).   If  an unmeasured  chemical  (or  group  of chemicals)  varies  consistently  in  the

                                              17

-------
environment  with  a  measured chemical,  then  the  AET  established  for  the  measured
contaminant will  indirectly apply  to,  or result  in  the management  of,  the  unmeasured
contaminant.   Jn such  cases,  a measured  contaminant would act as  a surrogate  for an
unmeasured  contaminant (or  group of  unmeasured  contaminants).   Because  all  potential
contaminants  cannot  be, measured  routinely,  management  strategies   must  rely  to  some
extent on "surrogate" chemicals.

      If  an  unmeasured  toxic chemical (or  group  of chemicals) does  not  always covary
with  t   measured  chemical (e.g.,  if  a  certain  industry  releases  an  unusual  mixture  of
contaminants), the  effect  should   be  mitigated  if a  sufficiently  large  and  diverse  dara
set is used  to  establish  AET,    Use  of  a large  data  set comprising  samples  from  a
variety  of  areas with  wide-ranging chemical concentrations  would decrease  the  likelihood
that  an  unrealistically  low AET   would be  set   Because  AET  are  set  by  the highest
concentration of  a  given  chemical in  samples without observed  biological effects.  AET
will  not  be  affected  by  less  contaminated  samples  in  which unmeasured  contaminants
cause biological effects.

      If an  unmeasured  toxic chemical does not covary with any of the measured  chemicals,
it   is  unlikely  that  the AET  (or any  other chemical-specific  approach)  could  predict
impacts  at  stations  where  the chemical is   inducing  toxic  effects.    The  frequency  of
occurrence  of stations  with  biological  effects  but no  chemicals  exceeding  AET is  the
subject of  validation tests (see Section 3 of this briefing document),


      Matrix  Effects  and  BioaYailabiIity~-Geochtmical associations  of contaminants  with
sediments  that  reduce bioavailabiUty of  those  contaminants  would affect  AET analogously
to  antagonistic effects  (i.e.,, they  would  increase  AET  relative  to  sediments  in  which
this  factor  was not  operative).   Sediment  matrices  observed  in Puget  Sound thai  may
reduce   bioavailabiUty  of   certain  contaminants   include  slag   material  (containing  high
concentrations of  various  metals   and  metalloids,  such  as  copper and arsenic)  and  coal
or soot  (which  may  contain  high  concentrations of largely  unavailable PAH, as  opposed
to  oil or  creosote,  in  which  PAH would  be  expected  to  be  far  more  bioavailable;  e.g.,
Farrington  and Teal  1982).   Many kinds  of matrices  may  occur  in  the environment and
a   large  proportion  may  be  difficult   to  classify  based  upon appearance   or   routinely
measured  sediment variables.   Hence,  the  use  of  matrix-specific  data sets to  generate
AET, although desirable, would be difficult to implement.

      To  address this  concern from  a technical perspective   (Le,,   representativeness  of
data  used   in  AET generation), the AET  database could be screened for  sediment with
chemical concentrations that  are  anomalously high relative  to those in other  nonimpacted
sediments   from  different  geographic   areas.-    From  a   management  perspective,  this
guideline  would  generate  more  protective  (sensitive)  sediment  quality  standards  that
may  also   be less  efficient  in  only  identifying  problem   sediments.   These sediments
would be  considered nonrepresentative  and  not  used in  AET  generation  unless  and  until
additional   data  could  substantiate  that  they are  representative.     Such  data   treatment
methods are discussed in the following section.

-------