UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                        WASHINGTON, D.C. 20460
                            FEB 221994,
                                                       OFFICE OF
                                                        WATER
                                                 EPA-823-B-94-001
MEMORANDUM

SUBJECT:


FROM:


.TO:
Use of the Mater-Effect Ratio
Standards
Tudor T. Davies, Director
Office of Science and Technology-

Water Management Division Directors, Regions I - X
State Water Quality Standards Program  Directors
 PURPOSE

      There are two purposes for this memorandum.

      The first is to transmit the Interim Guidance on the
 Determination and Use cf Water-Effect Ratios for Metals.  EPA
 committed to developing this guidance to support implementation
 of federal standards for those States included in the National
 Toxics Rule.           ;              •

      The second is to provide policy guidance on whether a
 State's application of-a water-effect ratio is a site-specific
 criterion adjustment subject to EPA review and
 approval/disapproval.       ,


 BACKGROUND    '                  -••/'."-•".

      In the early 1980's, members of the regulated community
 expressed concern that EPA's laboratory-derived_water quality
 criteria might not accurately reflect site-specific  conditions
 because of the effects of water chemistry and the ability  of
 species to adapt over time.  In response to these concerns,  EPA
 created three procedures to derive  site-specific criteria.   These
 procedures were published in the water  Quality  Standards
 Handbook. 1983.
                                                          Printed on Recycled Paper

-------
     . Site-specific  criteria are allowed by regulation and are
 subject to EPA review and approval.  The Federal water quality
 standards regulation-at section 13l.ll(b)(l) provides states with
 the opportunity to  adopt water quality criteria that are
 "...modified to- reflect site-specific conditions."  Under section
• 131.5(a)(2), EPA reviews standards to determine "whether a State-
 has adopted criteria to protect the designated water uses."

      On December 22, 1992, EPA promulgated the National Toxics
 Rule which established Federal water quality standards for 14
 States which had not met the requirements of Clean Water Act
 Section 303 (c) (2) (B).  As part of that rule, EPA gave the States
 discretion to adjust the aquatic life criteria for metals to
 reflect site-specific conditions through use of a water-effect
 ratio.   A water-effect ratio'is a means to account for a
.difference between the toxicity of the metal in laboratory
 dilution water and its toxicity in the water at the site.

      In promulgating the National Toxics Rule,  EPA committed to
 issuing updated guidance on the derivation of water-effect
 ratios.  The guidance  reflects new information since the
 previous guidance and is more comprehensive in order to provide
 greater clarity and increased understanding..   This new guidance
 should  help standardize procedures for deriving water-effect
 ratios  and make results more comparable and defensible.

     Recently,  an issue arose concerning the most appropriate
 form of metals upon which to base water quality standards.   On
 October 1,  1993,  EPA issued guidance on this issue which •
 indicated  that measuring the dissolved form of metal is the
 recommended approach.   This new policy however,  is prospective
 and  does not  affect the'criteria in the-National Toxics Rule'.
 Dissolved  metals  criteria are not generally numerically equal to
total recoverable criteria and the October 1,  1993..guidance
 contains recommendations for correction factors for fresh, water
 criteria.   The determination of site-specific criteria is
applicable to criteria  expressed as either total recoverable
metal or as dissolved metal.                              -

DISCUSSION                .      .              .   •    •   - .

     Existing guidance  and practice are that  EPA will approve
site- specific criteria  developed  using appropriate procedures.
That policy continues for the  options  set  forth  in the interim .
guidance transmitted -today,  regardless  of  whether the resulting
criterion  is  equal  to or more  or. less  stringent  than the EPA
national 304(a) guidance. .  This  interim guidance supersedes all
guidance concerning  water-effect .ratios previously issued by the
Agency.                                            •

-------
      Each of the three options', for-"deriving* a final water-effect
 ratio presented in this interim guidance meets the scientific and
 technical acceptability test for deriving site-specific criteria.

 Option 3 is the simplest, least restrictive and generally the
• least expensive approach for situations where simulated
 downstream water appropriately represents a "site."  It is a
 fully acceptable approach for deriving the water-effect ratio
 although it will generally provide a lower water-effect ratio
 than the other 2 options.  The other 2 options may be more costly
 and time consuming if more than 3 sample periods and water-effect
 ratio measurements are made, but are more accurate, and may yield
 a larger, but more scientifically defensible site specific
 criterion.

       Site-specific criteria, properly determined, will fully
 protect existing uses.  The waterbody or segment thereof to which
 the site-specific criteria apply must be clearly defined.   A site
 can be defined by the State and can be any size, small or large,
 including a watershed or basin.  However,  the site-specific
 criteria must protect the site as a whole.  It is likely to be
 more cost-effective to derive any site-specific criteria for as
 large an area as possible or appropriate.   It is emphasized that
 site-specific criteria are ambient water quality criteria
 applicable to a site.  They are not intended to be direct
 modifications to National Pollutant Discharge Elimination System
 (NPDES)   permit limits.  In most cases the «site" will be
 synonymous with a State's "segment"  in its water quality  .
 standards.  By defining sites on a larger scale, multiple
 dischargers can collaborate on water-effect ratio testing and
 attain appropriate site-specific criteria at a reduced cost.

      More attention has been given to water-effect ratios
 recently because of. the numerous discussions and meetings on the
 entire question of metals policy and because WERs were
 specifically applied in the National Toxics Rule.   In comments on
 the proposed National Toxics Rule,  the public questioned whether
 the EPA promulgation, should be based solely on the total
 recoverable form of a metal.  For the reasons set forth in the
 final preamble,  EPA chose to promulgate the criteria based on the
 total recoverable form with a provision for the application of a
 water-effect ratio.   In addition,  this approach was chosen
 because  of the unique difficulties of attempting to authorize
 site-specific criteria modifications for nationally promulgated
 criteria..    •    .       .'"          ,                              •

      EPA now recommends the use of dissolved metals for States
 revising their water quality standards.   Dissolved criteria may
 also be  modified by a site-specific  adjustment.

-------
      While the regulatory application of the water-effect ratio
 applied only to the 10 jurisdictions included in the final
 National Toxics Rule"for aquatic life metals criteria,  we
 understood that other.States would be interested in applying WERs
 to their adopted water quality standards.  The guidance upon
 which to base the judgment of the acceptability of the water-
' affect ratio applied by the State is contained in the attached
 Interim Guidance on The Determination and Use of Water-Effect
 Ratios for .Metals. It should be noted that this guidance also
 provides additional information on the recalculation procedure
 for site-specific criteria modifications.       ,

 Status of the Water-effect Ratio fWER) in non-National  Toxics
 Rule States

     .A central question concerning WERs is whether their use by a
 State results in a site-specific criterion subject to EPA review
 and approval under Section 303(c)  of the Clean Water Act?

      Derivation of a water-effect ratio by a State is a site-
 specific criterion adjustment subject to EPA review and
 approval/disapproval under Section 303(c).  There are two options
 by which this review can be accomplished.

      Option 1:  A State may derive and submit each individual
      water-effect ratio determination to EPA for review and
      approval.  This would be accomplished through the  normal
      review and revision process used by a State.

      Option 2:  A State can amend its water quality standards to
      provide a formal procedure which includes derivation of
      water-effect ratios, appropriate definition of sites,   and
      enforceable monitoring provisions to assure that designated
      uses' are protected.  Both this procedure and the resulting
      criteria would be subject to full public participation
      requirements.  Public review of a site-specific criterion
      could be accomplished in conjunction with the public review
      required for permit issuance.  EPA would review and
      approve/disapprove this protocol as a revised standard once.
      For public information, we recommend that once a year the
      State publish a list of site-specific criteria.

      An exception to this policy applies to the waters  of the
 jurisdictions included in the National Toxics Rule.  The EPA
 review is not required for the jurisdictions included in the
 National Toxics Rule where EPA established the procedure for. the
 State for application to the.criteria promulgated.  The National
 Toxics Rule was a formal rulemaking process with notice and
 comment by which EPA pre-authorized the use- of a correctly
 applied water-effect ratio.  That same process has not  yet taken
 place in States not included in the National Toxics Rule.

-------
However, the National Toxics. Rule does not affect State authority
to establish scientifically  defensible procedures to determine
Federally authorized WERs, to certify those WERs in NPDES permit
proceedings, or to deny their application based on the State's
risk management analysis

     As-described in Section."131.36 (b) (ill)- of the water quality
standards regulation (the official regulatory reference to the
National Toxics Rule), the water-effect ratio is a site-specific
calculation.  As indicated on page 60866 of the preamble to the
National Toxics Rule, the rule was constructed as a rebuttable
presumption. The water-effect ratio is assigned a value of .1.0
until a different water-effect ratio is derived from suitable
tests representative of conditions in the affected waterbody.  It
is the responsibility of the State to determine whether to rebut
the assumed value of 1.0 in the National Toxics Rule and apply
another value of the water-effect ratio in order to. establish a
site-specific criterion.  The site-specific criterion is then
used to develop appropriate NPDES permit limits.  The rule thus
provides a State with the flexibility to derive an appropriate
.site-specific criterion for specific waterbodies.

     As. a point of emphasis, although a water-effect ratio
affects permit limits for individual dischargers, it is the state
in all cases that determines if derivation of a site-specific
criterion based on the water-effect ratio is allowed and it is
the State that ensures that the calculations and data analysis
are done completely and correctly.

CONCLUSION

     This interim guidance explains and clarifies the use of
site-specific criteria. , It is issued as interim guidance because
it will be included as part of the process underway for review
and possible revision of the national aquatic life criteria
development methodology guidelines.  As part of that review, this
interim guidance is subject to amendment based on comments,
especially those from the users of the guidance.  At the end of
the guidelines revision process the guidance will be issued as
"final."

     EPA is interested in and encourages the submittal of high
quality datasets that can be used to provide insights into the
use of these guidelines and procedures.   .Such data and technical
comments should be submitted to Charles E. Stephan at EPA's
Environmental Research Laboratory at Duluth, MN.  A complete
address, telephone number and fax number for Mr. Stephan are
included in the guidance itself.  Other questions or comments
should be directed to the Standards and Applied Science Division
(mail code 4305, telephone 202-260-1315).

-------
     There is attached to this memorandum a simplified f],ow
diagram and an implementation procedure.  These are intended to
aid a user by placing the water-effect ratio procedure in the
context of proceeding from at site-specific criterion to a permit
limit.  Following these attachments is the guidance itself.

Attachments

cc: Robert Perciasepe, OW
    Martha G. Prothro., OW
    William Diamond, SASD      '
    Margaret Stasikowski, HECD              -
    Mike Cook, OWEC
    Cynthia Dougherty, OWEC
   . Lee Schroer,  OGC
    Susan Lepow,  OGC
    Courtney Riordan, ORD
    ORD (Duluth and Narragansett Laboratories)
    BSD Directors, Regions I - VIII, X
    BSD Branch,  Region IX
    Water Quality standards Coordinators, Regions I - X

-------
 CO
 CD


 CD
DC
LU

-------

-------
                WATER-EFFECT RATIO IMPLEMENTATION

PRELIMINARY ANALYSIS & PLAN FORMULATION     '
          - '   "  * '      -          .         '  "        »      *
     - Site definition                              .

       • How many discharges must be accounted for?  Tributaries?
         See page 17.
       • What is the waterbody type? (i.e., stream, tidal river,
         bay, etc.).  See page 44 and Appendix A.
       • How can these considerations best be combined to define
         the relevant geographic "site"?  See Appendix A § page
         82.     •     '      .       '.                '

     - Plan Development for Regulatory Agency Review

       • Is WER method 1 or 2 appropriate? (e.g., Is design flow
         a meaningful concept or are other considerations
         paramount?).  See page 6.
       • Define the effluent & receiving water sample locations
       • Describe the temporal sample collection protocols
         proposed.  See page 48.
       • Can simulated site water procedure be done, or is
         downstream sampling required?  See .Appendix A.
       • Describe the testing protocols — test species, test
         type, test length, etc.  See page 45, 50; Appendix I.
       • Describe the chemical testing; proposed.  See Appendix Ci
       • Describe other details of study - flow measurement,
         QA/QC, number of sampling periods proposed, to whom the
         results are expected to apply, schedule, etc.


SAMPLING DESIGN FOR STREAMS

     - Discuss the quantification of the design streamflow  (e.g.,
       7Q10) - USGS gage directly, by extrapolation from USGS
       gage, or ?

     - Effluents

      •••• measure flows to determine average for sampling day
       • collect 24 hour composite using "clean" equipment and
         appropriate procedures; avoid the use of the plant's
         daily composite sample as a shortcut.

     - Streams                                                  ..

       • measure flow  (use current meter or read from gage  if
         available) to determine dilution with effluent; and to
         check if within acceptable range for use  of the data
          (i.e., design flow to  10 times the design flow).
       • collect 24 hour composite of upstream water.

-------
LABORATORY PROCEDURES  (NOTE:  These are described an detail in
                       interim guidance).
     - Select appropriate primary & secondary tests

     - Determine appropriate cmcWER and/or cccWER

     - Perform chemistry using clean procedures, with methods
       that have adequate sensitivity to measure low
       concentrations, and use appropriate QA/QC

     - Calculate final water-effect ratio (FWER) for site.
       See page .36.                             .              .

IMPLEMENTATION

     - Assign FWERs and the site specific criteria for each metal
       to each discharger (if more.than one).

     - perform a waste load allocation and total maximum daily
       load (if appropriate) so that each discharger is provided
       a permit limit.             .
 S                      '        ,                .        , .-     ' .
     - establish monitoring condition for periodic evaluation of
       instream biology (recommended)

     - establish a permit condition for periodic testing of WER
       to verify site-specific criterion (NTR recommendation)  .

-------
          United States
          Environmental Protection
          Agency
           Office of Water
           Office of Science & Technology
           (Mail Code 4305)
February 1994
EPA-823-B-94-001
vvEPA
Interim Guidance
on Determination and Use
of Water-Effect Ratios
for  Metals

-------

-------
         Interim Guidance on

      Determination and Use of

   Water-Effect Ratios for Metals
            February 1994
U.S. Environmental Protection Agency

           Office of Water
  Office of Science and Technology
          Washington,  D.C.

 Office of Research and Development
 Environmental Research Laboratories
          Duluth,  Minnesota
     Narragansett, Rhode Island

-------

-------
                             NOTICES


This document has bejen reviewed by the Environmental Research
Laboratories, Duluth, MN and Narragansett, RI -(Office of Research
and Development) and the Office of .Science and Technology  (Office
of Water), U.S. Environmental Protection Agency, and approved for
publication.

              • ' *         .                        •
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.

-------

-------
                             FOREWORD

                 •'                .           . •
 This document provides interim guidance concerning the   .
 experimental determination of water-effect ratios (WERs)  for
 metals;  some aspects of the use of WERs are also addressed.  It
 is issued in support of EPA regulations and policy initiatives
 involving the application of water quality criteria and standards
"for metals.   This  document is agency guidance only.   It does not
 establish or affect legal rights or obligations.  It does not
 establish a  binding norm or prohibit alternatives not included in
 the document.  It  is not finally determinative of the issues
 addressed.  Agency decisions in any particular case will be made
 by applying  the law and regulations on the basis of specific  .
 facts when regulations are promulgated or permits are issued.

 This document is expected to be revised periodically to reflect
 advances in  this rapidly evolving area.  Comments,  especially
 those accompanied  by supporting data,  are welcomed and should.be
 sent to: Charles E. Stephan,  U.S. EPA,  6201 Congdon Boulevard,
 Duluth MN 55804 (TEL: 218-720-5510; FAX: 218-720-5539).
                                111

-------
               UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                          WASHINGTON, D.C. 20460


                             FEB 2 2 1994
                                                        OFFICE OF
                                                         WATER
     •  OFFICE OF SCIENCE AND TECHNOLOGY POSITION STATEMENT

     Section 131.11(b)(ii)  of the water quality standards
regulation  (40 CFR Part -131)  provides  the regulatory mechanism
for a State to develop site-specific criteria  for use in water
quality standards.  Adopting site-specific criteria in water
quality standards is a State option—not a requirement.   The
Environmental Protection Agency  (EPA)  in 1983  provided guidance
on scientifically acceptable methods by which  site-specific
criteria could be developed.

     The interim guidance provided  in  this document supersedes
all guidance concerning water-effect ratios and the Indicator
Species Procedure given in  Chapter  4 of the Water Quality
Standards Handbook issued by EPA in 1983 and in Guidelines for
Deriving Numerical Aquatic  Site-Specific Water Quality Criteria
bv Modifying National  Criteria.  1984.   Appendix  B also
supersedes  the guidance in  these earlier documents for the
Recalculation Procedure for performing site-specific criteria
modifications.

     This interim guidance  fulfills a  commitment made in the
final rule  to establish numeric  criteria for priority toxic
pollutants  (57 FR 60848, December 22,  1992, also known as the
"National Toxics Rule").  This guidance also is applicable to
pollutants  other than  metals with appropriate  modifications,
principally to chemical analyses.

     Except-for the jurisdictions subject to the aquatic life
criteria in the national toxics  rule,  water-effect ratios are
site-specific criteria subject to review and approval by the '
appropriate EPA Regional Administrator.   Site-specific criteria
are new or  revised criteria subject to the normal EPA review
requirements established in Clean Water Act §  303(c).   For the
States in the National Toxics Rule,  EPA has established tjiat
site-specific water-effect  ratios may  be applied to the criteria
promulgated in the rule to  establish site-specific criteria.   The
water-effect ratio portion  of theses criteria would still be
subject to  State review before the  development of total maximum
daily loads, waste load allocations or translation into NPDES
permit limits.  EPA would only review  these water-effect ratios
during its  oversight review of these State programs or review  of
State-issued permits.
                                                                    \
                                iv
                                                          Printed on Recycled Pape

-------
      Each of the three options for deriving a final water-effect
 ratio presented on page 36 of this interim guidance meets the
 scientific and technical acceptability test 'for deriving site-
 specific criteria specified in the water quality standards
 regulation (40 CFR 131.11(a)).  Option 3 is the simplest, least
 restrictive and generally the least expensive approach for
 situations where simulated downstream water appropriately
.represents a "site."   Option 3 requires experimental
 determination of three water-effect ratios with the primary test
 species that are determined during any season (as long as the
 downstream flow is between 2  and 10 times design flow
 conditions.)   The final WER is generally (but not always) the
 lowest experimentally determined WER.   Deriving a final water-
 effect ratio using option 3. with the use of simulated downstream
 water for a situation where this simulation appropriately
 represents a "site",  is a fully acceptable approach for deriving
 a  water-effect ratio  for use  in determining a site-specific
 criterion,  although it will generally provide a lower water-
 effect ratio than the other 2  options.

      As indicated in  the introduction to this guidance,  the
 determination of a water-effect ratio may require substantial
 resources.   A discharger should consider  cost-effective,
 preliminary measures  described in this guidance (e.g.,  use of
 "clean"'sampling and  chemical  analytical techniques or in non-NTR
 States, a recalculated criterion)  to determine if an indicator
 species site-specific criterion is really needed.   It may be that
 an appropropriate site-specific criterion is actually being
attained.   In many instances,  use of these other measures may
 eliminate the need for deriving final water-effect ratios.   The
 methods described in  this interim guidance should be sufficient
 to develop site-specific criteria that resolve concerns of
 dischargers when there appears to be no instream toxicity from a
 metal but,  where (a)  a discharge appears to exceed existing or ,
 proposed water quality-based permit limits,  or (b).an instream
 concentration appears to exceed an existing or proposed water
 quality criterion.

      This guidance describes  2 different methods for determining
 water-effect ratios.   Method  1 has 3 options each of which may
 only  require 3 sampling periods.   However options 1 and 2 may be
 expanded and require  a much greater effort.   While this position
 statement has discussed the simplest,  least expensive option for
method 1 (the single  discharge to a stream)  to illustrate that
 site  specific criteria are feasible even when only small
 dischargers are affected,  water-effect ratios may be calculated
using any of the other options described in the guidance if the
State/discharger believe that  there is reason to expect that a
more  accurate site-specific criterion will result from the
 increased cost and complexity  inherent in conducting the

-------
  additional -tests and analyzing the results.  Situations where
  this, could be the' case include, for example,- where seasonal
  effects in .receiving water quality or in discharge quality need
  to be assessed.

       In addition, EPA will consider other scientifically ,
.  .defensible approaches in developing final water-effect ratios as
  authorized in 40 CFR 131.11.  However, EPA strongly recommends
  that before a State/discharger implements any approach other than
  one described in this interim guidance', discussions be held with
  appropriate EPA regional offices and Office of Research and
  Development's scientists before actual testing begins.  These
  discussions would be to ensure that time and resources are not
  wasted on scientifically, and technically unacceptable approaches .
  It remains EPA's responsibility to make final decisions on the
  scientific and technical validity of alternative approaches to
  developing site-specific water quality criteria.

       EPA is fully cognizant of the continuing debate between what
  constitutes guidance and what is a regulatory requirement.
  Developing site-specific criteria is a State regulator/ option.
  Using the methodology correctly as described in this guidance
  assures the State that EPA will accept the result.   Other
  approaches are possible and logically should be discussed with •
  EPA prior "to implementation.      -              '

       The Of f ice of Science and Technology believes that this
  interim guidance advances the science of determining site-
  specific , criteria and provides policy guidance that States and
  EPA can use in this complex area.  It reflects the scientific
  advances in the past 10 years and the experience gained from
  dealing with these issues in real world situations.  This
  guidance will help improve implementation of water .quality
  standards and be the basis for future progress.
                               Tudor T.  Davies,  Director
                               Office of Science 'And Technology
                               Office of Water

-------
                             CONTENTS   •


                    ,                                     •    Page

Notices . . . . . . . . . .  . .../.,. .' .  .  .  .  .  .  ."  .  .  .  ii

Foreword  . . . . . • • • v  •• •  •  •	 iii .

Office of Science and Technology Position -Statement  .....  iv

Appendices  ........................  viii •

Figures . . . . ... . . ...  .  .  .  . .  .  .  .  .  .  .  .  ...  ix

Acknowledgments ..............  .  .  .  .  .  .  .  .  .  . . x

Executive Summary . . . . ...	  .  .  .  .  .  xi

Abbreviations . . ... . .  . .  .  ."  .  . .  .  ...  .  .  .  .  .  xiii

Glossary  .' . . . . ... «  . •  •  «  ...  ....  .  .  .  .  .  . xiv

Preface . . . . . . . . . ...  ...  . .  ...  .  .  .  .  .  .  . xvi


Introduction  ....-.....;	1


Method 1  . . -. . .... .  . ..... . •-.  .  .  (  .  .  .  .  .  .  17
   A. Experimental Design ...................  17
   B.. Background Information and Initial Decisions	44
   C. Selecting Primary and  Secondary  Tests  ...  .  .  .  .  .  .  45
   D. Acquiring and Acclimating Test Organisms   .......  47
   E. Collecting and Handling Upstream Water and Effluent  .  .  48
   F. Laboratory Dilution Water  .  ..... .  .  .  .  .  .  .-.'..  .  .  49
   G. Conducting Tests  ..........  ^  i  .......  50
   H. Chemical and Other Measurements  . .  .  .  .  .  ...  ...  55
   I. Calculating and Interpreting the Results   ...  .  .  .  .  57
   J. Reporting the Results  ...............  .  .  62 -


Method 2  .....................  	  65


References  .	  ...........  	  76
                               VI1

-------
          ~                  APPENDICES


                                                               Page

  A.  Comparison of WERs Determined Using. Upstream and
      Downstream Water  . .  . • . .  .-.".. •  •  •. •  »  v.	79

 " B.  The Recalculation Procedure	  90

  C.  Guidance Concerning the Use of "Clean Techniques" and
      QA/QC when Measuring Trace Metals	  98

  D.  Relationships. between WERs and the Chemistry and
      Toxicology of Metals  .......	;....... 109

'* E.  U.S. EPA Aquatic Life Criteria Documents for Metals . .  ,-. 134

  F.  Considerations Concerning Multiple-Metal,  Multiple-
      Discharge, and Special Flowing-Water Situations ..... 135

  G.  Additivity and the Two Components of a WER Determined
      Using Downstream Water  . ... .	 . . ....... 139

  H.  Special .Considerations Concerning the Determination
      of WERs with Saltwater Species	 145

  I.  Suggested Toxicity Tests for Determining WERs
      for Metals	 .. . 147
 •
  J.  Recommended Salts of Metals .........	153
                                 V111

-------
                    .        . FIGURES

 ;                •','..-     .                           •    Page
1.  Four Ways to Derive a Permit Limit   ...........  16
2'.  Calculating an Adjusted Geometric Mean   .........  71
3.  An Example .Derivation of a FWER . .  . .....  ...  .  .  .  72
4.  Reducing the Impact of Experimental Variation.......  73
5.  Calculating an LC50 (or EC50) by Interpolation  .  .  .  .  ..74
6.  Calculating a Time-Weighted Average  .;.........  75
Bl. An Example of the Deletion Process Using Three Phyla   .  .  97
Dl. A Scheme for Classifying Forms of Metal in Water   .  .... ill
D2. An Example of the Empirical Extrapolation Process  .... 125
D3. The Internal Consistency of the Two Approaches  .  .  .  .  .126
    I       ' .        -    ,        '           " -, -   •             . '
D4. The Application of the Two Approaches ... ,  . ...  .  . 128
D5. A Generalized Complexation Curve  ...........  .  . 131
D6. A Generalized Precipitation Curve ....... 	 132

-------
                          ACKNOWLEDGMENTS


 This document was written by:

      Charles E.  Stephan,  U.S. EPA, ORD, Environmental Research
           Laboratory,  Du'luth, MN.  .

      William H.  Peltier,  U.S. EPA, Region IV,  Environmental
           Services Division,  Athens,  GA.

      David J. Hansen,  U.S.  EPA,. ORD,  Environmental Research
           Laboratory/.Narragansett, RI.

      Charles G.  Delos,  U.S. EPA,  Office of Water,  Health
           and Ecological  Criteria Division,  Washington,  DC.

      Gary A.  Chapman,  U.S.  EPA,  ORD,  Environmental Research
           Laboratory (Narragansett),  Pacific Ecosystems ^Branch,
           Newport,  OR.


 The  authors thank all  the people who  participated  in  the open
 discussion of the experimental  determination of water-effect
 ratios on Tuesday evening, January-26,  1993  in Annapolis, MD
 Special thanks go to Herb Allen,  Bill Beckwith, Ken Bruland, Lee
Dunbar, Russ  Erickson,  and Carlton Hunt for  their  technical input
on this project,  although none  of them  necessarily agree with
everything in this  document.  Comments  by  Kent Ballentine, Karen
Gourdine, Mark Hicks, Suzanne Lussier,  Nelson Thomas, Bob Spehar,
Fritz Wagener, Robb Wood, and Phil Woods on  various drafts, or
portions  of drafts, were  also very helpful,  as were discussions
With several  other  individuals.       "   .

-------
                         EXECUTIVE SUMMARY


 A variety of physical and chemical characteristics of both the
 water ^ and the metal can influence the toxicity of a metal to
 aquatic organisms in a surface water.  When a site-specific
 aquatic life criterion is'derived for a metal, an adjustment
.procedure based on the toxicological determination of a water-
 effect ratio (WER) may be used to account for a difference
 between the toxicity of the metal in laboratory dilution water
 and its toxicity in the water at the site.  If there is a
 difference in toxicity and it is not taken into account, the
 aquatic life criterion for- the body of water will be more or less
 protective than intended by EPA's Guidelines for. Deriving
 Numerical National Water Quality Criteria for the Protection of
 Aquatic Organisms and Their Uses.  After a WER is determined for
 a site, a site-specific aquatic life criterion can be calculated
 by multiplying an appropriate national, state, or recalculated
 criterion by the WER. , Most WERs are expected to be equal to or
 greater than 1.0, but some might be less than 1.0.  Because most
 aquatic life criteria consist of- two numbers,  i.e.,  a Criterion
 Maximum Concentration (CMC)  and a Criterion Continuous
 Concentration (CCC),  either a cmcWER or a cccWER or both might be
 needed:for a site.  The cmcWER and the cccWER cannot be assumed
 to be equal, but it is not always necessary to determine both.

 In order to determine a WER,  side-by-side toxicity tests are
 performed to measure the toxicity of the metal .in two dilution
 waters.  One of the waters has to be a water that would be
 acceptable for use in laboratory toxicity tests conducted for the
 derivation of national water quality criteria for aquatic life.
 In most situations,  the second dilution water will be a simulated
 downstream water that is prepared by mixing upstream water and
 effluent in an appropriate ratio; in other situations,  the second
 dilution water will be a sample of the actual site, water to which
 the site-specific criterion is to apply.  The WER is calculated
 by dividing the endpoint obtained in the site water by the
 endpoint obtained in the laboratory dilution water.   A WER should
 be determined using a toxicity test whose endpoint is close to,
 but not lower than,  the CMC and/or CCC that is to be adjusted.
             .              "            -                      .1
 A total recoverable WER can be determined if the metal in both of
 the side-by-side toxicity tests is analyzed using the total
 recoverable measurement,  and a dissolved WER can be determined if
 the metal is analyzed in both tests using the dissolved
 measurement.   Thus four WERs  can be determined:
      Total recoverable cmcWER.                                  "
      Total recoverable cccWER.
      Dissolved cmcWER.       .
      Dissolved cccWER.              -        ,
 A total recoverable WER is  used to -calculate a total recoverable"
 site-specific  criterion from a .total recoverable national,  state,
                                XI

-------
 or recalculated aquatic.life criterion, whereas a dissolved WER
 is used to calculate a dissolved site-specific criterion from a
 dissolved criterion.  WERs are determined individually for each
 metal at each site; VpRs cannot be extrapolated from one metal to
 another, one effluent to another, or one site water to another.

 Because determining a WER requires substantial resources, the
.desirability of obtaining a WER should be carefully evaluated:
 1. Determine whether use of "clean techniques" for collecting,
    handling, storing,.preparing, and analyzing samples will
    eliminate the reason for considering determination of a WER,
    because existing data concerning concentrations of metals in
    effluents and surface waters might be erroneously high.
 2. Evaluate the potential for reducing the discharge of the
    metal.              .   .
 3. Investigate possible constraints on the permit limits, such as
    antibacksliding and antidegradation requirements arid human
    health and wildlife criteria.                •
 4. Consider use of the Recalculation Procedure.
 5. Evaluate the cost-effectiveness of determining a WER.
 If the determination of a WER is desirable,  a detailed workplan
 for should be submitted to the appropriate regulatory authority
 (and possibly to- the Water Management Division of the EPA
 Regional Office) for comment.   After the workplan is completed,
 the initial* phase should be implemented, the data should be
 evaluated,  and the workplan should be revised if appropriate.

 Two methods are used to determine WERs.  Method 1,  which is used
 to determine cccWERs that apply near plumes and to determine all
 cmcWERs,' uses data concerning three or more distinctly separate
 sampling events.  It is best if the sampling events occur during
 both low-flow and higher-flow periods.  When sampling does not
 occur during both low and higher flows, the site-specific
 criterion is derived in a more conservative manner due to greater
 uncertainty.  For each sampling event, a WER is determined using
 a selected toxicity test; for at least, one of the. sampling
 events,  a confirmatory WER is  determined using a different test.

 Method 2,  which is used to determine a cccWER for a large body of
 water outside the vicinities of plumes, requires substantial
 site-specific planning and more resources than Method 1.  WERs'
 are determined using samples of actual site water obtained at
 various times, locations, and depths to identify the range of
 WERs in the body .of water.  The WERs are used to determine how
 many site-specific CCCs should be derived for the body of water
 and what the one or more CCCs  should be.

 The guidance contained herein  replaces previous agency guidance
 concerning (a) the determination of WERs for use in the
 derivation of site-specific aquatic life criteria for metals and
 (b)  the Recalculation Procedure.  This guidance is designed to
 apply to metals,  but the principles apply to most pollutants.

                                xii   ••    • •' • '

-------
                           ABBREVIATIONS-

 ACR:   Acute-Chronic.Ratio                   :      •  .   .
 CGC:   Criterion Continuous Concentration
                i
. CMC:   Criterion Maximum Concentration
 CRM:   Certified Reference Material       .        .     '
 FAV:   Final Acute Value                               .
 FCV:   Final Chronic Value
 FW:    Freshwater
 FWER:  Final Water-Effect Ratio
 GMAV:  Genus Mean Acute Value      •
 HCME:  Highest Concentration of the Metal in the Effluent
 MDR:   Minimum Data Requirement
 NTR:   National Toxics Rule
 QA/QC: Quality Assurance/Quality Control
             ,' .     (••'.,'          *•         ,
 SMAV:  Species Mean Acute Value                 .
 SW:    Saltwater                                  .
 TDS:   Total Dissolved Solids
 TIE:   Toxicity Identification Evaluation
 TMDL:  Total Maximum Daily Load
 TOC:   Total Organic Carbon
 TRE:   Toxicity Reduction Evaluation
 TSD:   Technical Support Document
 TSS:   Total Suspended Solids
 WER:   Water-Effect Ratio
 WET:   Whole Effluent Toxicity
 WLA:   Wasteload Allocation
                    .           xiii

-------
                             GLOSSARY


 Acute-chronic ratio - an appropriate measure of the acute
      toxicity of a material divided by an appropriate
      measure of the chronic toxicity. of.the same material
      under the same conditions.      '    •

"Appropriate regulatory authority - Usually the State.water
      pollution control agency, even for States under the National
      Toxics-Rule; if, however, a State were to waive its section
      401 authority, the Water Management Division of the EPA
      Regional Office would become the appropriate regulatory
   1   authority.   .                                           .

 Clean techniques - a set of procedures designed to prevent
      contamination of samples so that concentrations of
      trace metals can be measured accurately and precisely.

 Critical species - a species that is commercially or
      recreationally important at the site,  a species that exists
      at the site and is listed as threatened or endangered under
      section 4 of the Endangered Species Act,  or a species for
    ,  which there is evidence that the loss of the species from
      the site is likely to cause an unacceptable.impact on a
      commercially or recreationally important species,  a
      threatened or endangered species,  the abundances of a
      variety of other species, or the structure or function  of
      the community.

 Design flow - the flow used for steady-state wasteload
   .   allocation modeling.

 Dissolved metal - defined here as "metal that passes through
      either a 0.45-nm or a 0.40-nm membrane  filter".

 Endpoint - the concentration of test material that is expected  to
      cause a specified amount of adverse effect.

 Final Water-Effect Ratio - the WER that is used in the
      calculation of a site-specific aquatic life criterion.  . •

 Flow-through test - a test in which test .solutions flow into
      the test chambers either intermittently (every few
      minutes)  or continuously and the excess flows .out.

 Labile metal - metal that is in water and will' readily
      convert from one form to another when in a
      nonequilibrium condition.

 Particulate metal - metal that is measured by the total
      recoverable method but not  by the dissolved method.

                                xiv

-------
 Primary test - the toxicity test used in the determination .
      of a Final Water-Effect Ratio (FWER);  the specification
      of the test includes the test species,-the life stage
      of the species,  the duration of the test,  and the
      adverse effect on which the endpoint is based.

 Refractory metal - metal that is in"water and will not
      readily convert from one form to another when in a
      nonequilibrium condition,  i.e.,  metal  that is in water
      and' is not labile.

 Renewal test - a test in which either the test solution in a
      test chamber is renewed at least once  during  the test
      or the test organisms are transferred  into a  new test
      solution of the same composition at least once  during
      the test.

 Secondary test - a toxicity test that is usually conducted
      along with the primary test only once  to test the
      assumptions that, within experimental variation,  (a)
      similar WERs will be obtained using tests  that  have
      similar sensitivities to the test material, and (b.)
      tests that are less sensitive to the test  material  will
      usually give WERs that are closer to 1.

 Simulated downstream water - a site water prepared by mixing
      effluent and upstream water in a known  ratio.

 Site-specific aquatic life criterion  - a water  quality
      criterion for aquatic life that  has been derived to be
      specifically appropriate to the  water quality
      characteristics  and/or species composition at a
      particular location.

 Site  water -  upstream water,  actual downstream water,  or
      simulated downstream water in which a toxicity  test is
      conducted side-by-side with the  same toxicity test  in a
      laboratory dilution water  to, determine a WER.

 Static test  -  a test  in  which'the solution and  organisms
      that  are  in a  test  chamber at  the beginning of  the  test
 ;     remain  in  the  chamber until  the  end of the test.

 Total recoverable metal  -  metal that  is  in aqueous solution
      after the  sample is appropriately acidified and
      digested and insoluble material  is  separated.

Water-effect ratio  -  an  appropriate measure of  the toxicity
     of a material  obtained in  a  site water divided  by the
      same measure of  the toxicity of  the same material
     obtained simultaneously in a laboratory, dilution water.
                                xv

-------
                             PREFACE


Several issues need cpnsideration when guidance such as this is
written:

1. Degrees of importance:  Procedures and methods are series of
   instructions, but some of the instructions are more important
   than others.  Some instructions are so important that, if they
   are. not followed, the results will be questionable or
   unacceptable; other instructions are less important, but
   definitely desirable.  Possibly the best way to express     :  .
   various degrees of-importance is the approach described in
   several ASTM Standards, such as in section' 3.6 of Standard
   E729 .(AS1M 1993a), which is modified here to apply to WERs:
      The words "must", "should", "may", "can", and "might" have
      specific meanings in this document.  "Must" is used to
      express an instruction that is to be followed, unless a
      site-specific consideration requires a deviation, and is
      used only in connection with instructions that directly
      relate to the validity of toxicity tests, WERs, FWERs, and
      the Recalculation Procedure.  "Should" is used to state
      instructions that are recommended and are to be followed if
      reasonably possible.  Deviation from one "should" will not
      invalidate a. WER, but deviation from several probably will.
      Terms such as "is desirable", "is often desirable", and
      "might be desirable" are used in connection with less
      important instructions.  "May" is used to mean "is (are)
      allowed to", "can" is used to mean "is (are) able to", and
      "might" is used to mean "could possibly".  Thus the classic
      distinction between "may" and "can" is preserved, and
      "might" is not used as a synonym for either "may" or "can".
   This does not eliminate all problems concerning the degree of
   importance, however.  For example,  a small deviation from a
   "must" might not invalidate a WER,  whereas a large deviation
   would.  (Each "must" and "must not".is in bold print for
   convenience,  not for emphasis, in this document.)

2. Educational and explanatory material;  Many people have asked
   for much detail in this document to ensure that as many WERs
   as possible are determined in an acceptable manner.   In
   addition,  some people want justifications for each detail.
   Much'of the detail that is desired by some people is based on
   "best professional judgment",  which is rarely considered an
   acceptable justification by people who disagree with a
   specified detail.  Even if details are taken from an EPA
   method or an ASTM standard,  they were often included in those
   documents on the basis of. best professional judgment.  In
   contrast,  some people want- detailed methodology presented
   without explanatory material.  . It was decided to include as
   much detail as is feasible,  and to provide rationale and
   explanation for major items.

                               xvi   .

-------
3. Alternatives;  When more than one alternative is both
   scientifically sound and appropriately protective, it seems
   reasonable to:present the alternatives rather than presenting
   the one that is considered best.  The reader can then select
   one based on cost-;effectiveness, personal preference, details
   of the particular situation, and perceived advantages and
   disadvantages.

4. Separation of "science*, "best professional •judgment1' and
   "regulatory decisions";  These can never be completely
   separated in this kind of document; for example, if data are
   analyzed for a statistically significant difference, the
   selection of alpha .is an important decision, but a rationale '
   for its selection is rarely presented, probably because the
   selection is not a scientific decision.  In this document, an
   attempt has been made to focus on good science, best
   professional judgment, and presentation of the rationale; when
   possible, these are separated from "regulatory decisions"
   concerning margin of safety, level of protection, beneficial
   use, regulatory convenience, and the goal of zero discharge.
   Some "regulatory decisions" relating to implementation,
   however, should be integrated with, not separated from,
   "science" because the two ought to be carefully considered
   together wherever science has implications for implementation.

5. Best professional -judgment;  Much of the guidance contained
   herein is qualitative rather than quantitative, and much
   judgment will usually be required to derive a site-specific
   water quality criterion for aquatic life.  In addition,
   although this version of the guidance for determining and
   using WERs attempts to cover all major questions that have
   arisen during use of the previous version and during
   preparation of this version, it undoubtedly does not cover all
   situations, questions, and extenuating circumstances that
   might arise in the future.  All necessary decisions should be
   based on both a thorough knowledge of aquatic toxicology and
   an understanding of this guidance; each decision should be
   consistent with the spirit of this guidance, which'is to make
   best use of "good science" to derive the most appropriate
   site-specific criteria.  -This guidance should be modified
   whenever sound .scientific evidence indicates that a site-
   specific criterion produced using this.guidance will probably
   substantially underprotect or overprotect the aquatic life at
   the site of concern.  Derivation of. site-specific criteria for
   aquatic life is a complex process and requires knowledge in
   many areas of aquatic toxicology; any deviation from this
   guidance should be carefully considered to ensure that it is '
   consistent with other parts of this guidance and with "good
   science".        .         .         '

6. Personal bias;  Bias can never be eliminated, and some
   decisions are at the fine line between "bias" and "best

                               xvii  -

-------
professional judgment".  The possibility of bias can be
eliminated only by adoption of an extreme position such as "no
regulation" or "no discharge".  One way to deal with bias is
to have, decisions made by a team of knowledgeable people.

Teamwork;  The determination of a" WER should be a cooperative
team effort beginning with the completion of the initial
workplan, interpretation of initial data, revision of the
workplan, etc.  The interaction of a variety of knowledgeable,
reasonable people will help obtain the .best results for the
expenditure of the fewest resources.  Members of the team
should acknowledge their biases so that the team can make best
use of the available information, taking into account its
relevancy to the immediate situation and its quality.
                           xvixi

-------
                            INTRODUCTION  -


 National aquatic life criteria for metals are intended to. protect
 the aquatic life in almost all surface waters of the. United
 States (U.S. EPA 1985).  This level.--of protection is accomplished
 in two ways.  First, the national dataset is required to contain
.aquatic species that have been found to be sensitive to a variety
 of pollutants.  Second, the dilution water and the metal salt
 used in the toxicity tests are required to-have physical and
 chemical characteristics that ensure that the metal is at least
 as toxic in the tests as it is in nearly all surface waters.   For
 example,  the dilution water is to be low in suspended .solids and
 in organic carbon,  and some forms of metal (e.g.,  insoluble metal
 and metal bound by organic complexing agents) cannot be used as
 the test material.   (The term "metal" is used herein to include
 both "metals" and "metalloids";)               .

 Alternatively* a national aquatic life criterion might not
 adequately protect the aquatic life at some sites.  An untested.
 species that is important at a site might be more  sensitive than
 any of the tested species.  Also, the metal might  be more toxic
 in site water than in laboratory dilution, water because,  for
 example,  the site water has a lower pH and/or hardness than most
 laboratory .waters.   Thus although a national aquatic life
 criterion is intended to be lower than necessary for most sites,
 a national criterion might not -adequately protect  the aquatic
 life at some sites.

 Because a national aquatic life criterion might be more or less
 protective than intended for the aquatic life in most bodies of
 water,  the U.S. EPA provided guidance (U.S.  EPA 1983a,1984)
 concerning three procedures that may be used to derive a site-
 specific criterion:
 1.  The Recalculation Procedure is intended to take into account
    relevant differences between the sensitivities  of the aquatic
    organisms in the national dataset and the sensitivities of
    organisms that occur at the site.
 2..The Indicator Species Procedure provides for the use of a
    water-effect ratio (WER) that is intended to take into account
    relevant differences between sthe toxicity of the metal in  '.
    laboratory dilution water and in site water.
 3.  The Resident Species Procedure is intended to take into
    account both kinds of differences simultaneously.
 A site-specific criterion is intended to come closer than the
 national  criterion to providing the intended level of protection
 to the aquatic life at the site,  usually by taking into account "
 the biological and/or chemical conditions (i.e., the species
 composition and/or water quality characteristics)  at the site.
 The fact  that the U.S. EPA has made these procedures available
 should not be interpreted as implying that the agency advocates
 that states derive site-specific criteria before setting state

-------
 standards.  Also,  derivation of  a  site-specific  criterion  does
 not  change  the  intended level of protection of the  aquatic life
 at the  site.  Because  a WER is expected to  appropriately take
 into account  (a) the4site-specific toxicity of the  metal,,  and  (b)
 synergism,  antagonism,  and additivity with  other constituents of
 the  site water, using  a WER is more-likely  to provide the
 intended level  of  protection than  not using a WER.

 Although guidance  concerning site-specific  criteria has been
 available since 1983  (U.S.  EPA 1983a,1984),  interest has
 increased in  recent years  as states have devoted more attention
 to chemical-specific water quality criteria for  aquatic life.  In
 addition, interest in  water-effect ratios  (WERs)  increased when
 the  "Interim  Guidance"  concerning  metals (U.S. EPA  1992) made.a
 fundamental change in  the  way that WERs  are experimentally
 determined  (see Appendix A),  because the change  is  expected to
 substantially increase the magnitude of  many WERs.  Interest was
 further focused on WERs when they  were integrated into some of
 the  aquatic life criteria  for metals that were promulgated by the
 National Toxics Rule  (57 FR 60848,  December 22,  1992),,  The
 newest  guidance issued by  the U.S.  EPA  (Prothro  1993) concerning
 aquatic life  criteria  for  metals affected the determination and
 use  of  WERs only insofar as it affected  the use  of  total
 recoverable and dissolved  criteria.

 The  early guidance concerning WERs (U.S. EPA 1983a,19B4)
 contained few details  and  needs revision, especially to take into
 account newer guidance concerning  metals (U.S. EPA  1992; Prothro
 1993) .  The guidance presented herein supersedes  all guidance
 concerning  WERs and the Indicator  Species Procedure given  in
 Chapter 4 of  the Water Quality Standards Handbook (U.S. EPA
 1983a)  and  in U.S. EPA (1984).  All guidance presented in U.S.
 EPA  (1992)  is superseded by that presented  by Prothro (1993) and
 by this document.  Metals  are specifically  addressed herein
 because of  the National Toxics Rule (NTR) ,and because of current
 interest in aquatic life criteria  for metals; although most of
 this,guidance also applies  to other pollutants,  some obviously
 applies only  to metals.

 Even though this document was prepared mainly because of the NTR,
 the guidance  contained herein concerning WERs is  likely to have
 impact beyond its  use with the NTR.  Therefore,  it  is appropriate
 to also present new guidance concerning  the Recalculation
 Procedure (see Appendix B)  because the previous  guidance (U.S.
EPA 1983a,1984) concerning  this procedure also contained few
details and needs  revision.   The NTR does not allow use of the
Recalculation Procedure in  jurisdictions subject  to the NTR.

The previous  guidance  concerning site-specific procedures  did not
allow the Recalculation Procedure  and the WER procedure to be
used together in the derivation of  a site-specific  aquatic life
criterion;  the only way to  take into account both species

-------
 composition and water quality  characteristics  in the
 determination of a  site-specific  criterion was to use  the
 Resident Species Procedure.  A specific  change contained herein
 is that, except In  -jurisdictions  that are subject to the NTR. the
 Recalculation Procedure and the WER  Procedure  mav now  be used
 together.  Additional reasons  for addressing both the
 Recalculation .Procedure and the WER  Procedure  in this  document
 are that both procedures are based directly on the guidelines for
 deriving national aquatic life criteria  (U.S.  EPA 1985) and, when
 the two are used together, use of the Recalculation Procedure has
 specific implications concerning  the determination of  the WER.

 This guidance is intended to produce WERs that may be  used to
 derive site-specific aquatic life criteria for metals  from most
 national and state,aquatic life criteria that  were derived from
 laboratory toxicity data.  Except in jurisdictions that are
 subject to the NTR, the WERs may  also be used  with site-specific
 aquatic life criteria that are derived for metals using the
 Recalculation Procedure described in Appendix  B.  WERs obtained
 using the methods described herein should not  be used  to adHust
 aquatic life criteria that were derived for metals in  other ways.
 For example, because they are  designed to be applied to criteria
 derived on the basis of laboratory toxicity tests, WERs
 determined using the methods described herein  cannot be used to
 adjust the residue-based mercury  Criterion Continuous
 Concentration (CCC) or the field-based selenium freshwater
 criterion.  For the purposes of the NTR, WERs may be used with
 the aquatic life criteria for  arsenic, cadmium, Chromium(III),
 chromium(VI), copper, lead, nickel,  silver, and zinc and with the
 Criterion Maximum Concentration (CMC) for mercury.  WERs may also
 be used with saltwater criteria for selenium.

 The concept of a WER is rather simple:
   Two side-by-side toxicity tests are conducted - one test using.
   laboratory dilution water and  the other using site water.  The
   endpoint obtained using site water is divided by the endpoint
   obtained using laboratory dilution water.  The quotient, is -the
   WER, which is multiplied times the national, state,  or
   recalculated aquatic life criterion to calculate the site-
   specific criterion.                                      ,
Although the concept is simple, the determination and use of WERs
 involves many considerations.

The primary purposes of this document are',to:
1. Identify steps that should be  taken before the determination
   of a WER is begun.                                      .
2. Describe the methods recommended by the U.S. EPA for the
   determination of WERs.
3. Address some issues .concerning the use of WERs.
4. Present new guidance concerning the Recalculation Procedure.

-------
 BeforeDetermining a WER

 Because a national criterion is intended to-protect aquatic life
 in almost all bodies., of water and because a WER is intended to
 account for a difference between the toxicity of a metal in a
 laboratory dilution water and its toxicity in a site water,
 dischargers who want higher permit .limits than those derived on
. the basis of an existing aquatic life criterion will probably
 consider determining a WER.  Use of a WER should be considered
 only as a last resort for at least three reasons:
 a. Even though some WERs will be substantially greater than 1.0,
    some will be about 1.0 and some will be less than 1.0.
 b.. The determination of a WER requires substantial resources.
 c. There are other things that a discharger can do that might be
    more cost-effective than determining a WER.

 The two situations in which the determination of a WER might
 appear attractive to dischargers are when (a)  a discharge appears
 to exceed existing or proposed water quality-based permit limits,
 and (b) an instream concentration appears to exceed an existing
 or proposed aquatic life criterion.  Such situations result from
 measurement of the concentration of a metal in an effluent or a
 surface water.  It would therefore seem reasonable to ensure that
 such measurements were not subject to contamination.   Usually it
 is much easier to verify chemical measurements by using "clean
 techniques" for collecting,  handling,  storing,  preparing,  and
 analyzing samples, than to determine a WER.   Clean techniques and
 some related QA/QC considerations are discussed in Appendix C.

 In addition to investigating the use of "clean techniques",  other
 steps that a discharger should take prior to beginning the
 experimental determination of a WER include:
 1. Evaluate the potential for reducing the discharge of the
    metal.                                 •   .  .
 2. Investigate such possible constraints on permit limits as
    antibacksliding and antidegradation requirements and human
    health and wildlife criteria.
 3. Obtain assistance from an aquatic toxicologist  who understands
    the basics of WERs (see Appendix D),  the  U.S. EPA's national
    aquatic life guidelines (U.S.  EPA 1985),  the guidance
    presented by Prothro (1993),  the national criteria document'
    for the metal(s)  of concern .(see Appendix E), the  procedures
    described by the U.S.  EPA (1993a,b,c)  for acute and chronic
    toxicity tests on effluents and surface waters,  and the
    procedures described by ASTM (1993a,b,c,d,e)  for acute and
    chronic toxicity tests  in laboratory dilution water.
4. Develop an initial definition of the site to which the site- '
    specific criterion is  to  apply.           ,
5. Consider use of the Recalculation Procedure  (see Appendix B).
6. Evaluate the cost-effectiveness  of  the determination,of  a WER.
    Comparative - toxicity tests  provide  the most  useful data,  but
    chemical analysis  of the  downstream water might be helpful

-------
    because the following are often true for some metals:
    a. The lower the percent of the total recoverable metal in the
       downstream-water that, is dissolved, the higher the WER.
    b. The higher the .concentration of total organic carbon (TOO
       and/or total suspended solids (TSS),  the higher the WER.
    It is also true that the higher the concentration of nontoxic
    dissolved metal,  the higher the WER.  Although some chemical
    analyses might provide useful information concerning the
    toxicities of some metals in water,  at the present only
    toxicity tes.ts can accurately reflect the toxicities of
    different forms of a metal,(see Appendix D).
 7.  Submit a workplan for the experimental determination of the
    WER to the appropriate -regulatory authority (and possibly to
    the Water Management Division of the EPA Regional Office)  for
    comment.  The workplan should, include detailed descriptions of
    the site; existing criterion and standard;  design flows;  site
    water; effluent;  sampling plan;  procedures  that will be used
    for collecting,  handling,  and analyzing  samples of site water
    and effluent;  primary and secondary  toxicity  tests;  quality
    assurance/quality control (QA/QC)  procedures;  Standard
    Operating Procedures (SOPs);  and data interpretation.
 After the workplan is completed,  the initial phase should be
 implemented; then the data obtained should  be  evaluated,  and  the
 workplan should be revised if appropriate.   Developing  and
 modifying the workplan and analyzing and interpreting the data
 should be a cooperative effort  by a team of 'knowledgeable people.
Two Kinds of WERs
              i
Most aquatic life criteria contain both a CMC and a CCC, and it
is usually possible to determine both a cmcWER and a cccWER.  The
two WERs 'cannot be assumed to be equal because the magnitude of a
WER will probably depend on the sensitivity of the toxicity test
used and .on the percent effluent in the site water, (see Appendix
D), both of which can depend on which WER is to be determined.
In some cases, it is expected that a larger WER can be applied to
the CCC than to the CMC, and so it would be environmentally
conservative to apply cmcWERs to CCCs.  In such cases it is
possible to determine a'• cmcWER and apply it to both the CMC and
the CCC in order to derive a site-specific CMC, a site-specific
CCC, and new permit limits.  If these new permit limits are
.controlled by the new site-specific CCC, a cccWER could be
determined using a more sensitive test, possibly raising the
site-specific CCC and the permit limits again.  A cccWER may, of
course, be determined whenever desired.  Unless the experimental
variation is increased, use of a- cccWER will usually improve the
accuracy of the resulting site-specific CCC.

In some cases, a larger WER cannot'be applied to the CCC than to
the CMC and so it might not be environmentally conservative to  "
apply a cmcWER to a CCC (see section A.4 of Method 1) .

-------
 Steady-state and Dynamic Models

 Some of the guidance contained herein specifically applies to
 situations, in which the permit limits were calculated using  ,
 steady-state modeling''; in particular, some samples are to be
 obtained when the actual stream flow; is close to the design flow.
 If permit limits were calculated using dynamic modeling,  the
_ guidance will have to be modified,  but- it is.unclear at present
"what^modifications are most appropriate.  For example, it might
 be useful to determine whether the magnitude of the WER is
 related to the flow of the upstream water and/or the effluent.    .


 Two Methods

 Two methods are used to determine WERs.  Method 1 will probably
 'be -used to determine all cmcWERs and most cccWERs because it can
 be applied to situations that are in the vicinities of plumes.
 Because WERs are likely to depend on the concentration of
 effluent in the water and because the percent effluent in a water
 sample obtained in the immediate vicinity of a plume is unknown,
 simulated downstream water is used so that the percent effluent
 in the sample is known.  For example, if a sample that was
 supposed to represent a complete-mix situation was accidently
 taken'in the plume upstream of complete mix, the sample would
 probably have a higher percent effluent and a higher WER than a
 sample taken downstream of complete mix; use of the higher WER to
 derive a site-specific criterion for the complete-mix situation
 would result in underprotection.  If the sample were accidently
 taken upstream of complete mix but outside the plume,
 overprotection would probably result.

 Method 1 will probably be used to determine all cmcWERs and most
 cccWERs in flowing fresh waters, such as rivers and streams.
 Method 1 is intended to apply not only to ordinary rivers and
 streams but also to streams that some people might consider
 extraordinary, such as streams whose design flows are zero and
 streams that some state and/or federal agencies refer to as
 "effluent-dependent", "habitat-creating", or "effluent-
 dominated" .  Method 1 is also used to determine cmcWERs in such.
 large sites .as oceans and large lakes, reservoirs, and estuaries
 (see Appendix F).                         .

 Method 2 is used to determine WERs that apply outside the area .of
 plumes in large bodies of water.  Such WERs will be cccWERs and
 will be determined using samples of actual site water obtained at
 various times, locations, and depths in order to identify the
 range of WERs that apply to the body of water.  These
 experimentally determined WERs are then used to decide how many
 site-specific criteria should be derived for the body of water
 and what the criterion (or criteria) should -be.  Method 2
 requires substantially more resources than Method 1.

-------
The complexity of each method increases when the number of metals
and/or the number of discharges is two or more:
a. The simplest situation is when -a WER is to be determined for
   only one metal,and only one discharge has permit limits for
   that metal.,   (This is the single-metal single-dispharge
   situation.)            .
b. A more complex .situation is when a WER is to be determined for
   only one metal, but more than one discharge has permit limits
 ;  for that metal.   (This is the single-metal multiple-discharge
   situation.)                            •
c. An even more complex situation is when WERs are to be
   determined for more than one metal, but only one discharge has
   permit limits for any of the metals.  (This is the multiple-
   metal single-discharge situation.)                        •
d. The most complex situation is when WERs are to be determined
   for more than one metal and more than one discharge has permit
   limits for some or all of the metals.  (This is the multiple-
   metal multiple-discharge situation.)
WERs need to be determined for each metal at each site because
extrapolation of a WER from one metal to another, one effluent to
another, or one surface water to another is too uncertain.

Both methods work well in multiple-metal situations, but special
tests or additional tests will be necessary to show that the
resulting combination of site-specific criteria will not ,be too
toxic.  Method 2 is better suited to multiple-discharge
situations than is Method 1.  Appendix F provides additional
guidance concernincf multiple-metal and multiple-discharge
situations, but it does not discuss allocation of waste loads,
which is performed when a wasteload allocation (WLA) or a total
maximum daily load (TMDL) is developed (U.S. EPA 1991a).


Two Analytical Measurements  -

A total recoverable WER can be determined if the metal in both of
the side-by-side toxicity tests is analyzed using the total
recoverable measurement; similarly, a dissolved WER can be
determined if the metal in both tests is analyzed using the
dissolved measurement.  A total recoverable WER is used to
calculate a total recoverable site-specific criterion from an '
aquatic life criterion that is expressed using the total
recoverable measurement, whereas a dissolved WER is used to
calculate a dissolved site-specific criterion from a criterion
that is expressed in terms of the dissolved measurement.  Figure
1 illustrates the relationships between total recoverable and
dissolved criteria, WERs, and the Recalculation Procedure.

Both Method 1 and Method 2 can be used to determine a total
recoverable WER and/or a dissolved WER.  The only difference in
the experimental procedure is whether the WER is based on
measurements of total recoverable metal or dissolved metal in the

-------
 test solutions.   Both total recoverable and dissolved
 measurements are to be performed for all tests to help judge the
 quality of the tests/  to provide a check on-the analytical.
 chemistry,  and to help understand the results-; performing both
 measurements also increases the alternatives available for use of
 the results.  For example,  a dissolved WER that is not useful
 with a total recoverable criterion might be Useful in the future
.if a dissolved criterion becomes available.  Also,  as explained
 in Appendix D.,  except for experimental variation,  use of a total
 recoverable WER with a total recoverable criterion should produce
 the same tcrtal recoverable permit limits, as use of .a dissolved
 WER with a dissolved criterion; the internal consistency .of the
 approaches and the data can be evaluated if both total  .
 recoverable and dissolved criteria and WERs are determined.   It
 is expected that' in many situations total recoverable WERs will
 be larger and more variable than dissolved WERs.
 The Quality of the Toxicitv Tests

 Traditionally, for practical reasons,  the requirements  concerning
 such aspects as acclimation of test organisms  to test temperature.
 and dilution water have not been as stringent  for toxicity  tests
 on surface waters and effluents as for tests using laboratory
 dilution water.  Because a WER is a ratio•calculated from the
 results of side-by-side tests,  it might seem that acclimation  is
 not important -for a WER as long as the organisms and conditions
 are identical in the two tests.  Because WERs  are used  to adjust
 aquatic life criteria that are derived from results of  laboratory
 tests,  the tests conducted in laboratory dilution water for the
 determination of WERs should be conducted in the same way as the
 laboratory toxicity tests used in the  derivation of aquatic life
 criteria.  In the WER process,  the tests in laboratory  dilution
 water provide the vital link between national  criteria  and  site-
 specific criteria, and so it is important to compare at .least
 some results obtained in the laboratory dilution water  With
 results obtained in at least one other laboratory.

 Three important principles for making  decisions  concerning  the
 methodology for the side-by-side tests are:
 1.  The  tests using laboratory dilution water should-be  conducted
    so that the results would be acceptable for use in the
 "  derivation of national criteria.
 2.  As much as is feasible, the tests using site  water should be
    conducted using the same procedures as the  tests using the
    laboratory dilution water.
 3.  All  tests should follow any special requirements that are
    necessary because the results are to be used  to calculate'a
    WER.  Some such special requirements are imposed because the
    criterion for a rather complex situation is being changed  ,
    based on few data,'so more assurance is required that the data
    are  high quality.
                                                          i
                                 8

-------
 The most important special requirement is that the concentrations
 of the metal are to be measured using both the total recoverable
 and dissolved methods in all toxicity tests-used for the
 determination of a WER.  This requirement is necessary because
 half of the tests conducted for the determination of WERs use a
 site water in which the concentration.of metal probably is not
 negligible.  Because it is likely that the concentration of metal
.in the laboratory dilution water is negligible, assuming that the
 concentration in both waters is negligible and basing WERs on the
 amount of metal added would produce an unnecessarily low value
 for the WER.  In addition, WERs are based on too few data to
 assume that nominal concentrations are accurate.  Nominal
 concentrations obviously cannot be used if a dissolved WER is to
 be determined.  Measured dissolved concentrations at the
. beginning and end of the test are used to judge the acceptability
 of the test, and it is certainly reasonable to measure the total
 recoverable concentration when the dissolved concentration is
 measured.  Further, measuring the concentrations might lead to an.
 interpretation of the results that allows a substantially better
 use of the WERs.


 Conditions for Determining a WER      '

 The appropriate regulatory authority might recommend that one or
 more conditions be met when a WER is determined in order to
 reduce the possibility of having to determine a new WER later:
 1. Requirements that are in the existing permit concerning WET
    testing, Toxicity Identification Evaluation (TIE), and/or
    Toxicity Reduction Evaluation (TRE) (U.S. EPA 1991a);
 2. Implementation of pollution prevention efforts, such as
    pretreatment, waste minimization, and source reduction.
 3. A demonstration that applicable technology-based requirements
    are being met.
 If one or more of these is not satisfied when the WER is
 determined and is implemented later, it is likely that a new WER
 will have to be determined because of the possibility of a change
 in the composition of the effluent.

.Even if all recommended conditions are satisfied, determination
 of a WER might not be possible if the effluent, upstream water',
 and/or downstream water are toxic to the test organisms.  In some
 such cases, it might be possible to determine a WER, but
 remediation of the toxicity is likely to be required anyway.  It
 is unlikely that a WER determined before remediation would be
 considered acceptable for use after remediation.  If it is
 desired to determine a WER before remediation and the toxicity is
 in the upstream water, it might be possible to use a laboratory
 dilution water or a water from a clean tributary in place of the
 upstream water; if a substitute water is used, its water quality
 characteristics should be similar to those of the upstream water
 (i.e., the pH should be within.0.2 pH units and the hardness,

-------
 alkalinity,  and concentrations of TSS and TOC should be within 10
 % or 5 mg/L,  whichever is greater,  of those in the upstream
 water).  If the upstream water is chronically toxic,  but not
 acutely toxic,  it might be possible to determine a cmcWER even if
 a cccWER cannot be determined; a cmcWER might not be useful,
 however,  if the permit limits are controlled by the CCC; in such
 a case, it would probably not be acceptable to assume that the
• cmcWER is an environmentally conservative estimate of the cccWER.
 If the WER is determined using downstream water and the toxicity
 is due. to the effluent,  tests at lower concentrations of the
 effluent might give an indication of the amount of remediation
 needed.
 .Conditions for Using a WER

 Besides requiring' that the WER be valid,  the appropriate
 regulatory authority might consider imposing other  conditions  for
 the approval of a site-specific criterion based  on  the WER:
 1.  Periodic reevaluation of the WER.
    a.  WERs determined in upstream water.take into account
       constituents contributed by point and nonpoint  sources and
       natural runoff; thus a WER should be  reevaluated whenever
       newly implemented controls or other changes substantially
       affect such factors as hardness,  alkalinity,  pH, suspended
       solids,  organic carbon,  or other  toxic materials.
    b.  Most WERs determined using downstream water are influenced
       more by the effluent.than the upstream water.   Downstream
       WERs should be 'reevaluated whenever newly  implemented
       controls, or other changes might substantially impact the
       effluent,  i.e.,  might impact the  forms and concentrations
       of the metal,  hardness,  alkalinity, pH, suspended solids,
       organic carbon,  or other toxic materials.  A  special
       concern is the possibility of a shift from discharge of
       nontoxic metal to discharge of toxic  metal such that the
       concentration of the metal does not increase; analytical
       chemistry might not detect the change but  toxicity tests
    ,   would.
    Even if no changes are known to have occurred, WERs should  be
    reevaluated periodically.   (The NTR  recommends that NPDES
    permits include periodic determinations  of WERs  in the      '
    monitoring requirements.)   With advance  planning,  it should
    usually be possible to perform such  reevaluations  under
    conditions  that are at least reasonably  similar  to those that
    control the permit limits (e.g., either  design-flow or high-
    flow conditions)  because there should  be a reasonably long
   period  of  time during which the reevaluation  can be performed.
    Periodic determination of WERs should  be designed  to answer
    questions,  not just generate data.
2.  Increased  chemical  monitoring of the upstream water, effluent,
    and/or  downstream water, as  appropriate,'for  water quality
    characteristics  that probably affect the toxicity  of the metal

                                10

-------
    (e.g.,, hardness,  alkalinity,  pH,  TOC,  and TSS)  to determine
    whether conditions change.  The conditions at the times the
    samples were obtained should be kept on record for reference.
    The WER. should be reevaluated whenever hardness,  alkalinity,
    pH, TOC,  and/or TS'S decrease below the values that existed
    when the WERs were determined.   •-
 3.  Periodic reevaluation of the environmental fate of the metal
.in the effluent (see Appendix A) .
 4.  WET testing.
 5.  Instream bioassessments.               .

 Decisions concerning the possible imposition of such conditions
 should take into account:
 a.  The ratio of the new and old criteria.  The greater the
    increase in the criterion, the more concern there should be
    about (1)  the fate of any nontoxic metal that contributes to
    the WER and (2) changes in water quality that might occur
    within the site.   The imposition of one or more conditions
    should be considered if the WER is used to raise the criterion
    by, for example,  a factor of  two,  and  especially if it is
    raised by a factor of five or more. The significance of the
    magnitude of the ratio can be judged by comparison with the
    acute-chronic ratio,  the factor of two that is  the ratio of
    the FAV to the CMC,  and the range of sensitivities of species
    in the criteria document for  the metal (see Appendix E).
 b.  The size of the site.                              •
 c.  The size of the discharge.
 d.  The rate of downstream dilution.
 e.  Whether the CMC or the CCC controls the permit  limits.
 When  WERs are determined using upstream water,  conditions on the
 use of a WER are more likely when the water contains an effluent
 that  increases the WER by adding TOC and/or TSS, because the WER-
 will  be larger and any decrease  in the discharge of  such TOC
 and/or TSS might decrease the WER and result in underprotection.
 A WER determined using downstream water is likely  to be larger
 and quite dependent  on the composition of the effluent;  there
 should be concern about whether  a change  in the effluent might
 result in underprotection at some time in the future.


 Implementation Considerations                                 «.

 In  some situations a discharger  might not want to  or might not be
 allowed to raise a criterion as  much as could be justified by ,a
 WER:
 1.  The maximum possible increase is  not needed and raising the
    criterion  more than needed might  greatly raise  the cost if a  •
    greater increase  would require more tests and/or  increase the
    conditions imposed on approval of the  site-specific criterion.
 2.  Such other constraints as antibacksliding or antidegradation
    requirements or human health  or wildlife criteria might limit
    the amount of increase regardless of the magnitude of the WER.

    .  •    '                     11         •     •• •      '  - •   '

-------
 3. The permit limits might be limited by-an aquatic life
    criterion that applies outside the site.  It is EPA policy
    that permit limits cannot be so high  that they inadequately
    protect a portion .-of the same or a different body of water
    that is outside the site; nothing contained herein changes
    this policy in any way.         •'
 If no increase in the existing discharge is allowed, the only use
. of a WER will be to determine whether an existing discharge needs
• to be reduced.  Thus a major use of WERs might be where
 technology-base.d controls allow concentrations in surface waters
 to exceed national, state, or recalculated aquatic life criteria.
 In this case, it might only be necessary to determine that the
 WER is greater than a particular value;  it might not be necessary
 to quantify the WER.  When possible, it  might be desirable to
 show that the maximum WER is greater than the WER that will be
 used in order to demonstrate that a margin of safety exists, but
 again it might not be necessary to quantify the maximum WER.

 In jurisdictions not subject to the NTR, WERs should be used to
 derive site-specific criteria, not just  to calculate permit
 limits, because data obtained from ambient monitoring should be
 interpreted by comparison with ambient criteria.  (This is not a .
 problem in jurisdictions subject to the  NTR because the NTR   :
 defines the ambient criterion as "WER x  the EPA criterion".)  If
 a WER is used to adjust permit limits without adjusting the
 criterion, the permit limits would allow the criterion to be
 exceeded.  Thus the WER should be used to calculate a site-
 specific criterion, which should then be used to calculate permit
 limits.  In some states, site-specific criteria can only be
 adopted as revised criteria in a separate, independent water
 quality standards review process.  In other states,  site-specific
 criteria can be developed in conjunction with the NPDES
 permitting process, as long as the adoption of a site-specific
 criterion satisfies the pertinent water  quality standards
 procedural requirements (i.e., a public  notice and a public
 hearing).  In either case, site-specific criteria are to be
 adopted prior to NPDES permit issuance.  Moreover, the .EPA
 Regional Administrator has authority ,to  approve or disapprove all
 new and revised site-specific criteria and to review NPDES
 permits to verify compliance with the applicable water quality
 criteria.           .                                         '. .

 Other aspects of the use of WERs in connection with permit
 limits, WLAs, and TMDLs are outside the  scope of this document.
 The Technical Support Document (U.S. EPA 1991a) and Prothro
 (1993) provide more information concerning implementation
 procedures.  Nothing contained herein should be interpreted as  "•
 changing the three-part approach that EPA uses to protect aquatic
 life: (1) numeric chemical-specific water quality criteria for
 individual pollutants, (2) whole effluent toxicity (WET) testing,
 and (3) instream bioassessments.
                                 12

-------
Even'though there are similarities between WET testing and the
determination of WERs, there are important differences.  For
example, WERs can be used to derive site-specific criteria for
.individual pollutants,, but WET testing cannot.  The difference
between WET testing and the determination of WERs is less when
the toxicity tests used in the determination .of the WER are ones
that are used in WET testing.  If a WER is used to make a large
change in a criterion, additional WET testing and/or instream
bioassessments are likely to be recommended.


The Sample—Specific WER Approach
   ;                  '            .             .   • ' -
A major problem with the determination and use of aquatic life
criteria for metals is that -no analytical measurement or
combination of measurements has yet been shown to explain the
toxicity of a metal to aquatic plants, invertebrates, amphibians,
and fishes over the relevant range of conditions in surface
waters (see Appendix D).  It is not just that insufficient data
exist to justify a relationship; rather, existing data possibly
contradict some ideas that could possibly be very useful if true.
For example, the concentration of free metal ion could possibly
be a useful basis for expressing water quality criteria for
metals if it could be feasible and could be used in a way that
does not result in widespread underprotection of aquatic life.
Some available data, however, might contradict the idea-that the
toxicity of copper to aquatic organisms is proportional to the
concentration or the activity of the cupric ion.  Evaluating the
usefulness of any approach based on metal speciation is difficult
until it is known how many of the species of the metal are toxic,
what the relative toxicities are, whether they are additive (if
more than one is toxic), and the quantitative effects of the
factors that have major impacts on the bioavailability and/or
toxicity of the toxic species.  Just as it is not easy to find a
useful quantitative relationship between the analytical chemistry
of metals and the toxicity of metals to aquatic life, it is also
not easy to find a qualitative relationship that can be used to
provide adequate protection for the aquatic life in almost all
bodies of water without providing as much overprotection for some
bodies of water as results from use of the total recoverable and
dissolved measurements.

The U.S.  EPA cannot ignore the existence of pollution problems •
and delay setting aquatic life criteria until all scientific
issues have been adequately resolved.  In light of uncertainty,
the agency needs to derive criteria that are environmentally
conservative in most bodies of water.  Because of uncertainty
concerning the relationship between the analytical chemistry and
the toxicity of metals, aquatic life criteria for metals are
expressed in terms of analytical measurements that result in the
criteria providing more protection than necessary for the aquatic
life in most bodies of water. .The agency has provided for the

   ' •  • •       •    '  '   .•  •     13

-------
  use of WERs to address the general conservatism,  but expects that
  some WERs will be less than 1.0 because national,  state,  and
  recalculated criteria are not necessarily environmentally
  conservative for all'.bodies of water.                   •  .   "•

  It has become obvious, however, that the determination and use of
  WERs is not a simple solution to the existing general
 . conservatism.  It is likely that a permanent solution will have
  to be based on an adequate quantitative explanation of how metals
  and aquatic organisms interact.  In the meantime,  the use of
  total recoverable and dissolved measurements to express criteria
  and the use of site-specific criteria  are intended to provide
  adequate protection for almost all bodies of water without
  excessive overprotection for too many  bodies of water.  Work
.  needs to continue on the permanent solution and,  just in  case,  on
  improved alternative approaches.

  Use of WERs to derive site-specific criteria is intended  to allow
  a reduction or elimination of the general overprotection
  associated with application of a national criterion to individual
  bodies of water,  but a major problem is that a WER will rarely be
  constant over time,  location,  and depth in a body of water, due to
  plumes, mixing,  and resuspension.  It  is possible that  dissolved
  concentrations and WERs will be less variable than total
  recoverable ones.   It might also be possible to reduce  the impact
  of the heterogeneity if WERs are additive across  time,  location,
  and depth (see Appendix G) . . Regardless of what approaches,
  tools,  hypotheses,  and assumptions are utilized,  variation will
  exist and WERs will have to be used in a conservative manner.
  Because of variation between bodies of water,  national  criteria
  are derived to be environmentally conservative for most bodies  of
  water,  whereas the WER procedure, which is intended to  reduce  the
  general conservatism of national criteria,  has to be conservative
  because of variation among WERs within a body of water.

  The conservatism introduced by variation among WERs is  due not to
  the concept of WERs,  but to the way they are used.   The reason
  that national criteria are conservative, in the first place is  the
  uncertainty concerning, the linkage of  analytical  chemistry and
 . toxicity;  the toxicity of solutions can be measured,  but  toxicity
  cannot  be modelled adequately using available chemical
  measurements.   Similarly,  the current  way that WERs are used
  depends on a linkage between analytical chemistry  and toxicity
  because WERs are  used to.derive site-specific criteria  that are
  expressed in terms  of chemical measurements.

  Without changing  the amount  or kind of toxicity testing that is "
  performed when WERs  are determined using Method 2,  a different
  way of  using the WERs could avoid some of the problems  introduced
  by the  dependence  on analytical chemistry.  .The "sample-specific
  WER approach"  could consist  of sampling a body of  water at a
  number  of  locations,  determining the WER for each sample,  and

                                 14

-------
 measuring the concentration of the metal in each 'sample.  Then
 for each individual sample, a quotient would be calculated by
 dividing the concentration of metal in the sample by the product
 of the national criterion times the WER obtained for that sample
 Except for experimental variation, when the quotient for a sample
 is less than-1, the concentration of metal in that sample is
 acceptable; when the quotient for a sample is greater than 1, the
. concentration of metal in that sample is too high.  As a check,
 both the tptal recoverable measurement and the dissolved
 measurement should be used because they should provide the same
 answer if everything is done correctly and accurately.  This
 approach can also be used whenever Method 1 is used; although  •
 Method 1 is used with simulated downstream water,  the sample-
 specific WER approach can be used with either simulated
 downstream water or actual downstream water.

 This sample-specific WER approach has several interesting
 features:               '
 1. It is not a different way of determining WERs;  it is merely a
    different way of using the WERs that are determined.
 2. Variation among WERs within a body of water is  not a problem.
 3. It eliminates problems concerning the unknown relationship
    between toxicity and analytical chemistry.
 4. It works equally well in areas that are in or near plumes  and
    in areas that are away from plumes.                       ,
 5. It works equally well in single-discharge and multiple-
    discharge situations.
 6. It automatically accounts for synergism,  antagonism,  and
    additivity between toxicants.
 This way of using WERs is equivalent to expressing the national
 criterion for a pollutant in terms of toxicity tests  whose
 endppints equal the CMC'and the CCC;.if the site water causes
 less adverse effect than is defined to be the  endpoint,  the
 concentration of that pollutant in the site water  does not  exceed
 the national criterion.   This sample-specific  WER  approach  does
 not directly fit into the current framework wherein criteria  are
 derived  and then permit  limits are calculated  from the criteria.

 If the sample-specific WER approach were  to produce a number  of
 quotients that  are greater than 1,  it  would seem that the
 concentration of metal in the discharge(s)  should be  reduced  '
 enough that  the  quotient is  not. greater than 1.  Although this
might  sound  straightforward,  the  discharger(s) would  find that  a
 substantial  reduction in the discharge of a metal would not
 achieve  the  intended result  if the reduction was due  to removal
 of nontoxic metal.   A chemical monitoring approach  that  cannot
differentiate between toxic  and nontoxic metal would  not detect "
 that only nontoxic metal had been removed, but the  sample-
 specific  WER  approach would.
                                15

-------
Figure  1:  Four Ways to  Derive a Permit  Limit
                          I Total Recoverable Criterion
                                    Recalculation
                                      Procedure
    Tofcal
Recoverable
                                                                    ^
                                                        and/or cccWER
                                                          \/
                                            Total Recoverable
                                          Site-specific Criterion
                          Total Recoverable Permit Limit
        Dissolved Criterion » (TR Criterion) (% dinaolved in toricfty tests)!
                                    Recalculation
                                      Procedure
                                            _v
                                                              _v
  Dissolved
  cmcWER
and/or cccWER
                                             Dissolved Site-
                                                   Criterion
                                                   \/
      Net % contribution from the total recoverable metal in the efiluent
      to the dissolved metal in the downstream water.  (This will probably
      change if the total recoverable concentration in the effluent
                                     \/
                       I Total Recoverable Permit Limit
 For both the total recoverable and dissolved measurements, derivation of an
 optional site-specific criterion is described on the right.  If both the
 Recalculation Procedure and the WER procedure are used, the Recalculation
 Procedure must be performed first. (The Recalculation Procedure cannot be
 used in jurisdictions that are subject to the National Toxics Rule.)
                                     16

-------
 METHOD 1: DETERMINING WERs FOR AREAS IN OR NEAR PLUMES


 Method 1 is based on .'the determination of WERs using simulated
 downstream water and so it can be used to determine a. WER that
 applies in the vicinity of a plume. -'" Use of simulated downstream
 -water ensures that the concentration of effluent in the site
.water is known,, which is important because the magnitude of the
 WER will often depend on the concentration of effluent in the
 downstream water.  Knowing the concentration of effluent makes it
 -possible to quantitatively relate the WER to the effluent.
 Method 1 can be used to determine either cmcWERs or cccWERs or
 both in single-metal,  flowing freshwater situations,  including
 streams whose design flow is zero and "effluent-dependent"
 streams (see Appendix F) .   As is also explained,in Appendix F,
 Method 1 is used when cmcWERs are determined for "large sites",
 although Method 2 is used when cccWERs are determined'for "large
 sites".  In addition,  Appendix F addresses special considerations
 regarding multiple-metal and/or multiple-discharge situations.

 Neither Method 1 nor Method 2 covers all important methodological
 details for conducting the side-by-side toxicity tests that are
 necessary in order to determine a WER.  .Many references are made
 to information published by the U.S.  EPA (1993a,b,c)  concerning
 toxicity tests on effluents and surface waters and by ASTM
 (1993a,b,c,d,e,f) concerning tests in laboratory dilution water.
 Method 1 addresses aspects of toxicity tests that (a)  need
 special attention when determining WERs and/or (b)  are usually
 different for tests conducted on effluents and tests  conducted in
 laboratory dilution water.  Appendix H provides additional
 information concerning toxicity tests with saltwater  species.


 A.  Experimental Design

    Because of the variety  of considerations that have important
    implications for the determination of a WER,  decisions
    concerning experimental design should be given careful
    attention and need  to answer the  following questions:
    1.  Should WERs be determined using upstream water,  actual
       downstream water,  and/or simulated downstream water?    '
    2.  Should WERs be determined when  the stream flow  is equal  to,
       higher than,  and/or  lower than  the design flow?
    3.  Which toxicity tests should be  used?
    4.  Should a cmcWER  or a cccWER or  both be determined?
    5.  How should a FWER be derived?
    6.  For metals whose criteria are hardness-dependent,  at what "
       hardness should  WERs be determined?
    The answers to these questions should be based- on  the  reason
    that ;WERs are determined,  but the  decisions should also take
    into .account some .practical consideration's.


                                17         •

-------
Should WERs be determined using upstream water, actual
downstream water, and/or simulated downstream water?

a. Upstream water provides the least complicated way of
   determining and using WERs because plumes, mixing
   zones/ and effluent variability do not have to be taken
   into account.  Use of upstream water provides the least
   useful WERs because it does not take into account the
   presence of the effluent, which is the source of the
   metal.  It is easy to assume that upstream water will
   give .smaller WERs than downstream water, but in some
   cases downstream water might give smaller WERs (see
   Appendix 6).  Regardless, of whether upstream water
  . gives smaller or larger WERs, a WER should be       .
   determined using the water to which the site-specific
   criterion is to apply (see Appendix A).

b. Actual downstream water might seem to be the most
   pertinent water to use when WERs are determined,  but
   whether this is time depends on what use is to be made
   of the WERs.  WERs determined using actual downstream
   water can be quantitatively interpreted using the
   sample-specific WER approach described at the end of
   the Introduction.  If, however,  it is desired to
   understand the gucintitative implications of a WER for
   an effluent of concern,  use of actual downstream water
   is problematic because the concentration of effluent in
   the water can only be known approximately.

   Sampling actual downstream water.in areas that are in
   or near plumes is especially difficult.  The WER
   obtained is likely to depend on where the sample is
   taken because the WER will probably depend on .the
   percent effluent in the sample (see Appendix D).   The
   sample.could be taken at the end of the pipe,  at the
   edge of the acute mixing zone,  at the edge of the
   chronic mixing zone,  or in a completely mixed
   situation.  If the sample is taken at the edge of a
   mixing zone,  the composition of the sample will
   probably differ from one point to another along the
   edge of the mixing zone.                            •

   If samples.of actual downstream water are to be taken
   close to a discharge,  the mixing patterns and plumes
   should be well known.   Dye dispersion studies
   (Kilpatrick 1992) are commonly used to determine
   isopleths of effluent concentration and complete mix; -
   dilution models (U.S.  EPA 1993d)  might also"be helpful
   when selecting sampling locations.  The most useful
   samples of actual downstream water are probably those
   taken just downstream of the point at which complete
   mix occurs or at the most distant point that is within

                         18

-------
the site to which the site-specific criterion is to
apply.  When samples are collected from a complete-mix
situation, it might be appropriate to composite samples
taken over a cross section'of the stream.  Regardless
of where it' is decided conceptually that a .sample
should be taken, it might .be difficult to identify
where the point exists in the stream and how it changes
with flow and over time.  In addition, if it is not
known exactly what the sample actually represents,
there is no way to know how reproducible the sample is.
These problems make it difficult to relate WERs
determined in actual downstream water to an effluent of
concern because the concentration of effluent in the
sample is not known; this is not a problem, however,, if
the sample-specific WER approach is used to interpret
the results.

Simulated downstream water would seem to be the most
unnatural of the three kinds of water, but it offers
several important advantages because effluent and
upstream water are mixed at a known ratio.  This is
important because the magnitude of the WER will often
depend on the concentration of effluent in the
downstream water.  Mixtures can be prepared to simulate
the ratio of effluent and upstream water that exists at
the edge of the acute mixing zone, at the edge of the
chronic mixing zone, at complete mix, or at any other
point of interest.  If desired, a sample of effluent
can be mixed with a sample on upstream water in
different ratios to simulate different points in a
stream.  Also, the ratio used can be one that simulates
conditions at design flow or at any other flow.

The sample-specific WER approach can be used with both
actual and simulated downstream water.  Additional
quantitative uses can be made of WERs determined using
simulated downstream water because the percent effluent
in the water is known, which allows quantitative
extrapolations to the effluent.  In addition, simulated
downstream water can be used to determine the variation
in the WER .that ,is due to variation in the effluent.-
It also allows comparison of two or more effluents and
determination of the interactions of two or more
effluents.  Additivity of WERs can be studied using
simulated downstream water (see Appendix 6); studies of
toxicity within plumes and studies of whether increased
flow of upstream water can increase toxicity are both •
studies of additivity of WERs.  Use of simulated
downstream water also makes it possible to conduct
controlled studies of changes in WERs due to aging and
changes in pH.
                      19

-------
                             (                         ,
      In Method 1,  therefore, WERs are determined using
      simulated downstream water that is prepared by mixing
      samples of effluent and upstream water in an appropriate
      ratio.   Most  importantly,  Method 1 can be used to .
      determine a WER that applies in the vicinity of a plume
      and can be quantitatively extrapolated to the effluent.

•  2.   Should WERs be determined when the stream flow is equal
     • to,  higher than,  and/or lower than the design flow?

      WERs are used in the derivation of site-specific criteria
      when it is desired that permit limits be based on a
      criterion that takes into account the characteristics of
      the water and/or, the metal at the site.   In most cases,
      permit  limits are calculated using steady-state models and
      are based on  a design flow.   It is therefore important
      that-WERs be  adequately protective under design-flow/
      conditions, which might be expected to require that some
      sets of samples of effluent, and upstream water be  obtained
      when the actual stream flow is close to the design flow.
      Collecting samples when the stream flow is close to the
      design  flow will  limit a WER determination to the  low-flow
      season  (e.g.,  from mid-July to mid-October in some places)
      and to  years  in which the  flow is sufficiently low.

      It is also important,  however,  that WERs that are  applied
      at design flow provide adequate protection at higher
      flows.   Generalizations concerning the impact of higher
      flows on WERs  are difficult  because such flows might (a)
      reduce  hardness,  alkalinity,  and pH,  (b)  increase  or
      decrease the  concentrations  of TOC and TSS,  (c)  resuspend
      toxic and/or .nontoxic metal  from the sediment,  and (d)
      wash additional pollutants into the water.   Acidic
      snowmelt,  for  example,  might lower the WER-both by
      diluting the WER  and by reducing the hardness,  alkalinity,
      and  pH;  if substantial labile metal is present,  the WER
      might be-lowered  more than the concentration of the metal,
      possibly resulting in increased toxicity at flows  higher
      than design flow.   Samples taken at higher flows might
      give smaller WERs because  the concentration of the
      effluent is more  dilute; however,  total  recoverable  WERs
     might be larger if the sample is taken just after  an event
      that greatly increases the concentration of TSS and/or TOC
     because  this might increase  both (1)  the concentration of
     nontoxic particulate metal in the water  and (2)  the
      capacity of the water to sorb and detoxify metal.

     WERs are not of concern when the stream  flow is lower than
     the  design flow because these are acknowledged times  of
     reduced  protection.  Reduced protection  might  not  occur,
     however, if the WER is  sufficiently high when the  flow is
     lower than design flow.

                              20

-------
.3 .;  Which toxicity, tests should be used?

    a. As explained in Appendix D, the magnitude of an
       experimentally determined WER is .likely to depend on
       the sensitivity of the toxicity test used.  This
       relationship between the magnitude of the WER and the
       sensitivity of the toxicity test is due to the aqueous
       chemistry of metals and is not related to the test
     •  organisms or the type of test.  The available data
       indicate that WERs determined with different tests do
       not differ greatly if the tests have about the same
       sensitivities, but the data also support the
       generalization that less sensitive toxicity tests
       usually give smaller WERs than more sensitive tests .
       (see Appendix D).       .           •
    b. When the CCC is lower than the CMC, it is likely that a
       larger WER will result from tests that are sensitive at
       the CCC than from tests that are sensitive at the CMC.
    c. The considerations concerning.the sensitivities of two
       tests should also apply to two endpoints for the same
       test.  For any lethality test,  use of the LC25 is
       likely to result in a larger WER than use of the LC50,
       although the difference might not be measurable in most
       cases and the LC25 is likely to be more variable than
       the LC50.  Selecting the percent effect to be used to
       define the endpoint might take into account (a) whether
       the endpoint is above or below the CMC and/or the CCC  '
       and (b) the data obtained when tests are conducted.
       Once the percent effect is selected for a.particular
       test (e.g., a 48-hr LC50 with 1-day-old fathead       '.
       minnows), the same percent effect must be used whenever
       that test is used to determine a WER for that effluent.
       Similarly, if two different tests with the same species
       (e.g.,  a lethality test and a sublethal test)  have
       substantially different sensitivities, both a cmcWER
       and a cccWER could be obtained with the same species.
    d. The primary toxicity test used in the determination of
       a WER should have an endpoint in laboratory dilution
       water that is close to, but not lower than/ the CMC
       and/or CCC to which the WER is to be applied.
    e. Because the endpoint of the primary test in laboratory
       dilution water cannot be lower than the CMC and/or CCC,
       the magnitude of the WER is likely to become closer to
       1, as the endpoint of the primary test becomes closer to
       the CMC and/or CCC (see Appendix D).
    f. The WER obtained with the primary test should be
       .confirmed with a secondary test that uses a species   -
       that is taxonomically different from the species used
  .     in the primary test.
       1)  The endpoint of the secondary test may be higher or
       ,   lower than the CMC,  the CCC,. or" the endpoint of the
          primary test.          '

             '       '        21      •

-------
2) Because  of  the  limited number  of  toxicity  tests that
   have  sensitivities  near  the CMC or CCC  for a metal,
   it'seems unreasonable  to require  that the  two
   species  be  further  apart taxonomically  than being in
   different orders.
Two different  endpoints with the  same species oust not
be used  as  the primary and  secondary tests, even if one
endpoint is lethal and the  other  is  sublethal.
If more  sensitive  toxicity  tests  generally give larger
WERs than less sensitive  tests, the  maximum value of a
WER will usually be obtained using a toxicity test
whose endpoint in  laboratory dilution water equals the
CMC or CMC.  If such a test is not used, the  maximum
possible WER probably  will  not be obtained.
No rationale exists to.support the idea that  different
species  or  tests with  the same sensitivity will produce
different WERs.  Because  the mode of action might
differ from species to species and/or from effect to
effect,  it  is  easy to  speculate that in some  cases the
magnitude of a WER will depend to some extent oh the
species, life  stage, and/or kind  of  test,  but no data
are available  to support  conclusions concerning the.
existence and/or magnitude  of any such differences.
If the-tests are otherwise  acceptable, both cmcWERs and
cccWERs  may .be determined using acute and/or  chronic
tests and using lethal and/or sublethal endpoints.  The
important consideration is  the sensitivity of the test,
not the  duration,  species,  life stage, or  adverse
effect used.                     :
There" is no reason to  use species that occur  at the
site; they may be  used in the determination of a WER if
desired,,but: .
1) It might be difficult  to determine which of the
   species that occur  at  the site are sensitive to the
   metal and are adaptable  to laboratory.conditions.
2) Species that occur  at  the.site might be harder to
   obtain in sufficient numbers for  conducting toxicity
   tests over  the  testing period.
3) Additional  QA tests will  probably be needed (see
   section C.3.b)  because data are not likely tq.be
   available from  other laboratories for comparison '
   with  the results in laboratory dilution water.
Because  a WER  is a ratio  of  results  obtained with the
same test in two different  dilution waters, toxicity
tests that are used in WET  testing,  for example,  may be
used, even if  the  national  aquatic life guidelines
(U.S. EPA 1985) do not allow use  of  the test  in the
derivation of  an aquatic  life criterion.   Of  course, a
test whose endpoint in laboratory dilution water is
below the CMC  and/or CCC  that is  to be adjusted cannot
be used  as a primary test.        '                    '
                      22

-------
1.' Because there is no rationale that suggest that it
   makes any difference-whether the test is conducted with
   a species that is warmwater or coldwater, a fish or an
   invertebrate, or resident or nonresident at the; site,.
   other than-the.fact that less sensitive tests are
   likely to give smaller WERs, such considerations as the
   availability of test organisms might be important in
   the selection of the test.  Information in Appendix I,
   a criteria document for the metal of concern  (see
   Appendix E), or any other pertinent source might be
   useful when selecting primary and secondary tests.
m. A test in which the test organisms are not fed might
   give a different WER than a test in which the organisms
   are fed just because of the presence of the food (see
   Appendix D) .  This might depend on the metal, the type
   and amount of food, and whether a total recoverable or
   dissolved WER is determined.                •
Different tests with similar sensitivities are expected to
give similar WERs, except for experimental variation.  The
purpose of the secondary test is to provide information
concerning this assumption and the validity of the WER.

Should a cmcWER or a cccWER or both be determined?

This question does not have to be answered if the
criterion for the site contains either a CMC or a CCC but'
not both.  For example, a body of water that is protected
for put-and-take fishing might have only a CMC, whereas a
stream whose design flow is zero might have only a CCC.

When the criterion contains both a CMC and a CCC, the
simplistic way to answer the question is to determine
whether the CMC or the CCC controls the existing permit
limits; which one is controlling depends on (a) the ratio
of the CMC to the CCC, (b) whether the number of mixing
zones is zero, one, or two, and.(c) which steady-state or
dynamic model was used in the calculation of the permit
limits.  A better way to answer the question would be to
also determine how much the controlling value would have
to be changed for the other value to become controlling;
this might indicate that it would not be cost-effective,to
derive, for example, a site-specific CMC (ssCMC) without
also deriving a site-specific CCC  (ssCCC).  There are also
other possibilities:  (1) It might be appropriate to use a.
phased approach, i.e., determine either the cmcWER or the
cccWER and then decide whether to determine the other.
(2) It might be appropriate and environmentally
conservative to determine a WER that can be applied to
both the CMC and the CCC.   (3) It is always allowable to
determine and use both a cmcWER and a cccWER, although
both can be determined only if toxicity tests with
appropriate sensitivities are available.

                         23              -

-------
Because the phased approach can always be used, it is only
important.to decide whether to use a different approach
when its use might be cost-effective*  Deciding whether to
use a different approach and selecting which one to use is
complex because a number of considerations need to be
taken into account:          :
a. Is the CMC equal to or higher than the CCC?
      If the CMC equals the CCC, two WERs cannot be
      determined if they would be determined using the
      same site water, but two WERs. could be determined if
      the cmcWER and' the cccWER would be determined using
      different site waters, e.g., waters that contain
      different concentrations of the effluent.
b. If the CMC is higher than the CCC, is there a toxicity
   test whose endpoint in laboratory dilution water is
   between the CMC and the CCC?
      If the CMC is higher than the .CCC and there is a
      toxicity test whose endpoint in laboratory dilution
      water is between the CMC and the CCC, both a cmcWER '
      and a cccWER can be determined.  If the CMC is
      higher than the CCC but no toxicity test has an
      endpoint in laboratory dilution water between the
      CMC and the CCC, two WERs cannot be determined if
      they would be determined using the same site water;
      two WERs could be determined if they were determined
      using different site waters, e.g., waters that
      contain different .concentrations of the effluent.
c. Was a steady-state or a dynamic model used in the
   calculation of the permit limits?
      It. is complex, but reasonably, clear, how to make a
      decision when a steady-state model was used, but it
      is not clear how a decision should be.made when a
      dynamic model was used.
d. If a steady-state model was used, were one or two
   design flows used, i.e., was the hydrologically based
   steady-state method used or was the biologically based
   steady-state method used?
      When the hydrologically based method is used, one
      design flow is used for both the CMC and the CCC,
      whereas when the.biologically based method is used,
      there is a CMC design flow and a CCC design flow.'
    .  When WERs are determined using downstream water, use
      of the biologically based method will probably cause
      the percent effluent in the site water used in the
      determination of the cmcWER to be different from the
      percent effluent in the site water used in the
      determination of the cccWER; thus the two WERs
     . should be determined using two different site
      waters.  This does not impact WERs determined using
      upstream water.
                         24

-------
e. Is there an acute mixing zone?  Is there a chroniq
   mixing zone?
      1. .When WERs are determined using upstream water,
         the presence or absence of mixing zones has no
         impact; the cmcWER and the cccWER-will both be
         determined using site water that contains zero
         percent effluent, i.e., the two WERs will be
         determined using the same site water.
      2. Even when downstream water is used, whether there
         is an acute mixing zone affects the point of
  •,-'.     application of the CMC or ssCMC, but it does not
         affect the determination of any WER.
      3. The existence of a chronic mixing zone has
         important implications for the determination of
   .      WERs when downstream water is used (see Appendix
         A).  When WERs are determined using downstream
         water, the cmcWER, should be determined using
         water at the edge of the chronic mixing zone,
         whereas the cccWER should be determined using
         water from a complete-mix situation.  (If the
         biologically based method is used, the two
         different design flows should also be taken into
         account when determining the percent effluent
         that should be in the simulated downstream
         water.)  Thus the percent effluent in the site
         water used in the determination of the cmcWER
         will be different from the percent effluent in
         the site water used in the determination of the
         cccWER; this is important because the magnitude
         of a WER will often depend substantially on the
         percent effluent in the water (see Appendix D).
f. In what situations would it be environmentally
   conservative to determine one WER and use it to adjust
   both the cmcWER and the cccWER?.
      Because (1) the CMC is never lower than the CCC and
      (2) a more sensitive test will generally give a WER
      closer to.l, it will be environmentally conservative
      to use a cmcWER to adjust a CCC when there are no
      contradicting considerations.  In this case, a
      cmcWER can be determined and used to adjust both the
      CMC and the CCC.  Because water quality can affect
      the WER, this approach is necessarily valid only if
      the cmcWER and the cccWER are determined in the same
      site water.  Other situations in which it would be
    - environmentally conservative to use one WER to
      adjust both the CMC and the CCC are described below.
These considerations have .one set of implications when
both the cmcWER and cccWER are to be determined using the
same site water, and another set of implications when the
two WERs are to be determined using different site waters,
e.g., when the site waters contain different
concentrations of effluent.

                      -  25   .

-------
When WERs are. determined using upstream water,  the  same
site water is used in the determination of  both the cmcWER
and the cccWER.  Whenever the two WERs  are  determined in
the same site, water,  any difference in  the  magnitude of
the. cmcWER arid the cccWER will probably be  due  to the
sensitivities of the  toxicity tests used.   Therefore:
a.  If more sensitive  toxicity tests generally give  larger
    WERs than  less  sensitive  tests,  the  maximum  cccWER (a
    cccWER determined  with a  test  whose  endpoint equals the
  •  CCC) .will  usually  be  larger than the maximum cmcWER
 .   because the CCC is never  higher than the CMC,.
b.  Because the CCC is never  higher than the CMC, the
    maximum cmcWER  will usually be smaller than  the maximum
   •cccWER and it will be environmentally conservative to
    use.the cmcWER  to  adjust  the CCC.
c.  A cccWER can be determined separately from a cmcWER
  .  only if there is a toxicity test with an endpoint in
    laboratory dilution water that is between the CMC and
    the  CCC.   If no such  test exists or  can  be devised,
    only a cmcWER can  be  determined, but  it  can  be used to
    adjust both the CMC and the CCC.
d.  Unless the experimental variation is  increased,  use of
    a cccWER,  instead  of.  a  cmcWER,.  to adjust the CCC will
  •  usually improve the accuracy of  the resulting site-
    specific- CCC.   Thus a cccWER may be determined and used
 .  whenever desired,  if  a  toxicity  test has an  endpoint in
    laboratory dilution water between the CMC and the CCC.
e. A cccWER cannot be used to  adjust :a CMC  if the cccWER
   was  determined  using  an endpoint that was lower than
  ,,the  CMC  in laboratory dilution water because it will
   probably reduce the level of protection.
f. Even if  there is a toxicity test that has an endpoint
   in laboratory dilution water that is between the CMC
 . and  the  CCC, it  is  not necessary to decide initially
   whether  to determine  a cmcWER  and/or a cccWER.  When
  .upstream water  is  used, it  is always allowcible to
   determine  a cmcWER and use  it  to derive a site-specific
   CMC and a  site-specific CCC  and then decide whether to
   determine  a cccWER.
g. If there is a toxicity test whose endpoint in
   laboratory dilution water is between the CCC and the,
   CMC,  and if this test is used as the secondary test in
   the determination  of  the cmcWER, this test will provide
   information that should be very useful for deciding
   whether to determine  a cccWER in addition to a cmcWER.
   Further, "if it is decided to determine a cccWER,  the
   same two tests used in the determination of  the cmcWER
   could then be used in the determination of-the cccWER,
   with a reversal of their roles as primary and secondary
   tests.  Alternatively, a cmcWER and a cccWER could be
   determined simultaneously if both tests are  conducted
   on each sample of site water.

                         26

-------
 When WERs are determined using downstream water.  the
 magnitude of each WER will probably depend on the
 concentration of effluent in the downstream water used
 (see Appendix ,.D) .   The. first important consideration is
 whether the design flow is greater than zero, and the
 second is whether there is a/chronic mixing zone.
 a. If the design flow is zero,  cmcWERs and/or cccWERs that
    are determined for design-flow conditions will both be
    determined in 100 percent effluent.  Thus this case is
    similar to using upstream water in that both WERs are
:    determined in the same site water.  When WERs  are
    determined for high-flow, conditions,  it will make a  -
    difference whether a chronic mixing zone needs to be
    taken into account,  which is the second consideration.
 b. If there is no  chronic mixing zone/ both WERs  will be
    determined for  the complete-mix situation; this case ,is
    similar to using upstream water in that both WERs are
    determined using the same site water.   If there is a
    chronic mixing  zone,  cmcWERs should be determined in
    the site water  that  exists at the edge of the  chronic
    mixing zone,  whereas cccWERs should be determined for
    the complete-mix situation (see Appendix A).   Thus the
    percent effluent will be higher in the site water used
    in the determination of the  cmcWER than in the site
    water used in the determination of the cccWER.   Because
    a site water with a  higher percent effluent will
    probably give a larger WER than a site water with a
    lower percent effluent,  both a cmcWER and a cccWER can
    be determined even if there  is no test whose endpoint
    in laboratory dilution water is between the CMC and the
    CCC.   There are opposing considerations,  however:
    1)  The site water used in the determination of the
       cmcWER will  probably have a higher  percent  effluent
       than the site water used  in the determination of the
       cccWER,  which will tend to cause the cmcWER to be
       larger than  the cccWER.          ,
    2)  If-there is  a toxicity test whose endpoint  in
       laboratory dilution water is between the CMC and the
       CCC,  use of  a more sensitive test in the
       determination of  the cccWER will tend to cause the
       cccWER to be larger than  the cmcWER.
One consequence of these opposing considerations  is  that
it  is  not known whether  use of  the cmcWER to adjust  the
CCC would be environmentally conservative;  if this
simplification is  not known to  be conservative, it should
not be used.   Thus it is important whether there  is  a
toxicity test whose endpoint in laboratory dilution  water"
is  between the CMC and the CCC:
a.  If  no toxicity  test has an endpoint in laboratory
    dilution water  between the CMC and the CCC, the two
  7  WERs  have to  be determined with the same test,  in which
    case  the cmcWER will  probably be larger because the

                         27            '

-------
 •" percent' effluent in the site water will be higher.
   Because of the difference in percent effluent in the
   ..site waters that should be used in the determinations
   of the two.WERs, .use of the cmcWER to adjust the CCC
   would not 'be environmentally conservative, but use of
   the cccWER to adjust the .CMC would be environmentally
   conservative.  Although both WERs could be determined,
   if would also be acceptable to determine only the
   cccWER and use it to adjust both the CMC and the CCC.
b. If there is a toxicity test whose endpoint in
   laboratory dilution water is between the CMC and the
   CCC, the two WERs could be determined using different
   toxicity tes.ts. . An environmentally conservative
   alternative to determining two WERs would be to
   determine a hybrid WER by using (1) a toxicity test
   whose endpoint is above the CMC (i.e., a toxicity test
   that is appropriate for the determination of a cmcWER)
   and (2) site water for the complete-mix situation
    (i.e.; site water appropriate for the determination of
   cccWER) .  It would be environmentally conservative to
   use this hybrid WER to. adjust the CMC and it would be
   environmentally conservative to use this hybrid WER to
   adjust the CCC.  Although both WERs could be
   determined, it would also be acceptable to determine
   only the hybrid WER and use it to adjust both the CMC
   and the CCC..  (This hybrid WER described here in
   paragraph b is the same as the cccWER described in
   paragraph a above in which no toxicity test had an
   endpoint in laboratory dilution water between the CMC
   and the CCC.)

How should a FWER be derived?

Background                       '-.-."

Because of experimental variation and variation in the
composition of surface waters and effluents, a single
determination of a WER does not provide sufficient
information to justify adjustment of a criterion.  After a
sufficient number of WERs have been determined in an
acceptable manner, a Final Water-Effect Ratio (FWER) is'
derived from the WERs, and the FWER is then used to
calculate the site-specific criterion.  If both a site-
specific CMC and a site-specific CCC are to be derived,
both a cmcFWER and a cccFWER have to be derived, unless an
environmentally conservative estimate is used in place of
the cmcFWER and/or the cccFWER.   .                       •

When a WER is determined using upstream water, the two
major sources of variation in the WER are .(.a) variability
in the quality, of the upstream water, much of which might'
be related to season and/or flow, and (b) experimental

                         28

-------
 variation.   When a WER is  determined in downstream water,
 the four major sources of  variation are (a)  variability in
 the quality of the upstream water,  much of which might  be
 related to  season and/or flow,  (b)  experimental  variation,
 (c)  variability in the composition  of the effluent,  and
 (d)  variability in the percent  effluent in the downstream
 water.   Variability and the,possibility of mistakes and
 rare events make it necessary, to  try to compromise between
 (1)  providing  a high probability  of adequate protection
 and (2)  placing too much reliance on the smallest
 experimentally determined  WER,  which might reflect
 experimental variation,  a  mistake,  or a rare event rather
 than a  meaningful difference in the WER.

 Various ways can be employed to address variability:
 a.  Replication can be  used to reduce the impact  of some
    sources  of  variation and to  verify the importance of
    others .-•-•-.
 b.  Because  variability in  the composition of the effluent
    might contribute substantially to the variability of
    the  WER,  it might be desirable to obtain  and  store two
    or more  samples of  the  effluent  at slightly different
    times, with the selection of the sampling times
    depending on such characteristics of the  discharge as
    the  average retention time,  in case an unusual WER is
    obtained with the first sample used.
 c.  Because  of  the possibility of  mistakes and  rare events,
    samples  of  effluent and upstream water should.be  large
    enough that portions  can be  stored for later .testing or
    analyses  if an unusual  WER is  obtained.
 d.  It might  be possible  to reduce the impact of  the
   variability in the  percent effluent in the  downstream
   water by. establishing a relationship between  the WER
   and  the percent effluent.
 Confounding  6f the sources can  be a problem when more than
 one  source  contributes substantial  variability.

When permit  limits are calculated using a.steady-state
model,  the limits  are  based on  a  design flow,  e.g.,. the
 7Q10.   It is usually assumed that a concentration of metal
 in an effluent that  does not  cause  unacceptable  effects, at
 the  design  flow will not cause  unacceptable effects  at
higher  flows because the metal  is diluted by the increased
 flow of  the  upstream water.   Decreased protection might
occur, however, if an  increase  in flow increases toxicity
more than it dilutes the concentration of metal.   When
permit limits  are  based  on a  national  criterion,  it  is
often assumed  that the criterion  is sufficiently
 conservative that  an increase in  toxicity will not be
great enough to overwhelm  the combination of dilution and
the  assumed  conservatism,  even  though it  is likely that
the national criterion is  not overprotective of  all  bodies

      '                   29       •         '.•••'.'

-------
of water.  When WERs are used to reduce the assumed
conservatism, there is more concern about the possibility
of increased toxicity at flows higher than the design flow
and it is important to  (1) determine some WERs that
correspond to higher flows or  (2) provide some
conservatism.  If the concentration of effluent in the
downstream water decreases as flow increases, WERs
determined at higher flows .are likely to be smaller than
WERs determined at design flow but the concentration of
metal will also be lower.  If the concentration of TSS
increases at high flows/ however, both the WER and the
concentration of metal might increase.  If they are
determined in an appropriate manner, WERs determined at
flows higher than the design flow can be used in two ways:
a. As environmentally conservative estimates of WERs
   determined at design flow.
b. To assess whether WERs determined at design, flow will
   provide adequate protection at higher flows.

In order to appropriately take into account seasonal and
flow effects and their interactions, both ways of using
high-flow WERs require that the downstream water used in
the determination of the WER be similar to that which
actually exists during the time of concern.  In addition,
high-flow WERs can be used in the second way only if the
composition of the downstream water .is known.  To satisfy
the requirements that (a) the downstream water used in the
determination of a WER be similar to the actual water and
(b) the composition of the downstream water be known, it
is necessary to obtain samples of effluent and upstream
water at the time of concern and to prepare a simulated
downstream water by mixing the samples at the ratio of the
flows of the effluent and the upstream water that existed
when the samples were obtained.

For the first way of using high-flow WERs, they are used
directly as environmentally conservative estimates of the
design-flow WER.  For the second way of using high-flow
WERs, each is used to calculate the highest concentration
of metal that could be in the effluent without, causing the
concentration of metal in the downstream water to exceed
the site-specific criterion that would be derived for that
water using the experimentally determined WER.  This
highest concentration of metal in the effluent (HCME) can
be calculated as :                                        •

        [(CCO (WER) (GFLCW + UFLOW)] - [ (uCONC)' (uFLOtf) ]
                               - -- -
where:
CCC =•   the national, state, or recalculated CCC  (or CMC)
        that is to be adjusted.

                         30

-------
 eFLOW = the flow of the effluent that was the basis of the
         preparation of the simulated downstream water.
         This should be the flow of the effluent that
         existed when the samples were taken.
 uFLOW = the flow of the upstream water that was the basis
        -of the preparation of the simulated downstream
      .,  water.   This should be the flow of the upstream
         water that  existed when the samples were taken.
 uCONC = the concentration of metal in the sample of
         upstream water used in the preparation of the
         simulated. downstream water.
 In order to calculate a HCME from an experimentally
 determined WER,  the only information needed besides the
 flows of the effluent and the upstream water  is the
 concentration of metal in the upstream water,  which should
 be measured anyway  in conjunction with the determination
 of the  WER.                                     •.-    .   ,

 When  a  steady-state model is used to derive permit limits,
 the limits on the effluent apply at all flows;  thus,  each
 HCME  can be used to calculate the highest WER (hWER)  that
 could be used to derive a site-specific criterion for the
 downstream water at -design flow so that there would be
 adequate protection at the flow for which the HCME was
 determined.   The hWER is calculated as:
      hWER = (gCME) < OFLOtfdf)  •*• ( uCONCdf ) ( uFLOWdf)
                  (CCC) (eFLOWdf
The suffix  "df" indicates that the values used for these
quantities  in the calculation of the hWER are those that
exist at design-flow conditions.  The additional datum
needed- in order to calculate the hWER is the concentration
of metal in upstream water at design^flow conditions; if
this is assumed to be zero, the hWER will be
environment ally conservative.  If a WER is determined when
uFLOW equals the design flow, hWER = WER.

The ,two ways of using WERs determined at flows higher than
design flow can be illustrated using, the following
examples.   These examples were formulated using the
concept of  additivity of WERs (see Appendix G) .  A WER
determined  in downstream water consists of two components,.
one due to  the effluent (the eWER) and one due to the
upstream water (the uWERK  If the eWER and uWER are
strictly additive, when WERs are determined at various
upstream flows, the downstream WERs can be calculated from
the composition of the downstream water (the % effluent
and the % upstream water)  and the two WERs (the eWER and
the uWER) using the equation:
                         31

-------
   HER «= (% effluent) (&WER) + (% upstream water) (uffSR)
                           100
              •                                     *
In the  examples below,  it .is assumed that:
a. A  site-specific CCC  is being derived.
b; The  national CCC is  2 ug/L.
c. The  eWER is  40.                              .
d. The  eWER and uWER are constant and strictly additive.
e; The  flow of  the effluent (eFLOW)  is always 10 cfs.
f. The  design flow of the upstream water (uFLOWdf) is 40
   cfs.
Therefore :                      .

      [(2 ug/L) (ttER) (10 CfB •«• UFLO»).3 -  [ ( uOQMQ ( uFLOM 1
                          10 ug/L     '               *
                   (10 cfs) * (uCONCdf) UP c£g)
                  (2 UI3/.L) (10 cfs •*• 40 CfB)      *


In the first  example,  the uWER is assumed to be 5 and so
the upstream  site-specific CCC (ussCCC)  = (CCC) (uWER) =
(2 ug/L) (5) = 10  ug/L.  uCONC is assumed to be 0.4 ug/L,
which means that  the assimilative capacity of the upstream
water is  9.6  ug/L.

eFLOW   uFLOW    At. Complete Mix        HCME      hWER
(cfs)    (cfs)   % Eff .  %  UPS.   WER     (ua/L)     _

 10       40      20.0   80.0   12.000     118.4     12.00
10
10
10
10.
10
10
63
90
190
490
990
' 1990
13.7
10.0
5.0
2.0
1.0
0.5
86.3
90.0
95.0
98.0
99.0
99.5
9.795
8.500
6.750
5.700
5.350
5.175
140.5
166.4
262.4
550.4
1030.4
1990.4
14.21
16.80
26.40
55.20
103.20
199.20
As' the flow of  the  upstream water increases,  the WER
decreases to a  limiting value equal to uWER.   Because the
assimilative capacity is greater than zero, the HCMEs and
hWERs increase  due  to the increased dilution of the
effluent.  The  increase in hWER at higher flows will not
allow any use of  the assimilative capacity of the upstream
water because the allowed concentration of metal in the
effluent is controlled by the lowest hWER, which is the  ..
design-flow hWER  in this example.  Any WER determined at a
higher flow can be  used as an environmentally conservative
estimate of the design-flow WER,  and the hWERs show that
t'he WER of 12 provides adequate protection at all .flows.
When uFLOW equals the design flow of 40 cfs,  WER = hWER.

                         32

-------
In the second example/ uWER is assumed to be 1, which
means that ussCCC = 2 ug/L.  uCONC is assumed to be 2
ug/L, so/that uCONC = ussCCC.  The assimilative capacity
of the upstream water is 0 ug/L.

eFLOW   uFLOW     At Complete Mix        HCME     hWER
(cfs)  • (cfs)   % Eff. % UPS.   WER      (ua/L)    	

 10       40     20.0.   80.0   8.800     80.00    8.800
 10     .  63     13.7   86.3   6.343     80.00    8.800
 10       90     10.0   90.0   4.900     ;80.00    8.800
 10      190      5.0   95.0   2.950     80.00    8.800
 10     .490      2.0   98.0   1.780   ;  80.00    8.800
 10      990      1.0   99.0   1.390"    80.00    8.800
 10     1990      0.5   99.5   1.195  -   80.00    8.800

All the WERs in this example are lower than the comparable
WERs in the first example because the uWER dropped from 5
to 1; the limiting value of the WER at very high flow is
1.  Also/ the HCMEs and hWERs are independent of. flow
because the increased dilution does not allow any more
metal to be discharged when uCONC = ussCCC, i.e., when the
assimilative capacity is zero.  As in the first example,
any WER determined at a flow higher than design flow .can
be used as an environmentally conservative estimate of the
design-flow WER and the hWERs show that the WER of 8.8
determined at design flow will provide adequate protection
at all flows for which information is available.  When
uFLOW equals the design flow of 40 cfs, WER = hWER.


In the third example, uWER is assumed to be 2, which means
that ussCCC = 4 ug/L.  uCONC is assumed to be 1 ug/L; thus
the assimilative capacity of the upstream water is 3 ug/L.

eFLOW   uFLOW     At Complete Mix        HCME     hWER
(cfs)   (cfs)   % Eff. % UPS.   WER    .  (uo/L)    	

 10       40     20.0   80.0   9.600      92.0     9.60
 10       63     13.7   86.3   7.206      98.9    10.29
 10       90     10.0   90.0   5.800  .107.0    11.10.
 10     ^190      5.0   95.0   3.900     137.0    14.10
 10      490      2.0  .98.0   2.760     227.0    23.10
 10      990      1.0   99.0   2.380     377.0    38.10
 10    .1990      0.5   99.5   2.190     677.0    68.10

All the WERs in this example are intermediate between the
comparable WERs in the first two examples because the uWER
is now 2, which is between 1 and 5; the limiting value of
the WER at very high flow is 2.  As in the other examples,
any WER determined at a flow higher than design flow can
be used as an environmentally conservative estimate of the

                         33   .  .             .            '

-------
 design-flow WER and the hWERs show that  the WER of  9.6
 determined at design flow will provide adequate protection
 at all flows for which information is  available.  When
 uFLOW equals .the design flow.of 40.cfs,  WER = hWER.

 If this third example is assumed to be subject  to acidic
 snowmelt in the spring so that the eWER  and uWER are  less-
 than-additive and result in a WER of 4.8 (rather than 5.8}
 at a uFLOW of 90 cfs,  the third HCME would be 87 ug/L, and
 the third hWER would be 9.1,  This hWER  is lower than the
 design-flow WER of 9.6,  so the site-specific criterion
 would have to be derived using the WER of  9.1,  rather than
 the design-flow WER- of 9.6,  in order to  provide the
 intended level of protection.   If the  eWER and  uWER were
 less-than-additive only to the extent  that the  third  WER
 was 5.3,  the third HCME would be 97 ug/L and the third
•hWER would be 10.1.   In this case,  dilution by  the
 increased flow would more than compensate  for the WERs
 being less-than-additive,  so that the  design-flow WER of
 9.6 would provide adequate protection  at a uFLOW of 90
 cfs.  Auxiliary information might indicate whether an
 unusual WER is real or is an accident; for example, if the
 hardness,  alkalinity,, and pH of snowmelt are all low, this
.information would support a low WER.

 If the eWER and uWER were more-than-additive so that  the
 third WER was 10,  this WER would not be  an environmentally
 conservative estimate of the .design-flow WER.   If a WER
 determined at a higher flow is to be used  as an estimate
 of the design-flow WER and there is reason to believe that
•the eWER and the uWER might be more-than-additive, a  test
 for additivity can be performed (see Appendix 6).

 Calculating HCMEs and hWERs is straightforward  if the WERs
 are based on the total recoverable measurement.  If they
 are'based on the dissolved measurement,  it is necessary to
 take into account the percent of the total recoverable  .
 metal in the effluent that becomes dissolved in the
 downstream water.

 To ensure adequate protection,  a group of  WERs  should '
 include one or more WERs corresponding to  flows near  the
 design flow,  as well as one or more WERs corresponding to
 higher flows.
 a. Calculation of hWERs from WERs determined at various
    flows and seasons identifies the highest WER that  can
    be used in the 'derivation of a site-specific criterion'
    and still provide adequate protection at all flows for
    which WERs are available.  Use of hWERs eliminates the
    need to assume that WERs determined at  design flow will
    provide adequate protection at higher flows. Because
    hWERs are calculated to apply at design flow, they

                         34

-------
 •'  apply to the flow on which the permit limits are based.
   The lowest of the hWERs ensures adequate protection at
   all flows, if hWERs are.available -for a sufficient
   range of flows, seasons, and other conditions. .  '
b. Unless addltivity is assumed, a WER cannot be
   extrapolated from one flow- to another and therefore it-
   is not possible to predict a design-flow WER from a WER
   determined at other conditions.  The largest WER is
   likely to occur at design flow because, of the flows
   during which protection is to be provided, the design
   flow is the flow at which the highest concentration of
   effluent will probably occur in the downstream water.
   This largest WER has to be experimentally determined;
   it cannot be'.predicted.

The examples also illustrate that if the concentration of
metal in the upstream water is below the site-specific
criterion for that water, in the limit of infinite
-dilution of the effluent with upstream water, there will
be adequate protection.  The concern, therefore, is for
intermediate levels of dilution*  Even if the assimilative
capacity is zero, as in the second example, there is more
concern at the lower or intermediate flows, when the
effluent load is.still a major portion of the total load,
than, at higher flows when the effluent load is a minor
contribution.
The Options

To ensure adequate protection over a range of flows, two
types of WERs need to be determined:
Type 1 WERs are determined by obtaining samples of
         effluent and upstream water when the downstream
         flow is between one and two times higher than
         what it would be under design-flow conditions.
Type 2 WERs are determined by obtaining samples of
         effluent and upstream water when the downstream
         flow is between two and ten times higher than
         what it would be under design-flow conditions.
The only difference between the two types of samples is.
the downstream flow at the time the samples are taken.
For both types of WERs, the samples should be mixed at the
ratio of the flows that existed when the samples were
taken so that seasonal and flow-related changes in the
water quality characteristics of the upstream water are
properly related to the flow at which they occurred.  The
ratio at which the samples are mixed does not have to be
the exact ratio that existed when the samples were taken,
but the ratio, has to be known, which is why simulated
downstream water is used.  For each Type 1 WER and each
Type 2 WER that is determined, a hWER is calculated.

                         35

-------
Ideally, sufficient numbers of both types of WERs would be
available and each WER would be sufficiently precise and
accurate-and the Type 1 WERs would be sufficiently similar
that the FWER,. could be the geometric mean of the Type 1
WERs, unless 'the FWER had to be lowered because of one or
more hWERs.  If an adequate number of one or both types of
WERs is not available, an environmentally conservative WER
or-hWER should be used as the FWER.

Three Type 1 and/or Type 2 WERs, which were determined
using acceptable procedures and for which there were at
least three,weeks between any two sampling events, must be
available in order for a FWER to be derived.  If three or
more are available, the FWER should be derived from the
WERs and hWERs using the lowest numbered option whose
requirements are satisfied:
1. If there are two or more Type 1 WERs:
   a. If at least nineteen percent of all of the WERs are
      Type 2 WERs, the derivation of the FWER depends on
      the properties of the Type 1 WERs:
      1) If the range of the Type 1 WERs is not greater
         than a factor of 5 and/or the range of the ratios
         of .the Type 1 WER to the concentration of metal
         in the simulated downstream water is not greater
         than a factor of 5, the FWER is the lower of (a)
         the adjusted geometric mean (see Figure 2) of all
         of the Type 1 WERs and (b) the lowest hWER.
      2). If the range of the Type 1 WERs is greater than a
         factor of 5 and the range of the ratios of the
         Type 1 WER to the concentration of metal in the
         simulated downstream water is greater than a
         factor of 5, the FWER is the lowest of (a) the
         lowest Type 1 WER, (b) the lowest hWER, and (c)
         the geometric mean of all the Type 1 and Type 2
         WERs, unless an analysis of the joint
         probabilities of the occurrences of WERs and
         metal concentrations indicates that a higher WER
         would still provide the level of protection
         intended by the criterion.  (EPA intends to
         provide guidance concerning such an analysis.)
   b. If less than nineteen percent of all of the WERs are
      Type 2 WERs, the FWER is the lower of (1) the lowest
      Type 1 WER and (2) the lowest hWER.
2. If there is one Type 1 WER, the FWER is the lowest of
   (a) the Type 1 WER, (b) the lowest hWER,  and (c) the
   geometric mean of all of the Type 1 and Type 2 WERs.
3. If there are no Type 1 WERs, the FWER is the lower of •
   (a) the lowest Type 2 WER and (b)  the lowest hWER.
If fewer than three WERs are available .and a site-specific
criterion is to be derived using a WER or a FWER, the WER
or FWER has to be assumed to be 1.  Examples of deriving
FWERs using these options are presented in Figure 3.

                       .  36

-------
 The  options  are  designed  to  ensure  that:
 a. The  options apply  equally well to  ordinary  flowing
   waters  and to streams  whose  design flow  is  zero.
 b. The  requirements for deriving the  FWER as something
   other than the lowest  WER are not  too  stringent.
 c. The  probability is high that the criterion  will be
   adequately protective  at  all flows,  regardless of the
   amount  of data that are available.
 d. The  generation of  bpth types of  WERs is  encouraged
   because environmental  conservatism is  built in if both
   types of  WERs are  not  available  in acceptable numbers.
 e. The  amount of conservatism decreases as  the quality and
   quantity  of the.available data increase.
 The  requirement  that  three WERs be  available is based on a .
 judgment 'that fewer WERs  will not provide sufficient
 information; The requirement that  at least nineteen
 percent of all of the available WERs  be Type 2 WERs is
 based on a judgment concerning  what constitutes an
 adequate mix of  the two types of WERs:  when there are five
 or more WERs, at least one-fifth should be  Type 2 WERs.

 Because each of  these options for deriving  a FWER is
"expected to  provide adequate protection,  anyone who
 desires to determine  a FWER  can generate  three or more
 appropriate  WERs and  use  the option that  corresponds to
 the  WERs that are available. The options that utilize the
 least useful WERs are expected  to provide adequate
 protection because of the way the FWER is derived  from the
 WERs.  It  is intended that,  on  the  average, Option la will
 result  in  the highest FWER,  and so  it is  recommended that
 data generation  should be designed  to satisfy  the
 requirements 6f  this  option  if  possible.  For  example, if
 two  Type 1 WERs  have  been determined, determining  a third
 Type 1  WER will  require use  of  Option Ib, whereas
 determining  a Type 2  WER  will require use of Option la.

 Calculation  of  the FWER as an adjusted geometric mean
. raises  three issues:
 a. The  level of  protection would be greater if the lowest
   WER, rather  than  an adjusted mean, were  used as the
   FWER.  Although true,  theTintended level of protection
   is provided by the national  aquatic life criterion
   derived according to the  national guidelines; when
    sufficient data are available  and it is  clear how the
    data should be used,  there is  no reason  to add a
    substantial  margin of  safety and thereby change the
    intended level of protection.   Use of an adjusted
    geometric mean is acceptable if sufficient data are•
    available concerning the WER to demonstrate that the
    adjusted geometric mean will provide the intended level
    of protection.  Use of the lowest of three or more WERs
   would be justified, if,  for example, the criterion had

   1   '  -   •        "       37          -.         "   .'.'.'"'.

-------
    been lowered to protect a commercially important
    species  and a WER determined with that species was
    lower than WERs determined with other species.
b.  The level  pf protection would be greater if  the
    adjustment' was to a probability of 0.95 rather than to
    a probability of 0.70.   As above,  the intended level of
  .  protection is provided  by the national aquatic life
    criterion  derived according to the national  guidelines.
    There is no need to substantially increase the level of
    protection when site-specific criteria are derived.
c.  It  would be easier to use the, more common arithmetic
    mean,  especially because the geometric mean  usually
    does not provide much more protection than the
    arithmetic mean.   Although true,  use  of the  geometric
 •   mean rather than the arithmetic mean  is justified on
    the basis  of statistics and mathematics;  use of the
  •  geometric  mean is also  consistent  with the intended
    level of protection.  Use of the arithmetic  mean is
    appropriate when the values can range  from minus
    infinity to plus infinity.   The geometric mean (GM) is
    equivalent to using the arithmetic mean of the
    logarithms of the values.   WERs cannot  be negative, but
    the logarithms of WERs  can.   The distribution of the
    logarithms of WERs is therefore more likely  to be
    normally distributed than is the distribution of the
 •   WERs.  Thus,  it is better to use the GM of WERs.  In
    addition,  when dealing  with quotients,  use of the GM
    reduces  arguments about the correct way to do some
    calculations because the same answer is obtained in
    different  ways.   For example,  if WER1  = (Nl)V(Dl) and
    WER2  = (N2)/(D2),  then  the GM of WER1  and WER2 gives
    the same value as [ (GM  of Nl and N2)/(GM of  Dl and D2)]
    and also equals the square root of
 .   {[(Nl) (N2)]/[(D1HD2)]}.

Anytime the FWER is  derived as  the lowest  of a  series of
experimentally determined  WERs  and/or hWERs, the magnitude
of  the FWER will depend at least  in part  on  experimental
variation.  There are at least  three  ways  that  the
influence of  experimental  variation on the FWER can be
reduced:                                           '    •
a. A WER determined with a primary test can  be  replicated
    and the .geometric mean  of the  replicates  used as the
   value of the WER for that determination.  Ttien the FWER
   would be the lowest  of  a number of geometric means
   rather than the lowest  of a  number of  individual WERs.
   To be true replicates,  the replicate determinations of
    a WER should not  be based on the same test in
    laboratory dilution water,  the same sample of site
   water, or  the same sample of effluent.
b. If,   for  example,  Option 3  is to be used with three Type
   2 WERs and the endpoints of  both the primary and

                         38

-------
       secondary tests in laboratory dilution water are above
       the CMC and/or CCC to which the WER is to apply, WERs
       can be determined with both the primary and secondary
       tests for each of the three sampling times.  For each
       sampling time, the geometric mean of the WER obtained
       with the primary test and ..the WER obtained with the
       secondary test could be calculated; then the lowest of
       these three geometric means could be used as the FWER.
       The three WERs cannot .consist of some WERs determined
       with one of the tests and some WERs determined with the
       other test; similarly the three WERs cannot consist of
       a combination of individual WERs obtained with the
       primary and/or secondary tests and geometric means of
       results of primary and secondary tests.
    c. "As mentioned above, because the variability of the
       effluent might contribute substantially to the
       variability of the WERs, it might be desirable to
       obtain and store more than one sample of the effluent
       when a WER is to be determined in case an unusual WER
       is obtained''with the first sample used.
    Examples of the first and second ways of reducing the
    impact of experimental variation are presented in Figure
    4.  The availability of these alternatives does not mean
    that they are necessarily cost-effective.

6.  For metals whose criteria are hardness-dependent, at what
    hardness should WERs be determined?

    The issue of hardness bears on such topics as acclimation
    of test organisms to the site water, adjustment of the
    hardness of the site water, and how an experimentally
    determined WER should be used.  If all WERs were
    determined at design-flow conditions, it might seem that
    all WERs should be determined at the design-flow hardness.
    Some permit limits, however, are not based on the hardness
    that is most likely to occur at design flow; in addition,
    conducting all tests at design-flow conditions provides no
    information concerning whether adequate protection will be
    provided at other flows.  Thus, unless the hardnesses, of
    the upstream water and the effluent are similar and do not
    vary with flow,  the hardness of the site water will not, be
    the same for all WER determinations.

    Because the toxicity tests, should be begun within 36 hours
    after the samples of effluent and upstream water are
    collected, there is little time to acclimate organisms to
    a sample-specific hardness.  One alternative would be to •
    acclimate the organisms to a preselected hardness and then
    adjust the hardness of the site water, but adjusting the
    hardness of the site water might have various effects on
    the toxicity of the metal due to competitive binding and
    ionic impacts on the test organisms and on the speciation

              ."••->'     39              "

-------
  f   •  '    "                              '         >»,
 of the metal; lowering hardness without also diluting the
 WER-is especially problematic.  The least objectionable
 approach -'is to acclimate the organisms to a laboratory
 dilution water with a hardness in the range of 50 to 150
 mg/L and then' use this water as the.laboratory dilution
 water when the WER is determined.  In this way,  the test
 organisms  will be acclimated to the laboratory dilution
 water'as specified by ASTM (1993a,b,c,d,e).

 Test organisms may be acclimated to.the site water for a
 short  time as long as this does not cause the tests to
 begin more than 36 hours after the samples, were collected.
 Regardless of what acclimation procedure is used,  the
 organisms  used for the toxicity test  conducted using site
 water are  unlikely to. be acclimated as well as would be'
 desirable.  This is a general problem with toxicity tests
 conducted  in site water (U.S. EPA 1993a,b,c;  ASTM 1993f),
 and its impact on the results of tests is unknown.

 For the practical reasons given above,  an experimentally
 determined WER will usually be a ratio of endpoints
 determined at two different hardnesses and will  thus
 include contributions from a variety  of differences
.between the two waters,  including hardness.   The
 disadvantages of differing hardnesses are that (a)  the
 test organisms probably will not be adequately acclimated
 to  site water and (b)  additional calculations will  be
 needed to  account for the differing hardnesses;  the
 advantages are that it allows the generation of  data
 concerning the adequacy of protection at various flows of
 upstream water and it provides a way  of overcoming  two
 problems with the hardness equations:  (1)  it  is  not known
 ho*w applicable they are to hardnesses outside the range of
 25'to  400  mg/L and (2)  it is not known how applicable they
 are to unusual combinations of hardness,  alkalinity,  and
 pH  or  to unusual ratios of calcium and magnesium.

 The additional calculations that are  necessary'to account
 for the differing hardnesses will also overcome  the
 shortcomings  of the hardness equations.   The  purpose of
 determining a WER is to determine how much metal can be- in
 a site water  without lowering the intended level of
protection.   Each experimentally determined WER  is
 inherently referenced to the hardness  of  the  laboratory
 dilution water that was used in the determination of the
WER, but the  hardness  equation.can be  used to calculate
adjusted WERs that are referenced to  other.hardnesses for-
the laboratory dilution water.   When used to  adjust WERs,
a hardness equation for a CMC or CCC  can be used to
reference  a WER to any hardness for a  laboratory dilution
water, whether it  is inside or outside  the range of 25 to
400 mg/L,  because  any inappropriateness in the equation

                         40       -  ' •'            ..-•'.

-------
will be automatically compensated for when the adjusted
WER is used in the derivation of .a FWER and permit limits.

For.example, the hardness equation for the freshwater CMC
for copper gives CMCs of 9.2, 18, and 34 ug/L at
hardnesses of 50, 100, and 200 mg/L, respectively.  If
acute toxicity tests with Ceriodaphnia reticulata gave an
EC50 of 18 ug/L using a laboratory dilution water with a
hardness of 100 mg/L and an EC50 of 532.2 ug/L in a site
water, the resulting WER would be 29.57.  It can be
assumed that, within experimental variation, ECSOs of 9.2  .
and 34 ug/L and WERs of 57.85 and 15.65 would have been
obtained if laboratory dilution waters with hardnesses of
50 and 200 mg/L, respectively, had been used, because the
EC50 of 532.2 ug/L obtained in the site water does not
depend on what water is used for the laboratory dilution
water.  The WERs of 57.85 and 15.65 can be considered to
be adjusted WERs that were extrapolated from the
experimentally determined WER using the hardness equation
for the copper CMC.  If used correctly, the experimentally
determined WER and all of the adjusted WERs will result in
the same permit limits because they are internally
consistent and are all based on the EC50 of 532.2 ug/L
that was obtained in site water.

A hardness equation for copper can be used to adjust the
WER if the hardness of the laboratory dilution water used
in the determination of the WER is in the range of 25 to
400 mg/L  (preferably in the range of about 40 to 250 mg/L
because most of the data used to derive the equation are
in this range).  However, the hardness equation can be
used to adjust WERs to hardnesses outside the range of 25  -
to 400 mg/L because the basis of the adjusted WER does not
change the fact that the EC50 obtained in site water was
532.2 ug/L.  If the hardness of the site water was 16
mg/L, the hardness equation would predict an EC50 of 3.153
ug/L, which would result in an adjusted WER of 168.8.
This use of the hardness equation outside the range of 25
to 400 ma/L is valid only if the calculated CMC is used
with the corresponding adjusted WER.  Similarly, if the
hardness of the site water had been 447 mg/L, the hardness
equation would predict an EC50 of 72.66 ug/L, with a
corresponding adjusted WER of 7.325.  If the hardness of
^447 mg/L were due to an effluent that contained calcium
chloride and the alkalinity and pH of the site water were
what would usually occur at a hardness of 50 mg/L rather
than 400 mg/L, any inappropriateness in the calculated
EC50 of 72.66 ug/L will be compensated for in the adjusted
WER of 7.325, because the adjusted WER is based on the    §;
EC50 of 532.2 ug/L that was obtained using the site water.'
                         41

-------
    In the above examples  it was  assumed that  at  a hardness  of
    100 mg/L  the EC50  for  C_. reticulata equalled  the CMC,
    which is"a very reasonable simplifying assumption. .If,
    however,  the VJER had been determined with  the more.
    resistant Daphnia  pulex and ECSOs  of 50 ug/L  and 750 ug/L
    had been  obtained  using a laboratory dilution'water and  a
    site  water,  respectively,  the CMC  given by th€s hardness
    equation  could not be  used as the  predicted EC50.  A new
    equation  would have  to be derived  by changing the
    intercept so that  the  new equation gives an EC50 of 50
    ug/L  at a hardness of  100 mg/L;  this new equation could
    then  be used to calculate adjusted ECSOs,  which  could then  .
    be used to. calculate corresponding adjusted WERs:

           Hardness         EC50          WER
             fmcr/L)          (ua/L)             .
               16             8.894         84.33
               50            26.022         28.82
              100            50.000*        15.00*
              200            96.073          7.81
              447           204.970          3.66

    The values marked with an asterisk  are the assumed
    experimentally determined values; the  others were
    calculated from these values.  At each hardness  the
    product  of the EC50  times the WER equals  750 ug/L because
    all of the WERs are  based on the same  EC50 obtained using
    site water. Thus use of the WER allows application of  the
    hardness equation for a metal to conditions to which  it
    otherwise might not  be applicable.

    HCMEs can then be calculated using  either the
    experimentally determined WER or an adjusted WER as long
    as  the WER is  applied to the CMC that  corresponds to  the
    hardness on which the WER is based.  For  example, if  the
  •  concentration  of copper in the upstream water was 1 ug/L
    and the  flows  of the effluent and upstream water were 9
    and 73 cfs, respectively, when the  samples were  collected,
    the HCME calculated  from the WER of 15.00 would  be:
                                                           r
HCME *  (17.73 Uff/L) (15) (9  •«• 73 cfs) - (1 ug/L) (73 cfs) m 2415    ^
                         9 Cfs

    because  the CMC is 17.73 ug/L at a  hardness of 100 mg/L.
    (The value of  17.73  ug/L is used for the  CMC instead  of 18
    ug/L to  reduce roundoff error in this  example.)  If the  •
    hardness of the site ,water was actually 447 ug/L, the HCME
    could also be  calculated using the  WER of 3.66 and the  CMC
    of  72.66 ug/L  that.would be obtained from the CMC hardness
    equation:            .


                             42

-------
HCME =  (72.66 Uff/L) (3.66) (9 +-73 Cfs) - (1 Uff/L) (73 cfs)
     Either WER can be used in the calculation of  the HCME  as
     long -as the CMC and the WER correspond to the same
     hardness and therefore to each other, because:

              (17.73 Uff/L) (15) = (72.66 Uff/L) (3.66) ... "  .

     Although the HCME will be correct as long as  the hardness,
     CMC,- and WER correspond to each other, the WER used in the
     derivation of the FWER must be the one that is calculated
     using a hardness equation to be compatible with the
     hardness of the site water.  If the hardness  of the site
     water was 447 ug/L, the WER used in the derivation of  the
     FWER has to be 3.66; therefore, the simplest  approach  is
     to calculate the HCME using the WER of 3.66 and the
     corresponding CMC of 72.66 ug/L, because these correspond
     to the hardness of 447 ug/L, which is the hardness of  the
     site water.
                :                     :

     In contrast, the hWER should be calculated using the CMC
     that corresponds to the design hardness... If  the design
     hardness is 50 mg/L,  the corresponding CMC is 9.2 ug/L.
     If the design flows of the effluent and the upstream water
     are 9 and 20 cfs, respectively, and the concentration  of
     metal in upstream water at design conditions  is 1 ug/L,
     the hWER obtained from the WER determined using the site
     water with a hardness of 447 mg/L would be:

       hWER -  (2415 ug/L) (9 Cfs) * (1 Uff/L) (20 Cfs)  - 01 =A
              :     (9.2 Uff/L) (9 Cfs + 20 Cfs)  "    81-54.

     None of these calculations provides a way of  extrapolating
     a WER from one site-water hardness to another.  The only
     extrapolations that are possible are from one hardness of
     laboratory dilution water to another; the adjusted WERs
     are based on predicted toxicity in laboratory dilution
     water, but they are all based on measured toxicity in  site
     water.  If a WER is to apply to the design flow and the
     design hardness,  one or more toxicity tests have to be ,
     conducted using .samples of effluent and upstream water
     obtained under design-flow conditions and mixed at the .
     design-flow ratio to produce the design hardness.  A WER
     that is specifically appropriate to design conditions
     cannot be based on predicted toxicity in site water; it
     has to be based on measured toxicity in site water that  •
     corresponds to design-flow conditions.  The situation  is
     more complicated if the design hardness is not the
     hardness that is most likely to occur when effluent and
     upstream water are mixed at the ratio of ,the  design flows.
                              43

-------
B.' Background Information and Initial Decisions
 \
1.  Information should be obtained  concerning the effluent and
    the operating., and discharge  schedules of the discharger.

2.  The spatial extent of the site  to which the WER and the
    site-specific criterion are  intended to apply should be
    defined  (see Appendix A).  Information concerning
    tributaries, the plume, and  the point of complete mix
    should be obtained.  Dilution models  (U.S. EPA 1993d) and
    dye dispersion studies  (Kilpatrick 1992) might provide
    information that is useful for  defining sites for cmcWERs.

3.  If the Recalculation Procedure  (see Appendix B) is to be
    used, it should be performed.                       .

4.  Pertinent information .concerning the calculation of the
    permit limits should be obtained:
    a. What are the design flows, i.e., the flow of the
       upstream water (e.g., 7Q10)  and the flow of the
       effluent that are used in the calculation of the permit
       limits?  (The design flows for the CMC and CCC might be
       the same or different.)
    b. Is there a CMC (acute) mixing zone and/or a CCC
       (chronic) mixing zone?    -       '  '    .   ,
    e. What are the dilution(s) at  the edge(s) of the mixing
       zone(s)?
    d. If the criterion is hardness-dependent, what is the
       hardness on which the permit limits are based?  Is this
       a hardness that is likely to occur under design-flow
       conditions?                    -    •

5.  It should be decided whether to determine a cmcWER and/or
    a cccWER.                                '-..'•

6.  The water quality criteria document (see.Appendix E) that
    serves as the basis of the aquatic life criterion should
    be read to identify any chemical or .toxicological
    properties of the metal that are relevant.         .

7.  If the WER is being determined by or for a discharger,  dt
    will probably'be desirable to decide what is the smallest
    WER that is desired by the discharger (e.g., the smallest
    WER that would not require a reduction in the amount of
    metal discharged).  This "smallest desired WER" might be
    useful when deciding whether to determine a WISH.  If a WER
    is determined, this "smallest desired WER" might be useful
    when selecting the range of concentrations to be tested in
    the site water; ,

8.  Information should be read concerning health and safety
    considerations regarding collection and handling of

                             44

-------
       effluent and surface water samples and conducting toxicity
       tests (U.S. EPA 1993a; ASTM 1993a) .,  Information should
       also be read concerning safety and handling of the
       metallic salt .that will be used in t^e preparation, of-the
       stock solution.

   9.  The proposed work should be "discussed with the appropriate
       regulatory authority  (and possibly the Water Management
       Division of the EPA Regional Office) before deciding how
      .'to proceed with the development of a detailed workplan.

   1.0. Plans should be made to perform one or more range finding
       tests in both laboratory dilution water and site water
       (see section G,7).


C. Selecting Primary and Secondary Tests

   1.  For each WER  (cmcWER and/or cccWER)  to be determined, the
       primary and secondary tests should be selected using the
       rationale presented in section A. 3,  the information in
       Appendix I, the information in the criteria document for
       the metal  (see Appendix E), and any other pertinent
       information that is available.  When a specific test
       species is not specified,' also select the species.
       Because at least three WERs must be determined with the
       primary test, but only one must be determined with the
       secondary, test, selection of the tests might be influenced
       by the availability of the species  (and the life stage in
       some cases) during the planned testing period.
       a. The description of a "test" specifies not only the test
          species and the duration of the test,but also the life
          stage of the species and the adverse effect on which
          the results are to be based, ail of vwhich can have a
          major impact on the sensitivity of the test.
       b. The endpoint  (e.g., LC50, EC50, IC50)  of the primary
          test in laboratory dilution water should be as close as
          possible, but it must not be below, the CMC and/or CCC
          to which the WER is to be applied, because for any two
          tests, the test that has the lower endpoint is likely
          to give the higher WER  (see Appendix D) .   .      " ' ,
          NOTE: If both the Recalculation Procedure and a WER are
                to be used in the derivation of the site-specific
                criterion, the Recalculation Procedure must be
                completed first because the recalculated CMC
                and/or CCC must be used in the selection of the
                primary and secondary tests.                    •
       c. The endpoint  (e.g., LC50, EC50, IC50)  of the secondary
          test in laboratory dilution water should be as close as
          possible, but may be above or below, the CMC and/or GCC
          to which the WER is to be applied.


                •.'..••.      45

-------
   1} Because few toxicity tests have endpoints close to
      the CMC and CCC and because the major use of the
      secondary test is confirmation •(see section I.7.b),
      the endpoint of the secondary test may be below the
      CMC or CCC.  If the endpoint of the secondary test
      in laboratory dilution.--water is above the CMC and/or
      CCC, it might be possible to use the results to
      reduce the impact of experimental variation (see
      Figure 4).  If the endpoint of the primary test in
      laboratory dilution water is above the CMC and the
      endpoint of the secondary test is between the CMC
      and CCC, it should be possible to determine both a
      cccWER and a,cmcWER using the same two tests.
   2) It is often desirable to conduct the secondary test
      when the first primary test is conducted in case 'the
      results are surprising; conducting both tests the
      first time also makes it possible to interchange the
      primary and secondary tests, if desired, without
      increasing the number of tests that need to be
      conducted.   (If results of one or more, rangefinding
      tests are not available,.it might be desirable to
      wait and.conduct the secondary test when more
     * information is available concerning the laboratory
      dilution water and the site water.)

•The primary and secondary tests oust be conducted with
species in different taxonomic orders; at least one
species must be an animal and, when feasible, one species
should be a vertebrate and the other should be an
invertebrate.  A plant cannot be used if nutrients and/or
•chelators need to be added to either or both dilution
waters in order to determine the WER.  It is desirable to
use a test and species for which the rate of success is
known to be high and for which the test organisms are
readily available.   (If the WER is to be used with a
recalculated CMC and/or CCC, the species used in the
primary and secondary tests do not have to be on the list
of species that are used to obtain the recalculated CMC
arid/or CCC.)              .

There are advantages to using tests suggested in Appendix
I or other tests of comparable sensitivity for which data
are available from one or more other laboratories.
a. A good indication of the sensitivity of the test is
   available.  This helps ensure that the endpoint in
   laboratory dilution water is close to the;CMC and/or  .
   CCC and aids in the selection of concentrations of the
   metal to be used in the rangefinding and/or definitive
   toxicity tests, in laboratory dilution water.  Tests
   with other species such as species that occur at the
   site may be used, but it is sometimes more difficult  to
   obtain, hold, and test such species.

                         46      -

-------
       b. When a WER is determined and used, the results of-the
          tests in laboratory dilution water provide the
          connection between the data used in the derivation of
          the national! criterion and the data obtained in. site
          water, i.e'., the results in laboratory dilution water
          are a vital link in the derivation and use of a WER.
          It is, therefore, important to be able to judge the
          quality of the results in laboratory dilution water.
          Comparison of results with data from other laboratories
          evaluates all aspects of the test methodology
          simultaneously, but for the determination of WERs, the .
          most important aspect is the quality of;the laboratory
          dilution water because the dilution water is the most
          important difference between the two side-by-side tests
          from which the WER is calculated.  Thus,  two tests must
          be conducted for which data are available on the metal
          of concern in a laboratory dilution water from at least
          one other laboratory-  If both the primary and
          secondary tests are ones for which acceptable data are
          available from at least one other laboratory, these are
          the only two tests that have to be conducted.  If,
          however, the primary and/or secondary tests are ones
          for which no results are already available for the
          metal of concern from another laboratory, the first or
          second time a WER is determined at least two additional
          tests must be conducted in the laboratory dilution
          water in addition to the tests that are conducted for
          the determination of WERs (see sections F.5 and 1.5).
          1) For the determination of a WER, data are not
             required for a reference toxicant with either the
             primary test or the secondary test because the above
         "    requirement provides similar data for the metal fpr
             which the WER is actually being determined.
          2) See Section 1.5 concerning interpretation of the
             results of these tests before additional tests are
             conducted.               .            .


D. Acquiring and Acclimating Test Organisms

 ' 1.  The test organisms should be obtained, cultured, held, •.
       acclimated, fed, and handled as recommended by the U.S.
       EPA  (1993a,b,c) and/or by ASTM (1993a,b,c,d,e).  All test
       organisms must be acceptably acclimated to a laboratory
       dilution water that satisfies the requirements given in
       sections F.3 and F.4; an appropriate number of the
       organisms may be randomly or impartially removed from the
   ,    laboratory dilution water and placed in the site water
       when it becomes available in order to acclimate the
       organisms to the site water for a while just before the   :
       tests are begun.


                                47    "    .'••'

-------
   2.  The organisms used in a pair of side-by-side tests must be
       drawn from the same, population and tested under identical
       conditions.                         •                  .
                                                         •
                    !

E. Collecting and Handling Upstream- Water and Effluent

 .  1.  Upstream water will usually be mixed with effluent to
       prepare simulated downstream water.  Upstream water may
       also be used as a site water if a WER is to be determined
       using upstream water in addition to' or instead of
       determining a WER using downstream water. , The samples of
       upstream "water must be representative; they must not be
       unduly affected by recent runoff events (or other erosion
       or resuspension events) that cause higher levels of TSS
       than would normally be present, unless there is particular
       concern about such conditions.

 .  2.  The sample of effluent used in the determination of a WER
       must be representative; it must be collected during a
       period when the discharger is operating normally.
       Selection of the date and time of sampling of the effluent
       should take into account the discharge pattern of the
       discharger.  It might be appropriate to collect effluent
       samples during the middle of the week to allow for
       reestablishment of steady-state conditions after shutdowns
       for weekends and holidays; alternatively, if end-of-the-
       week slug discharges are routine, they should probably be
       evaluated.  As mentioned above, because the variability of
       the effluent might contribute substantially to the
       variability of the WERs, it might be desirable to obtain
       and store more than one sample of the effluent when WERs
       are to be determined in case an unusual WER is obtained
       with the first sample used.

   3.  When samples of site water and effluent are collected for
       the determination of the WERs with.the primary test, there
       must be at least three weeks between one sampling event
       and the next.  It is desirable to obtain samples in at
       least two different seasons and/or during times of
       probable differences in the characteristics of the site
       water and/or effluent.

   4.  Samples of upstream water and effluent must be collected,
       transported, handled, and stored as recommended by the
       U.S. EPA (1993a).  For example, samples of effluent should
       usually be composites, but grab samples are acceptable if
       the residence time of the effluent is sufficiently long.
       A sufficient volume should be obtained so that some can be
       stored for additional testing or analyse's if em unusual
       WER is obtained.  Samples must be stored at 0 to 4°C in
       the dark with no air space in the sample container.

                                48

-------
   5.  At the time of collection, the flow of both the upstream
     •  water and the effluent must be either measured or
       estimated by means of correlation with a nearby, U.S.G.S.
       gauge, the pH..of both upstream water and effluent must be
       measured, and' samples' of both upstream water and effluent
       should be filtered for measurement of dissolved metals.
       Hardness, TSS, TOC, and total recoverable and dissolved
       metal must be measured in both the effluent and the
       upstream water.  Any other water quality characteristics,
       such as total dissolved solids (IDS) and conductivity,
       that are monitored monthly or more often by the permittee
       and reported in the Discharge Monitoring Report must also
       be measured.  These and the other measurements provide
       information concerning the representativeness of the
       samples and the variability of the upstream water and
       effluent.       ;

   6.  "Chain of custody" procedures (U.S., EPA 1991b) should be
       used for all samples of site water and effluent,
       especially if the data might be involved in a legal
       proceeding.

   7.  Tests must be begun .within 36 hours after the collection
       of the samples of the effluent and/or the site water,
       except that tests may be begun more than 36 hours after
       the collection of the samples .if it would require an
       inordinate amount of resources to transport the samples to
       the laboratory and begin the tests within 36 hours.

   8.  If acute and/or chronic tests are to be conducted with
       daphnids and if the sample of the site water contains
       predators, the site water must be filtered through a 37-jim
       sieve or screen to remove predators.


F. Laboratory Dilution Water

   1.  The laboratory dilution water must satisfy the
       requirements given by U.S. EPA (1993a,b,c) or ASTM
       (1993a,b,c,d,e).  The laboratory dilution water must be a
       ground water, surface water, reconstituted water, diluted
       mineral water, or dechlorinated tap water that has been
       demonstrated to be acceptable to aquatic .organisms.  If a;
       surface water is used for acute or chronic tests with
       daphnids and if predators are observed in the sample of
       the water, it must be filtered through a 37-pn sieve or
       screen to remove the predators.  Water prepared by such  •
       treatments as deionization and reverse osmosis must not be
       used as the laboratory dilution water unless salts,
       mineral water, hypersaline brine, or sea salts are added
       as recommended by U.S. EPA  (1993a) or ASTM  (1993a).

                                         /
    •.-..•'         .• .             49         •    . '   .     ••''''

-------
    2.. The concentrations of both TOG and TSS must be less than 5
        mg/L.

    3.  The hardness pf the laboratory dilution water should be
        between 50 arid 150 mg/L and must be between 40 and 220
        mg/L. .If the criterion for .the metal is hardness-
        dependent, the hardness of the laboratory dilution water
_  •      must not be above the hardness of the site water,  unless
       ' the hardness of the site water is below 50 mg/L.

    4.  The.alkalinity and pH of the laboratory dilution water
        must be appropriate for its hardness; values for
        alkalinity and pH that are appropriate for some hardnesses
        are given by U.S. EPA (1993a)  and ASTM (1993a); other
        corresponding values should be determined by
        interpolation.  Alkalinity should be adjusted using sodium
        bicarbonate,  and pH should be.adjusted .using aeration,
        sodium hydroxide, and/or sulfuric acid.

    5.  It would seem reasonable that,  before any samples  of site
        water, or effluent are collected,  the toxicity tests  that
        are to be conducted in the laboratory dilution  water for
        comparison with results of the same  tests from, other .
        laboratories  (see sections C.3.b  and 1.5)  should be
        conducted. These should be performed at  the hardness,
        alkalinity, and pH specified in sections  F.3 and F.4.
                                            i

G.  Conducting Tests

    1.   There  must be no'differences between the  side-by-side
        tests  other than  the  composition  of  the dilution water,
        the concentrations of metal  tested,  and possibly the water
        in which the  test organisms  are acclimated just prior to
        the beginning of  the  tests.

    2.  More than one test using site water may be conducted side-
       by-side with  a test using laboratory  dilution water;  the
       one test  in laboratory dilution water will be used in the
       calculation of several WERs, which means  that it is  very
       important  that that one  test be acceptable.            ,,

   3.  Facilities  for conducting toxicity tests  should be set up
       and test  chambers  should be  selected  and  cleaned as
       recommended by the U.S.  EPA  (1993a,b,c) and/or ASTM
        (1993a,b,c,d,e).                                          .

   4.  A stock solution  should  be prepared using an inorganic
       salt that is highly soluble in water.
       a. The salt does not have to be one that was used in tests
          that were used  in the derivation of  the national
          criterion.  Nitrate salts are generally acceptable;

                                50

-------
   chloride and sulfate salts of many metals are also
  "acceptable  (see Appendix J).  It is usually desirable
   to avoid use of a hygroscopic salt.  The salt used
   -should meet A.C.S. specifications for reagent-grade, if
   such specifications are available; use of a better
   grade is usually not worth the extra cost.  No salt
   should be used until information concerning safety and
   handling has been read. ,
b. The stock solution may be acidified (using metal-free
   nitric acid) only as necessary to get the metal into
   solution.           .          .  ' .
c. The same stock solution must be used to add metal to
   all tests conducted at one time.

For tests suggested in Appendix I,  the appendix presents
the recommended duration and whether the static, or renewal
technique should be used; additional information is
available in the references cited in the appendix.
Regardless of whether or not or how often test solutions
are renewed when these tests are conducted for other
purposes, the following guidance applies to all tests that
are conducted for the determination of WERs:
a. The renewal technique must be used for tests that last
   longer than 48 hr.
b. If the concentration of dissolved metal decreases by
   more than 50 % .in 48 hours in static or renewal tests,
   the test solutions must be renewed every 24 hours.
   Similarly, if the concentration of dissolved oxygen
   becomes too low, the test solutions must be renewed
   every 24 hours.  If one test in a pair of tests is a
   renewal test, both tests must be renewal tests.
c. When test solutions are to be renewed, the new test
•''  solutions must be prepared from the original unspiked
   effluent and water samples that have been stored at 0
   to 4°C in the dark with no air space in the sample
   container.
d. The static technique may be used for tests that do not
   last longer than .48 hours unless the above
   specifications require use of the renewal technique.
If a test is used that is not suggested in Appendix I, the
duration and technique recommended for a comparable teat
should be used.             •

Recommendations concerning temperature, loading, feeding,
dissolved oxygen, aeration, disturbance,  and controls
given by the U.S. EPA (1993a,b,c) and/6r ASTM
(1993a,b,c,d,e) must be followed. . The procedures that are
used must be used in both of the side-by-side tests.

To aid in the selection of the concentrations of metals   „
that should be used in the test solutions in site water, a
static rangefinding test should be conducted for 8 to 96

        '"•' '  •           51      '   • ."•             •    •   •

-------
8
 hours,  using a dilution factor of 10  (or 0.1)  or 3.2  (or
 0.32)  increasing from about a factor  of  10  below the  value
 of the eridpoint given in the criteria document for the
 metal  or in Appendix I of this document  for tests with  "
 newly  hatched' fathead minnows.  If the test is not in the
 criteria document and no other data are  available,  a  mean
 acute  value or other data for a taxonomically  similar
 species should be used as the predicted  value.   This
 rangefinding test will provide information  concerning the
 concentrations that  should be used to bracket  the endpoint
 in the definitive test and will provide  information
 concerning whether the control survival  will be
 acceptable.   If dissolved metal is measured in one or more
 treatments at the beginning and end of the  rangefinding
 test,  these data will indicate whether the  concentration
 should be expected to decrease by more than 50  %  during
 the definitive test.   The rangefinding test may be
 conducted in either  of two ways:  \     •
 a.  It may be conducted using the samples of effluent  and
    site water that will be used in the definitive test.
    In  this case,  the duration of the  rangefinding test
    should be as long as possible within  the limitation
    that the definitive test, must begin within  36  hours
    after the samples  of effluent and/or  site water were
   •collected,  except  as per section E.7.
 b.  It may be conducted using one set  of  samples of
    effluent  and upstream water with the  definitive  tests
    being conducted using samples obtained at a  later  date.
    In this case the  rangefinding test might give  better
    results because it can last longer, but  there  is the
    possibility that  the quality of the effluent and/or
    site water might  change.   Chemical analyses  for
  *  hardness  and pH might indicate whether any major
    changes occurred  from one sample to the  next.
 Rangefinding tests are especially desirable, before  the
 first set of, toxicity tests.   It might be desirable to
 conduct rangefinding  tests before each individual
 determination of a WER to obtain additional information
 concerning the effluent,  dilution water,  organisms, etc.,
 before  each set of side-by-side tests are begun.
                                                       *
 Several considerations are important  in  the selection of
 the dilution factor  for definitive tests.   Use  of
 concentrations that are close together will reduce  the
 uncertainty  in the WER but will require  more
 concentrations to cover a range within which the  endpoints
might occur.   Because of the resources necessary  to
 determine a  WER,  it  is important that endpoints in both
 dilution waters be obtained whenever  a set  of  side-by-side
 tests are conducted..  Because static  and renewal  tests can
be used to determine  WERs,  it is relatively e,asy  to use
more treatments than  would be used in flow-through tests.

                         52

-------
    The dilution factor for total recoverable metal must be
    between 0.65 and 0.99, and the recommended factor is 0.7.
    Although-factors between 0.75 and 0.99 may be used, their
    use will probably not be cost-effective.  'Because there is
    likely to be more uncertainty in the predicted value of
    the endpoint in site water, 6 or 7 concentrations are
    recommended in the laboratory-dilution water, and 8 or 9
    in the simulated downstream water, at a dilution factor of
    0.7.  It might be desirable to use even more treatments in
    the first of the WER determinations,, because the design of
    subsequent tests can be based on the results of the first
    tests if the site water, laboratory dilution water, and
    test organisms do not change too much.  The cost of adding
    treatments can be minimized if the concentration of metal
    is measured only in samples from treatments that will be
    used in the calculation of the endpoint.

9.  Each test must contain a dilution-water control.  The
    number of test organisms intended to be exposed to each
    treatment, including the controls, must be at least 20.
    It is desirable that the organisms be distributed between
    two or more test chambers per treatment.  If test
    organisms are not randomly assigned to the test chambers,
    they must be assigned impartially (U.S. EPA 1993a; ASTM
    1993a) between all test chambers for ;a pair of side-by-
    side tests.  For example, it is not acceptable to assign
    20 organisms to one treatment,  and then assign 20
    organisms to another treatment, etc.  Similarly, it is not
    acceptable to assign all the organisms to the test using
    one of the dilution waters and then assign organisms to
    the test using the other dilution water.  The test
    chambers should be assigned to location in a totally
    random arrangement or in a randomized block design.

10. For the test using site water,  one of the following
    procedures should be used to prepare the test solutions
    for the test chambers and the "chemistry controls" (see
    section H.I):                                ,
    a. Thoroughly mix the sample of the effluent and place the
       same known volume of the effluent in each test chamber;
       add the necessary amount of metal, which will be    •
       different for each treatment; mix thoroughly; let stand
       for 2 to 4 hours; add the necessary amount of upstream
       water to each test chamber;  mix thoroughly; let stand
       for 1 to 3 hours.
    b. Add the necessary amount of metal to a large sample of
       the effluent and also maintain an unspiked sample of •
       the effluent; perform serial dilution using a graduated
       cylinder and the well-mixed spiked and .unspiked samples
       of the effluent; let stand for 2 to 4 hours; add the
       necessary amount of upstream water to each test
       chamber; mix thoroughly; let stand for 1 to 3 hours.

                             53         .''.'•        '  '

-------
    c. Prepare a large volume of  simulated downstream water by
       mixing effluent and upstream water in the desired
       ratio; place the same known volume of the simulated
       downstream, water in each test chamber; -add the .
       necessary "amount of metal, which will be different for
       each treatment; mix thoroughly and let stand for 1 to 3
       hours.
    d. Prepare a large volume of  simulated downstream water by
       mixing effluent and upstream water in the desired
       ratio,; divide it into two  portions; prepare a large
       volume of the highest test concentration of metal using
       one portion of the simulated downstream water; perform
       serial dilution using a graduated cylinder and the
       well-mixed,spiked and unspiked samples of the simulated
       downstream water; let stand for 1 to 3 hours.
    Procedures "a" and "b" allow  the metal to equilibrate
    somewhat with the effluent before the solution is diluted
    with upstream water.

11. For the test using the laboratory dilution water, either
    of the following procedures may be used to prepare the
    test solutions for the test chambers and the '"chemistry
    controls* (see section H.I) :
    a. Place the same known volume of the laboratory dilution
       water in each test chamber; add the necessary amount of
       metal, which will be different for each treatment; mix
       thoroughly; let Stand for  1 to 3 hours.
    b. Prepare a large volume of  the highest test
       concentration in the laboratory dilution water; perform
       serial dilution using a graduated cylinder and the
       well-mixed spiked and unspiked samples of the
       laboratory dilution water; let stand for 1 to 3 hours.

12. The test organisms, which have been acclimated as per
    section D.I, must be added to the test chambers for the
    site-by-side tests at .the same time.  The time at which
    the test organisms are placed in the test chambers is
    defined as the beginning of the tests, which must be
    within 36 hours of the collection of the samples, except
    as per section E.7.       •
                                  ) '•            •          •'  •
13. Observe the test organisms and. record the effects and
    symptoms as specified by the  U.S. EPA (1993a,b,c) and/or
    ASTM (1993a,b,c,d,e).  Especially note whether the.
    effects, symptoms, and time course of toxicity are the
    same in the side-by-side tests.

14. Whenever solutions are renewed, sufficient solution should
    be prepared to.allow for chemical analyses.
                             54

-------
H.  Chemical  and  Other Measurements

   •1.  To reduce the  possibility  of  contamination  of  test
       solutions befpre or  during tests,  thermometers and. probes
       for measuring  pH and dissolved oxygen must  not be placed
       in test chambers that will provide data  concerning effects
       on test organisms or data  concerning the concentration of
       the metal.  Thus measurements of pH, dissolved oxygen/ and
       temperature before or during  a test must be performed
       either on "chemistry controls" that, contain test organisms
       and-are fed the same as the other  test chambers or on
       aliquots  that  are removed  from the test  chambers.  The
       other measurements may be  performed on the  actual test
       solutions at the beginning and/or  end of the test or the
       renewal.

   2.  Hardness  (in fresh water)  or  salinity (in salt water), pH,
       alkalinity, TSS, and TOC must be measured on the upstream
       water, the effluent, the simulated and/or actual
       downstream water, and the  laboratory dilution water<
       Measurement of conductivity and/or total .dissolved solid.s
       (TDS) is  recommended in fresh water.

   3.  Dissolved oxygen, pH, and  temperature mast be measured
       during the test at the times  specified by the U.S. EPA
       (1993a,b,c) and/or ASTM (1993a,b,c,d,e).  The measurements
       must  be performed on the same schedule for both of the
       side-by-side tests.  Measurements  must be performed on
      . both  the .chemistry controls and actual test solutions at
       the end of the test.

   4.  Both  total recoverable and dissolved metal must be
       measured  in the. upstream water, the effluent, and
       appropriate test solutions for each of the tests.
•  ,     a. The analytical measurements should be  sufficiently
          sensitive and precise that variability in analyses will
          not greatly increase the variability  of the WERs.  If
          the detection limit of  the analytical  method that will
          be used to  determine the metal  is greater than one-
          tenth  of the CCC  or. CMC that is to be  adjusted, the
          analytical  method should probably be  improved or  .  •
          replaced (see Appendix  C).  If  additional sensitivity
          is needed,   it is .often useful to separate the metal  •  .
          from the matrix because this will simultaneously
          concentrate the metal and  remove interferences.
          Replicate analyses should  be performed if necessary to
          reduce the  impact of analytical variability.
          1) EPA methods (U.S.. EPA 1983b,1991c)  should usually/be
             used for both  total  recoverable and dissolved
             measurements,  but in some cases alternate methods
             might have to be used in order to  achieve the
             necessary sensitivity.  Approval for use of

                               '.55  ••  '    •'         '

-------
   alternate methods is to be requested from the
   appropriate regulatory authority.
All measurements of metals must be performed using
appropriate QA/QC,techniques.  Clean techniques, for
obtaining/ handling, storing, preparing, and analyzing
the samples should be used when necessary to achieve
blanks that are sufficiently low  (see Appendix C).
Rather than measuring the metal in all test solutions,
it is often possible to store samples and then analyze
only those that are needed to calculate the results of
the toxicity tests.  For dichotomous data (e.g.,
either-or data; data concerning survival), the metal in
the following must be measured:
1) all concentrations in which some, but not all, of
   the test organisms were adversely affected.
2) the highest concentration that did not adversely
  . affect any test organisms.
3) the lowest concentration that adversely affected all
   of the test organisms,
4) the controls.
For data that are not dichotomous (i.e., for count and
continuous data)', the metal in the controls and in the
treatments that define the concentration-effect curve
must be measured; measurement of the concentrations of
metals in other treatments is desirable.
In each treatment in which the concentration of metal
is to be measured, both the total recoverable and
dissolved concentrations must be measured;
1) Samples must be taken for measurement of total
   recoverable metal once for a static test, and once
   for each renewal for renewal tests; in renewal  ,
   tests, the samples are to be taken after the
   organisms have been transferred to the new test
   solutions.  When total recoverable metal is measured
   in a test chamber, the whole solution in the chamber
   must be mixed before the sample is taken for
   analysis; the solution in the  test chamber must not
   be acidified before the sample is taken.  The sample
   must be acidified after it is  placed in the sample
   container.
2) Dissolved metal must be measured at the beginning-
   and end of each static test; in a renewal test, the
   dissolved metal must, be measured at the beginning of
   the test and just before the solution is renewed the
   first time.  When dissolved metal is measured in a
   test chamber,  the whole solution in the test chamber
   must be mixed before a sufficient amount is removed
   for filtration; the solution in the test chamber
   must not be acidified before the sample is taken.
   The sample must be filtered within one hour after  it
   is taken, and  the filtrate must be acidified  after
   filtration.

                      56

-------
   5.  Replicates, matrix spikes, and other. QA/QC checks must be
       performed as required by the U.S. EPA (1983a,1991c).


I. Calculating and Interpreting the Results

   1.  To prevent roundoff error in' subsequent calculations,  at
       least four significant digits must be retained in all
       endpoints, WERs,  and FWERs.  This requirement is not based
       on mathematics or statistics and does not reflect the
       precision of the value; its purpose is to minimize concern
       about the effects of rounding off on a site-specific
       criterion.  All of these numbers are intermediate values
       in the calculation of permit limits and should not be
       rounded off as if they were values of-ultimate concern.

   2.  Evaluate the acceptability of each toxicity test
       individually.                                    .
       a. If the procedures used deviated from those specified
    .  :    above, particularly in terms of acclimation,
          randomization, temperature control, measurement of
          metal, and/or disease or disease-treatment, the test
          should be rejected; if deviations were numerous and/or
          substantial, the test must be rejected.
       b. Most tests are unacceptable if more than 10 percent of
          the organisms in the controls were adversely affected,
          but the limit is higher for some tests; for the tests
          recommended in Appendix I, the references given should
          be consulted.
       c. If an LC50 or EC50 is to be calculated:
          1) The percent of the organisms that were adversely
             affected must have been less than 50 percent,  and
       .      should have been less' than 37 percent, in at least
             one treatment other than the. control.
          2) In laboratory dilution water the percent of the
             organisms that were adversely affected must have
             been greater than 50 percent, and should have been
             greater than 63 percent, in at least one treatment.
             In site water the percent of the organisms that were
             adversely affected should have been greater than 63
             percent in at least one treatment.  (The LC50 or •••
             EC50 may be a "greater than" or "less than" value in
             site water, but not in laboratory dilution water.)
          3) If there was an inversion in the data  (il.e., if a
             lower concentration killed or affected a greater
             percentage of the organisms than a higher
             concentration), it must not have involved more than
             two concentrations that killed or affected between
             20 and 80 percent of the test organisms.
          If an endpoint other than an LC50 or EC50 is used or  if
          Abbott's formula is used, the above requirements will
          have to be modified accordingly.

                . '          ' •    57

-------
 if--,
    d. Determine whether  there was anything unusual about the
       test results that  would make them questionable.
    e. If solutions were  not renewed.every 24 hours, the
       concentration of dissolved metal must not have .
       decreased by more.than 50 percent from the beginning to
       the end of a static test or from the beginning to the
       end of a renewal in a renewal test in test
       concentrations that were used in the calculation of the
       results of the test.

3.  Determine whether the effects, symptoms, and time course
    of toxicity was the same in the side-by-side tests in the
    site water and .the laboratory dilution water.  For
    example, did mortality occur in one acute test, but
    immobilization in. the other?  Did most deaths occur before
    24 hours in one test, but after 24 hours in the other?  in
    sublethal tests, was  the most sensitive effect the same in
    both tests?  If the effects, symptoms,  and/or time course
    of toxicity were different, it might indicate that the
    test is questionable  or that additivity, synergism,  or
    antagonism occurred in site water.  Such information might
    be particularly useful when comparing tests that produced
    unusually low or high WERs with tests that produced
   .moderate WERs.                          "

4.   Calculate the results of each test:
    a. If the data for the most sensitive effect are
       dichotomous, the endpoint must be calculated as.a LC50,
       EC50,  LC25, EC25,  etc.,  using methods described by the
       U.S.  EPA (1993a)  or ASTM (1993a) .   If two or more
       treatments affected between 0 and 100 percent in .both
       tests in a side-by-side pair,  prbbit analysis must be
      • used to calculate results of both tests,  unless the
      ' probit model is rejected by the goodness of fit test in
       one or both of the acute tests.  If  probit analysis
       cannot be used,  either because fewer than two
       percentages are between 0 and 100  percent or because
       the model does not fit the data, computational
       interpolation must be used (see Figure 5); graphical
       interpolation must not be used.
       1)  The same endpoint (LC50,  EC25,  etc.)  and the same-
          computational . method must be .used for both tests
          used .in the calculation of a WER.
       2)  The selection of the percentage used to define the
          endpoint might  be influenced by the percent effect
          that occurred in the tests and  the correspondence
          with the CCC and/or CMC..
       3)  If  no treatment, killed or affected more than 50
          percent of the  test organisms and the test was
          otherwise acceptable,  the LC50  or EC50 should be
          reported to be  greater than the highest test
          concentration.

                            58   .  "    '   .

-------
   4) If no treatment other than the control killed or
      affected less than 50 percent of the test organisms
      and the test was otherwise acceptable, the LC50 or
      EC50 should be reported to be less than the lowest
      test concentration.                        '
b. If the data for the most .sensitive effect are not
   dichotomous, the endpoint must be calculated using a
   regression-type method (Hoekstra and Van Ewijk 1993;
   Stephari and Rogers 1985)/ such as linear interpolation
   (U.S. EPA 1993b,c) or a nonlinear regression method
   (Barnthouse et al.' 1987;  Suter et al. 1987; Bruce and
   Versteeg 1992).  The selection of the percentage used
   to define the endpoint might be influenced by the
   percent effect that occurred in the tests and the
   correspondence with the CCC and/or CMC.  The endpoints
   in the side-by-side tests must be based on the same
   amount of the same adverse effect so that the WER is a
   ratio of identical endpoints.  The same computational
   method must be used for bo.th tests used in the
   calculation of the WER.
c. Both total recoverable and dissolved results should be
   calculated for each test.
d. Results should be based on the time-weighted average
   measured metal concentrations (see Figure 6).

The acceptability of the laboratory dilution water must be
evaluated by comparing results obtained with two sensitive
tests using the laboratory dilution water with results
that were obtained using a comparable laboratory dilution
water in. one or more other laboratories  (see sections
C.S.b and F.5).
a. If, after taking into account any known effect of
   hardness on tpxicity, the new values  for the endpoints
   of both of the tests are  (1) more than a factor of 1.5
   higher than the respective means of the values from the
   other laboratories or  (2) more than a factor of 1.5 ,
   lower than the respective means of values from the
   other laboratories or  (3) lower than  the respective
   lowest values available  from other, laboratories or  (4)
   higher than the respective highest values available
   from other laboratories, the new and  old data must be
   carefully evaluated to determine whether the laboratory
   dilution water used in the WER determination was
   acceptable.  For  example, there might have been an
   error in the  chemical measurements, which might mean
   that the results  of all  tests performed  in the WER
   determination need to be adjusted  and that the WER
   would not change.  It  is also possible'that  the metal
   is more or  less toxic  in the laboratory  dilution  water
   used in the WER determination.  Further,  if  the new
   data were based on measured concentrations but  the  old
   data were based on nominal  concentrations,  the  new data

                         59

-------
        should probably be considered to be better than the
        old.   Evaluation of results of"any other toxicity tests
        on the same or a different metal using the same
        laboratory,dilution water might be useful.
     b.  If, after taking, into account any known effect of
        hardness on toxicity,  the.--new values for the endpoints
        of -the two tests are not either both higher or both
        lower in comparison than data from other laboratories
        (as per section a above)  and .if both of the new values
        are within a factor of 2 of the respective means of the
        previously available values or are within the ranges of
        the values,  the laboratory dilution water used in the
        WER determination is acceptable.
     c.  A control chart approach may be used if sufficient data
        are available.                             .
     d.  If the comparisons do not indicate that the laboratory
        dilution water,  test method,  etc.,  are acceptable,  the
        tests probably should be considered unacceptable,
        unless -other toxicity data  are available to indicate
        that  they are acceptable
     Comparison of results of tests between laboratories
     provides a check on all aspects of, the test procedure;,, the
    .emphasis here is on the quality of the laboratory dilution
     water because all  other aspects of the side-by-side tests
     on  which the WER is based must be the same,  except
     possibly for the concentrations of metal  used and the
     acclimation just prior to the  beginning of the tests.

6.   If  all the necessary tests  and the laboratory dilution
    water are .acceptable,  a WER must  be.calculated by dividing
     the endpoint obtained using  site  water by the endpoint
     obtained using  laboratory dilution water.
     a.  If both a primary test and  a secondary test were
        conducted using both waters, WERs must be calculated
        for both tests.
    b.  Both  total recoverable and  dissolved WERs must be
        calculated.
    c.  If jthe detection limit of the  analytical  method used to
       measure  the metal  is above  the endpoint  in laboratory
        dilution water,  the detection  limit must  be used as  the
        endpoint,  which will result  in a lower WER than would
       be obtained  if  the actual concentration had been
       measured.  If the  detection limit of the  analytical
       method used  is  above the  endpoint in site water, a WER
        cannot be determined.          ..

7.  Investigation of the  WER.                      ,       •  •
    a. The results of  the chemical measurements  of hardness,
       alkalinity, pH,  TSS, TOC, total recoverable metal,
       dissolved metal, etc., on the  effluent  and the upstream
       water  should  be  examined  and compared  with previously
       available values for the  effluent and  upstream water,

                             60

-------
.respectively,  to determine whether the samples were
 representative'and to get some indication of the
 variability in the composition,  especially as it,might
 affect the toxicity of the metal and the WER, and  to
 see if the WER correlates with one or more of the
.measurements.              -'
 The WERs obtained with the primary and secondary tests
 should be compared to determine whether the WER
 obtained with  the secondary test confirmed the WER
 obtained with  the primary test.   Equally sensitive
 tests are expected to ,give WERs that are similar (e.g.,
 within a factor of 3),  whereas a test that is less
 sensitive will probably give a smaller WER than a  more
 sensitive test (see Appendix D)-.   Thus a WER obtained
 with a primary test is considered confirmed if either
 or both of the following are true:
 1)  the WERs obtained with the primary and secondary
    tests are within a factor of 3.
 2)  the test, regardless of whether it is the primary or
    secondary test,  that gives a higher endpoint in the
    laboratory  dilution water also gives the larger WER.
 If the WER obtained with the secondary test does not
 confirm the WER obtained with the primary test,  the
 results should be investigated.   In addition,  WERs
 probably should be determined using both tests the next
 time samples are obtained and it would be desirable to
 determine a WER using a third test.   It is also
 important to evaluate what the results imply about the
 protectiveness,of any proposed site-specific criterion.
 If the WER is  larger than 5,  'it  should be investigated.
 1)  If the endpoint obtained using the laboratory
    dilution water was lower than previously reported  .
    lowest value or was more than,a factor of two lower
    than an existing Species Mean Acute Value in a '
    criteria document,  additional.-tests dn the
    laboratory  dilution water are probably desirable.
 2)  If a total  recoverable WER was larger than 5 but the
    dissolved WER was not,  is the metal one whose WER is
    likely to be affected by >TSS  and/qr TOC and was the
    concentration of TSS and/or TQC high?  Was there a
    substantial difference between the total recoverable
    and dissolved concentrations  of the metal in.the
    downstream  water?
 3)  If both the total recoverable and dissolved WERs
    were larger than 5,  is it likely that there is
    nontoxic dissolved metal in the downstream water?
 The adverse effects and the time-course of effects in
 the side-by-side tests should be compared.   If they are
 different,  it  might indicate that the site-water test
 is' questionable or that additivity,  syhergism,  or
 antagonism occurred in the site  water.   This might be
 especially important if the WER obtained with the  .

 '.-'••:      '•          61             • ... '..  „• ' ' •

-------
 8.
           secondary, test did not confirm the WER obtained with
           the primary test or if the WER was very large or small.

        If at least one WER determined with the primary test was
        confirmed by a WER that was simultaneously determined with
        the secondary test,  the cmcFWER and/ or the cccFWER should
        be derived as described in section A.5.  .

    9.   All data generated during the determination of the WER
        should be examined to see if there are any implications
        for the national or site-specific aquatic life criterion
        a.  If there are data for a species for which data were not
           previously available or unusual data for a species for
           which data were/available,  the national criterion might
           need to be revised.
      ,  b.  If the primary test  gives an LC50  or EC50  in laboratory
           dilution water that -is the same as the' national .CMC
           the resulting site-specific CMC should be  similar to
         .  the LC50 that was  obtained with the primary test  using
           downstream water.  Such relationships  might  serve as a
           check on the applicability of  the  use  of WERs /
        c.  If data indicate that  the site-specific criterion would
           not adequately protect  a critical  species, the  site-
           specific criterion probably should be  lowered.
                                      ,-                        i,

J. Reporting the Results-

   A report  of the experimental determinatior* of a WER to the
   appropriate .regulatory authority must include the following :
   1.  Name(s) of the investigator (s) , name and location of the
       laboratory, and dates of initiation and termination of the
2.  A description  of  the  laboratory dilution water,  including
    source, preparation,  and any demonstrations that an
    aquatic species can survive,  grow, and reproduce in it
3.  The name, location, and  description of the discharger, a
    description of the effluent,  and the design flows of the
    effluent and the  upstream water.
4.  A description  of  each sampling  station,  date, and time,
    with an explanation of why they were selected, and the
    flows of the upstream water  and the effluent at  the time
    the samples were  collected.
5.  The procedures used to obtain,  transport, and store the
    samples of the upstream water and the effluent .
6,  Any pretreatment, such as filtration, of the effluent,
    site water, and/or laboratory dilution water.
7.  Results of all chemical and physical measurements on
    upstream water, effluent, actual and/or simulated
    downstream water, and laboratory dilution water, including
    hardness (or salinity) ,. alkalinity, pH,  and concentrations
    of total recoverable metal, -dissolved metal,  TSS, and TOC
                             62

-------
 8.  Description of  the  experimental  design,  test  chambers,
    depth  and..volume  of solution in  the  chambers,  loading and
    lighting, and numbers  of  organisms and chambers  per
    treatment.    ••--               •:                   .
 9.  Source and  grade  of the metallic salt,  and how the  stock
    solution was prepared, including any acids or bases used.
 10. Source, of'the test  organisms, -scientific name and how .
    verified, age,  life stage, means and ranges of weights
    and/or lengths, observed  diseases, treatments, holding  and
    acclimation procedures, and  food.
 11. The average and range  of  the temperature,  pH;  hardness  (or
    salinity),  and  the  concentration of  dissolved oxygen (as %
    saturation.and  as mg/L) during acclimation, and  the method
    used to measure them.
 12. The following must  be  presented  for  each toxicity test:
    a. The average  and  range  of  the  measured concentrations  of
       dissolved oxygen, as % saturation and as mg/L.
    b. The average  and  range  of  the  test temperature and the ,
       method used  to measure it.
    c. The schedule for taking samples of test solutions and
       the methods  used to obtain, prepare,  and store them.
    d. A summary table  of  the total  recoverable and  dissolved
       concentrations of the  metal in each treatment,,
       including all  controls, in.which  they were measured.
    e. A summary table  of  the values of  the  tpxicoldgical
       variable(s)  for  each,treatment, including all controls,
       in  sufficient  detail to allow an  independent
       statistical  analysis of the data.
    f. The endpoint and the method used.to calculate it.
    g. Comparisons -with other data obtained  by conducting the
    ' - . same test on the same  metal using laboratory  dilution
       water in the same and  different laboratories;  such data
       may be from a  criteria document or from another  source.
    h. Anything unusual about the test,  any  deviations  from
     .  the procedures described  above >• and-any other relevant
       information.
13. All differences,  other than  the  dilution water arid  the
    concentrations of metal in the test  solutions,, between the
    side-by-side tests  using  laboratory  dilution water  and  -
    site water.   , . ,
14. Comparison  of results  obtained with  the  primary  and
    secondary tests.
15,' The WER and an explanation of  its calculation.        -
A report of the derivation of a FWER must include the
following:        ,                 ^
1.  A report-of the determination of each WER that was,
    determined for the derivation of the EWER; all WERs
  •  determined with secondary tests must be reported along
    with all WERs that were determined with the primary test,

     .••..•'   ' •      " . ."  :•'  63.' .  •  '- .   .   .  .  - •    "

-------
The  design flow of  the  upstream water  and  the  effluent and
the  hardness used in  the  derivation  of the permit  limits
if the  criterion for  the  metal  is hardness-dependent.
A summary  table must ,be presented that contains the
following  for each  WER  that was derived:
a. the  value of the WER and the two  endpoints  from which
   it was  calculated.         .                     ,
b. the  hWER calculated  from.the WER.
c. the  test and species that was used.
d. the  date the samples of effluent  and site water were
   collected.
e. the  flows of the effluent and upstream .water when the
   samples were taken.
f. the  following information concerning the laboratory
   dilution water, effluent, upstream water,  and actual
   and/or simulated downstream water: hardness (salinity),
   alkalinity,  pH, and concentrations of total recoverable
   metal, dissolved metal, TSS,  and TOC.
A detailed explanation of how the FWER was derived from
the WERs that are in the summary table.
                        64

-------
 METHOD 2: DETERMINING CCCWERS FOR AREAS AWAY FROM PLUMES

                              ;,      *.>'••'"..       '

 Method,2 might be viewed as a simple process wherein samples of
-site water are obtained from locations within a large body of
 fresh or salt water (e.g.,  an ocean or a large lake,  reservoir,
 or estuary),  a WER is determined for each sample,  and the FWER is
 calculated as the geometric mean of some or all of the WERs.  In
 reality.  Method 2 is not likely to .produce useful  results unless
 substantial resources are devoted to planning and  conducting the
 study.   Most sites to which Method 2 is applied will have long
 retention times,  complex mixing patterns,  and a number of
 dischargers.   Because metals are persistent,  the long retention
 times mean that the sites are likely to be defined to cover
 rather  large areas;  thus such sites will herein be referred to
 generically as "large site's,".  Despite the differences between
 them, all large sites require similar special considerations
 regarding the determination of WERs.  Because Method 2 is based
 on samples Of actual surface water (rather than simulated surface
 water),  no sample should; be taken in the vicinity  of  a plume and
 the method should be used to determine cccWERs,  not cmcWERs.  If
 WERs are to be determined for more than one metal,. Appendix F
 should  be read.                      .

 Method  2  uses many of the same methodologies as Method 1,  such as
 those for toxicity tests and chemical analyses.  Because the
 sampling plan is  crucial to Method 2 and the plan  has "to be based
 on site-specific  considerations,  this description  'of  Method 2
 will be more  qualitative than the description of Method l.:

 Method  2  is based on use of actual surface water samples,  but use
 of simulated  surface water  might provide information  that is
 useful  for some purposes:
 1.  It might be desirable to compare the WERs for two  discharges
    that  contain the  same metal.   .This might, be accomplished by
   .selecting  an appropriate dilution water and preparing two
  . simulated  surface waters,  one that contains a known
    concentration  of  one effluent and one that contains a known
    concentration  of  the other effluent.   The relative magnitude
    of the two WERs is'likely to,be more useful than the absolute
    values of  the  WERs themselves.
 2.  It might be desirable to determine whether the  eWER for a
    particular effluent is additive with the WER of the .site water
    (see Appendix  G).   This  can be studied by determining WERs for
    several  different known  concentrations  of  the effluent in site
    water.   •/_-;._    .   -''•"•,''-.'•.   -      •    -'-
 3.  An event such  as  a rain  might affect the WER because of a
    change in  the  water quality,  but it .might  also  reduce the WER
    just by  dilution  of refractory metal or TSS.  A proportional
    decrease in the WER and  in the concentration of the metal
    (such-as by dilution of  refractory metal)  will  not result in
    underprotection;  if,  however,  dilution decreases the WER

       •'   ,  "   ,-       .'  '"'    •  .-65         "'.''••   v  /  .  '

-------
    proportionally more than it decreases the concentration of
    metal in the downstream water,  underprotection is likely to
    occur.  This is essentially a determination of whether the WER
    is additive when the effluent is diluted with rain water (see
    Appendix- G) .           ,         ..,•'"
 4.  An event that increases TSS might increase the total
    recoverable concentration of the metal and the total
    recoverable WER without having much effect on either  the
    dissolved concentration or the dissolved WER.
 In  all four cases,  the use of simulated surface water is useful
 because it allows for the determination of WERs using known
 concentrations  'of effluent.

 An  important step in the determination of any WER is to  define
 the area to be  included in the site.   The major principle that
 should be applied when defining the area is the samp for all
 sites:  The site should be neither  too  small nor too' large.  If
 the area selected is too small,  permit limits might  be
 unnecessarily controlled by  a criterion for an area  outside the
 site, whereas too large an area might  unnecessarily  incorporate
 spatial complexities that are not  relevant.to the discharge(s) of
 concern and thereby unnecessarily  increase the cost  of
 determining the WER.   Applying this principle is  likely  to be
 more difficult  for  large sites than for flowing-water  sites.

 Because WERs  for large sites  will  usually be~determined  using
 actual, rather  than-simulated,  surface water,  there  are  five
major considerations  regarding experimental  design and data
 analysis:

 1< Total recoverable WERs  at  large sites might vary  sovmuch
   across time,  location,  and depth that  they are not  very
   useful.  An assumption  should be developed that an
   appropriately defined WER will  be much more  similar 'across
   time, location, and depth .within the site  than will a total
 •  recoverable WER.  If  such an  assumption cannot be used, it  is
   likely that either  the FWER will have to be set equal to the
   lowest WER and be overprotective for most of the site or
   separate site-specific criteria will have to be derived for
   two or more sites.                                    ,
   a. One assumption that is likely to be worth jtestirig is that
      the dissolved WER varies much less across time, location,
      and depth within a site than the total recoverable WER.  If
      the assumption-proves valid, a dissolved WER cart be applied
      to a dissolved national water-quality criterion to derive a
      dissolved site-specific water quality criterion that will
      apply to the whole site.
  b. A second assumption that might be worth testing is that the
      WER correlates with a water quality characteristic such as
      TSS or TOG across time, location, and depth.
  c. Another assumption that might be. worth testing is that the
      dissolved and/or total recoverable WER is mostly due to

                                66

-------
    nohtoxic metal rather than to a water quality
    .characteristic, that reduces toxicity.  If this is true and
    if there is variability in the WER, the WER will correlate
    with the concentration of metal in the site water. .This is
    similar to the first assumption, but this one can allow use
    of both total recoverable' and dissolved WERs, whereas the
    first one only allows use of a dissolved WER.
 If WERs are too variable* to be useful and no way can be found
 to deal with the variability,  additional sampling will
 probably be required in order to develop a WER and/or, a site-
 specific water quality criterion that is either (a) spatially
 and/or temporally dependent or (b) constant and
.environmentally conservative for nearly all conditions.

 An experimental design should be developed that tests whether
 the assumption is of practical value across the range of
 conditions that occur at different times,  locations, and
 depths within the site.  Each design has to be formulated
 individually to fit .the specific site.  The design should try
 to take into account the times,  locations,  and depths at which
 the extremes of the physical,  chemical,  and biological
 conditions occur within the site.,  which will require detailed
 infonnatipn concerning the site.   In addition,  the
 experimental design should balance available resources with
 the need for adequate sampling.
 a. Selection of the number and timing of sampling events
    should take into account seasonal,  weekly,  and daily
  /•considerations.  Intensive  sampling should occur>during the
    two most extreme seasons, with confirmatory sampling durin'g
    the other two seasons.   Selection of  the day and time of
    sample collection should take into account the discharge
    schedules of the major industrial and/or municipal
    discharges.   For example, it  might be appropriate to
    collect samples during the  middle of  the week to allow for
    reestablishment.of steady-state conditions after shutdowns
    for weekends and holidays;  alternatively,  end-of-the-week
    slug discharges are routine in  some situations.   In coastal
    sites,  the tidal cycle might  be important if facilities
    discharge,  for example,  over  a  four-hour period beginning
    at slack high tide.   Because  the highest concentration of
    effluent in the surface water probably occurs at ebb tide,
    determination of WERs using site water samples  obtained at
    this time might result  in inappropriately large WERs that
    would result, in underprotection at  other times;  samples
    with_unusually large WERs might be  especially useful for
    testing assumptions.   The importance  of  each consideration
    should be determined on a case-by-case basis.
b.  Selection of the .number and locations of stations to be
    sampled within a sampling event should consider the site as
    a  whole and  take into account sources of water  and
    discharges,  mixing patterns, and currents (and  tides in
    coastal areas).   If  the site has been adequately

'••  ...    '-      •'',.-       67       •"..•'•      :

-------
    characterized, an acceptable design can probably be
    developed .using existing information concerning  (1) sources
    of the metal and other pollutants and-(2) the spatial and
    temporal distribution of concentrations of the metal and
   . water quality factors that might affect the toxicity of the
    metal.  Samples should not be" taken within or-near mixing
    zones or plumes of dischargers; dilution models  (U.S. EPA
    1993} and dye dispersion studies (Kilpatrick 1992) can
    indicate areas that should definitely be avoided.  Maps,
    .current charts, hydrodynamic models, and water quality '
    models used to allocate waste loads and derive permit
    limits are likely to be helpful when determining when and
    where to obtain site-water samples.  Available information
    might provide an indication of the acceptability of site
    water for testing selected species.  The larger and more
    complex the site,  the greater the number of sampling
    locations that will be needed.
 c. In addition to determining the horizontal location of each
    sampling station,  the vertical location (i.e.,  depth)  of
    the_sampling point needs to be selected.   Known mixing
    regimes,  the presence of vertical stratification of TSS
    and/or salinity,  concentration pf metal,  effluent plumes,
    tolerance of test  species,  and the  need to obtain samples
    of site water that span the range of site conditions should
    be considered when selecting the depth at which the sample
    is to be  taken.  Some decisions concerning depth cannot  be
    made  until information is obtained  at  the time  of sampling;
    for example,  a conductivity meter,  salinometer,  or
    transmissometer might be useful for determining where  and
    at what depth to collect samples.   Turbidity might
    correlate with TSS.and both might relate  to the"toxicity of
    the metal in site  water;  salinity can  indicate whether the
    test  organisms and the site water are  compatible.
 Because  each site is  unique,  specific  guidance cannot  be  given
 here  concerning either the selection of the  appropriate number
 and locations of sampling stations  within  a  site or the
 frequency  of sampling.   All  available  information concerning
 the site should be utilized to ensure  that the times,
 locations, and  depths  of samples span  the  range of  water
 quality  characteristics  that might  affect  the toxicity of the
metal:
    a. High and  low concentrations of TSS.
    b. High and  low concentrations of effluents.
    c. Seasonal  effects.
    d. The range  of tidal  conditions in saltwater situations.
The sampling plan should provide the data  needed to allow an
evaluation of the usefulness of the assumption(s) that  the
experimental design is intended to test.   Statisticians should
play a key role  in experimental design and data analysis, but
professional judgment that takes into account pertinent   ,
biological, chemical,  and toxicological considerations  is at
least as important as'rigorous statistical analysis when

                             68

-------
    interpreting the data and determining the degree to which the
    data correspond to the assumption(s).

 3. The details of .each sampling design should, be formulated with
    the aid of people who understand the site and people who have
    a working knowledge of WERs.  Because of the complexity of
    designing a WER study for large sites,  the design team should
    utilize the combined expertise and experience of individuals
    from the appropriate EPA Region,  states,, municipalities,
   , dischargers,  environmental groups,  and others who can
    constructively contribute to the design of- the study.
    Building a team of cooperating aquatic toxicologists, aquatic "
    chemists,  limnologists," oceanographers,  water quality
.   modelers,  statisticians,  individuals  from other key
    disciplines,, as  well as regulators  and those regulated, who
   have knowledge of the site and the  site-specific procedures,
  •  is central to success of the derivation .of a WER for a large
    site.   Rather than submitting the workplan to the appropriate'
   regulatory authority (and possibly  the Water Management
   Division of the  EPA Regional Office)  for comment at the  end,
   they should be members of the team from the beginning.

 4. Data from one sampling event should always be analyzed prior
   to the next sampling event with the goal of improving the
   sampling design  as the study progresses.  For example,  if the
   toxicity of the  metal in surface water samples is related tp
   the concentration of TSS,  a water quality characteristic  such
   as turbidity might be measured at the time of collection  of
   water samples and used in the selection'of the concentrations
   to l>e: used in the WER toxicity tests  in site water.   At a
   minimum,  the team that interprets the results of one sampling
   event  and plans  the next  should include an aquatic
   toxicologist,  a  metals chemist,  a statistician,  and a modeler
   ,or other user of the data.                               '

5. The, final  interpretation  of the data  and the derivation of the
   FWER(s)  should be performed by a team.   Sufficient data are
   likely to  be available to allow a quantitative estimate of
   experimental  variation, differences between species,  and
   seasonal differences.-  It will be necessary to decide whether
   .one site-specific criterion can be  applied to the whole area
;   or whether separate site-specific criteria need to be derived
   for two or more  sites.  The interpretation of the data might
   produce two or more alternatives  that  the appropriate
   regulatory authority could subject  to.a cost-benefit analysis.

Other aspects of the determination of  a WER for a large site are
likely to be  the same as described for Method 1.   For example:
a. WERs should be determined,using two or more sensitive species;
   the suggestions  given in  Appendix I should be considered  when
   selecting  the tests and species to  be  used.


       .   '   :    •        '. \     69     .•""-•-   '   --    ; '   .

-------
Chemical  analyses  of  site water,  laboratory  dilution water,
and  test  solutions should follow  the requirements  for the
specific  test" used and  those .given  in  this document.
If tests  in many surface water  samples are compared to. one
test in a laboratory  dilution water, it is very important that
that one  test be acceptable.  Use' of  (1) rangefinding tests,
(2)  additional treatments beyond  the standard five
concentrations plus controls, and (3)  dilutions that are
functions of the known  concentration-effect  relationships
obtained  with the  toxicity test and metal of concern will help
ensure that the desired endpoints and  WERs can be calculated.
Measurements of the-concentrations  of  both total recoverable"
and  dissolved metal should be targeted to .the test
concentrations whose  data will be.used in the calculation of
the  endpoints.
Samples of site water and/or effluent  should be collected,
handled,  and transported so that the tests can begin as soon
as is feasible.
If the large site  is  a  saltwater site, the considerations
presented in Appendix H ought to be given attention.        :
                         •   70

-------
 Figure 2:  Calculating an Adjusted Geometric Mean


 Where n = the number of experimentally determined WERs in, a set,
 the  "adjusted geometric mean"  of the set is calculated as
 follows:                             ,"'

 a. Take the  logarithm pf each of the WERs.  The logarithms can be
   to any  base,  but natural logarithms (base e) are preferred for
   reporting purposes.
 b. Calculate x = the arithmetic mean of the logarithms.
 c. Calculate, s = the sample standard deviation of the
 ,  logarithms:                            .
                                (x - x )2
                                 n - l   '             ,      -

   Calculate  SE• = the standard error of the arithmetic mean:
   SE - s/Jn .     __.'.•
   Calculate  A* x - (fc0.7) (SE) ,  where  t0-7 is the value of Student's
   t  statistic  for a  one-sided probability of 0.7D with n - l
   degrees of freedom.  The values of, t0>7  for some common
   degrees of freedom (df)  are:.                ^
 '-•". v  ••'•"              1         0.727
                           2         0.617   .
      ,                     30.584          '
                           4         0.569

      .                     5  ' ..."    0.559
                           6         0.553           •
                           7  .:.      0.549
                  •;.\.-;.' .8 _-       0.546                       ,

  :    -    \                 9  ."•.'.   0.543    '
                          10         0.542
                          .11         0.540
 J            '             12         Q-.539        .          -  ,
         "      -      '    -  -        .-"•••'    •   \ •
 .  The values of  t0>7 for more degrees  of freedom are available,
   for example, on page  T-5  of Natrella (1966).
f. Take the antilogarithm of A.-

This adjustment of the geometric mean  accounts  for the fact that
the means of fifty percent of the  sets of WERs  are expected to be
higher than the actual mean; using the one-sided value of t for
0.70 reduces the percentage  to thirty.:
                                71

-------
 Figure 3: An Example Derivation of a FWER


 This example assumes-that cccWERs were determined monthly, using
 simulated downstream water that was prepared -by mixing upstream
 water with effluent at the ratio that existed when the samples
 were obtained.   Also^ the flow of the effluent is always 10 cfs,
 and the design  flow of the upstream water is 40 cfs.   (Therefore,
 the downstream  flow at design-flow conditions is 50 cfs.)   The  '
 concentration of metal in upstream water at design flow is 0 4
 ug/L,  and the CCC is 2 ug/L.   Each FWER is derived from the WERs
 and hWERs that  are available  through that month.
Month
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan.
Feb.
eFLOW
(cfs)

 10
 10
 10
 10
 10 ,
 10
 10
 10
 10
 10 .
 10
 1.0
uFLOW
(cfs)

 850
 289
 300
 430
 120
  85
  40
  45
 150
 110
 180
 244
uCONC
(ucr/L)

 0.8,
 0.6
 0.6
 0.6
 0.4
 0.4 •
 0.4
 0.4
 0.4
 0.4
 0.6
 0.6
 WER
 5.2a
 6.0C
 5.8°
 5.7C
 7.0C
10.5e
12. Oe
11. Oe
 7.5C
 3.5C
 6.9°
 6.1C
  HCME
 (ua/L)

 826.4
 341.5
 341.6
 475.8
 177.2
 196.1
,118.4
 119.2'
 234.0
  79.6
 251.4
 295.2
                                                   hWER
82.80
34.31
34.32
47.74
17.88
19.77
12.00
12.08
23.56
 8.12
25.30
29.68
                          FWER
 1.0b
 1.0b
 1..0b
 5.7d
 5.7d
 6.80f
10.699
10.88g
10.889
 8.12h
 8.12h
 8.12h
   Neither Type 1 nor Type 2; the downstream flow  (i.e., the sum
   of the eFLOW and the uFLOW) is > 500 cfs.
   The total number of available Type 1 and Type 2 WERs is less
   than 3.                        .
   A Type 2 WER; the downstream flow is between 100 and 500'cfs
   No Type 1 WER is available; the FWER is the lower of the
   lowest Type 2 WER and the lowest hWER.              v
   A Type 1 WER; the downstream flow is between 50 and 100 cfs
   One Type 1 WER is available; the FWER is the geometric mean of
   all Type 1 and Type 2 WERs.
   Two or more Type 1, WERs are available and the range is less
   than a factor of 5; the FWER is the adjusted geometric mean
   (see Figure 2)  of the Type 1 WERs,  because all the hWERs are
   higher.                           ..-•  .
   Two or more Type 1 WERs are available and the range is not
   greater than a factor of 5; the FWER is the lowest hWER
   because the lowest hWER is lower than the adjusted geometric
   mean of the Type 1 WERs.
                               72

-------
 Figure 4: Reducing the  Impact  of  Experimental Variation
When the  FWER is  the,-lowest  of,  for example,  three WERs, the
impact  of experimental variation can be  reduced by conducting
additional primary  tests.  • If;tne endpoint  of the secondary test
is  above  the  CMC  or CCC  to which the FWER is  to be applied, the
additional tests  can also  be conducted with the secondary  test;
Month
April
May
June

Lowest
   .Case 1

   (Primary
     Test)

    4.801
    2.552
    9.164

    2.552
        (Primary
          Test)

         4.801
         2.552
         9.164
           Case,2

          (Primary
            Test)

           3.565
           4.190
           6.736
        Geometric
          Mean

          4.137
          3.270
          7.857

          3'. 270
Month
April
May
June

Lowest
           Case 3

(Primary  (Second.
  Test)      Test)
  4.801
  2.552;
  9.164
3.163
5.039
7.110
Geo.
 Mean

3'. 897
3.586
8.072

3.586
                               Case 4

                    (Primary  (Second.
                      Test)      Test)
4.801
2.552
9.164
3.163
2.944
7.110
Geo,
 Mean

3.897
2.741
8.072

2.741
Case 1 uses the individual WERs obtained with the primary test
for the three months, and the-FWER is the lowest of the three
,WERs.  In Case 2, duplicate primary tests were conducted in each
month, so that a geometric mean could be calculated for each
month; the FWER is the lowest of the three geometric means.

In Cases 3 and 4, both a primary test and a secondary test were
conducted each month and the endpoints for both tests in
laboratory dilution water are above the CMC or CCC to which the.
FWER is to be applied.  In both of these cases, therefore, the
FWER is the lowest of -the three geometric means.

The availability of these alternatives does not mean that they
are necessarily cost-effective.
                                73

-------
 Figure 5: Calculating an LC50  (or EC50) by Interpolation


 When fewer than two treatments kill some but not all of', the
 exposed test organisms, a statistically sound estimate of an LC50
 cannot be calculated.  Some programs" and methods produce LCSOs
 when there are fewer than two  "partial kills % but such results
 are obtained using interpolation, not statistics.  If (a) a test
 is otherwise 'acceptable, (b) a sufficient number of organisms are
 exposed to each treatment,  and (c) the concentrations are
 sufficiently close together, a test with zero or one partial kill
 can provide all the information that is needed concerning the
 LC50.  An LC50 calculated by interpolation should probably be
.called an "approximate LC50" to acknpwledge the lack of a
 statistical basis for its calculation,  but „. this does not imply
 that such an LC50 provides  no useful toxicological information.
 If desired,  the binomial test can be used to calculate a
 statistically sound probability that the true LC50 lies between
 two tested concentrations (Stephan 1977} .

 Although more complex interpolation methods can be used,  they
 will not produce a more useful LC50 than the method described.
 here.  Inversions in the data between two test concentrations
 should be removed by pooling the mortality data for those two
 concentrations and calculating a percent mortality that  is then
 assigned to  both concentrations.   Logarithms to a base other than
 10 can be used if desired.   If PI and P2  are the percentages of
 the test organisms that died when exposed to concentrations Cl
 and C2,  respectively,  and if   Cl < C2,    PI < P2,    0 < PI •£ 50
 and   50 £  P2 £'100,.   then:  -
                            p - 50 ~
                                P2 - PI

                    C = Log Cl + P(Log C2 - Log Cl)

                             LC50 = 10C
If PI = 0 and P2  =  100,  LC50
If PI = P2 = 50,  LC50 = t/(£D (C2)  .
If PI = 50, LC50  = Cl.
If P2 = 50, LC50  = C2.                    ,
If Cl = 4 mg/L, C2 = 7 mg/L,  PI  =  15  %, and  P2  =  100  %,
   then LC50 = 5.036565 mg/L.                   . .  .   :

Besides the mathematical requirements given  above* the following
toxicological recommendations are  given in sections G.8 and I 2-
a. 0.65 < C1/C2 < 0.99.                                    .
b. 0 £ PI < 37.   •
c. 63 < P2 £ 100.
                                74

-------
Figure  6:  Calculating a  Time-Weighted Average
                      ' •            -  '     "         J~     "'      '
                  ;      ,      :•    •-',';,.  1  .'••  .        .*"•:.   ' ,
If a  sampling plan  (e.,g.,  for measuring metal in a  treatment in a
toxicity test)  is designed so that  a  series  of values  are
obtained over time  in such a  way  that each value contains the
same  amount of  information (i.e./represents the same  amount of
time)/  then the most  meaningful average is the arithmetic
average.   In most cases, however, when  a  series  of  values is
obtained over time, some values contain more information than
others; in these cases the most meaningful average  is  a  time-,
weighted average  (TWA).  If each  value  contains  the same amount
of information, the arithmetic average  will  equal the  TWA.

A TWA is obtained by  multiplying  each value  by a weight, and  then
dividing the sum of the  products  by the sum  of the  weights.   The
simplest approach is  to  let each  weight be the duration  of time
that  the sample represents.   Except for the  first and  last
.samples, the period, of time represented by a sample starts
halfway to the  previous  sample and  ends halfway  to  the next
sample.  The period of time represented, by the first sample
starts  at  the beginning  of the test,  and  the period of time
represented by  the  last  sample ends at  the end of the  test.   Thus
for a 96-hr toxicity  test,  the sum  of the weights will be 96 hr.

The following are hypothetical examples of grab  samples  taken
from  96-hr flow-through  tests for two common sampling  regimes:

Sampling   Cone.    Weight   Product  -  Time-weighted average
time  (hr)  (mcr/L).   (hr)1   (hr) (ma/L)'         (ma/L)

    0       12        48
   96       14        48
                      96       1248      "   1248/96.= 13.00

    0        8        12
   24        6        24
   48        7        24
   72        9        24
   96        8       12.       	
                      96       720           720/96  = 7.500


When all the weights  are the  same,  the  arithmetic average equals
the TWA.   Similarly,  if  only  one  sample is taken, both the
arithmetic average  and the TWA equal  the  value of that sample.

The rules  are more  complex for composite  samples  and for samples
from renewal tests.   In  all cases.,  however,  the  sampling plan can
be designed so  that the  TWA equals  the  arithmetic average.
                                75

-------
                             REFERENCES
 ASTM.   1993a.   Guide /for Conducting Acute Toxicity Tests with
 Fishes, Macroinvertebrates,  and Amphibians.   Standard E729
 American Society for Testing and Materials,  Philadelphia, PA. -

 ASTM.   1993b.   Guide for Conducting Static Acute  Toxirity Tests
 Starting with Embryos of Four Species of  Saltwater Bivalve
 Molluscs.  Standard  E724.  American Society  for Testina  and
 Materials, Philadelphia,  PA.         .                  H      :

 ASTM.   1993c.   Guide for Conducting Renewal  Life-Cycle Toxicity
 Tests with Daphnia macma.  Standard E1193 .   American  Society for
 Testing and Materials, Philadelphia, PA.

 ASTM.   1993d.   Guide for Conducting Early  Life-Stage  Toxicitv
 Tests with Fishes.   Standard E1241.  American Society for Testino
 and Materials,  Philadelphia, PA.

 ASTM.  1993e.   Guide  for Conducting Three-Brood,  Renewal Toxicity
 Tests with Cenodaphnia dubia.  Standard E1295.  American Society
 for Testing and Materials, Philadelphia, , PA.         .'."..

 ASTM.  1993 f.  Guide for Conducting Acute Toxicity Tests on
 Aqueous Effluents with Fishes, Macroinvertebrates, and
 ?™thouse<  L-W'*'  G'w- Suter'  A. E. Rosen, and J.J. Beauchamp.
 1987.  . Estimating Responses of Fish Populations to Toxic
.Contaminants.   Environ. .Toxicol. Chem. 6:811-824.
              and ?-J-  Versteeg«   1992-   A Statistical Procedure
      o        Contlnuous Toxicity. Data,,  Environ.  Toxicol., Chem.
 11:1485-1494.               .

 Hoekstra,  J.A. ,  and P.H.  Van Ewijk.   1993.   Alternatives for the
 No-Observed-Ef feet . Level .  Environ.  Toxicol. Chem,  12:487-194.

 Kilpatrick,  F.A.  1992.   Simulation  of  Soluble Waste Transport
 ^ „?£     P  in Surface Waters Using  Tracers. ,  Open-File Report
 ?i7Ji  m 5*S'-,Geolog:LCal  Survey'  Books  and  Open-File Reports, Box
 25425, Federal Center,  Denver, CO 80225.
          **'G- . 1966-  EJtperimental  Statistics.  National  Bureau
of Standards Handbook  91.   (Issued August  1,  1963;  reprinted
October 1966 with corrections) .  U>S. Government Printing  Office,
Washington, DC.
                                76

-------
  Prothro,  M.G.   1993.   Memorandum titled "Office of Water Policy
  and Technical  Guidance on Interpretation and Implementation of
  'Aquatic Life Metals Criteria". •  October 1.  •
                     .'"*.'              "         _.        •-
  Stephan,  C.E.   1977.   Methods  for. Calculating an LC50.   In:
  .Aquatic Toxicology and Hazard  Evaluation.   (F.L.  Mayer  arid J.L.
  Hamelink,  eds.)  ASTM STP 634.   American Society for. Testing and
  Materials,  Philadelphia,  PA.   pp.  65-84.

  Stephan,  C.E.,  and J.W. Rogers.   1985.   Advantages of Using
  Regression Analysis to Calculate Results of  Chronic Toxicity
  Tests.   In:  Aquatic Toxicology and Hazard Assessment: Eighth
 .Symposium.   (R.C.  Banner  and D.J.  Harisen, eds.)   ASTM STP 891.
  American Society for  Testing and Materials,  Philadelphia,  PA.
  pp.  328-338.

  Suter,  G.W., A.E.  Rosen,  E. Linder,  and D.F.  Parkhurst.   1987.
  Endpoints for  Responses of Fish  to Chronic Toxic  Exposures.
  Environ.  Toxicol.  Chem. 6:793-809.

  U.S. EPA.   1983.a.   Water  Quality Standards Handbook.  Office of
  Water Regulations  and.Standards,  Washington,  DC.          .

  U.S. EPA.   1983b.   Methods for Chemical Analysis  of Water and
  Wastes.   EPA-600/4-79-020.  National Technical Information
  Service,, Springfield,  VA.

  U.S. EPA.   1984.   Guidelines for Deriving Numerical Aquatic  Site-
  Specific  Water Quality Criteria  by Modifying  National Criteria.
  EPA-600/3-84-099   orPB85-121101.   National  Technical
  Information  Service,  Springfield,  VA.

  U.S. EPA.   1985.   Guidelines for Deriving Numerical National
  Water Quality  Criteria for the Protection of  Aquatic  Organisms
.and Their Uses.  PB85 -227 04 9..:. National. Technical Information
  Service,  Springfield,  VA.

  U.Si EPA.  1991a.   Technical Support pocument for Water  Quality-
  based Toxics Control.   EPA/505/2-90-boi  or   PB91-127415.
  National  Technical Information Service,  Springfield,  VA.

  U.S. EPA.  1991b.   Manual  for  the Evaluation  of Laboratories
  Performing Aquatic Toxicity Tests.   EPA/600/4-90/031.  National
  Technical Information Service, Springfield, VA.

  U.S. EPA.  1991c-.   Methods for the Determination  of Metals in
  Environmental  Samples.  tEPA-60Q/4-91-OiO.  National Technical
'Information  Service,  Springfield,  VA.
                                 77

-------
U.S. EPA.   1992.   Interim Guidance  oh  Interpretation and
Implementation  of .Aquatic Life  Criteria  for Metals.  Office of
Science and Technology, Health  and  Ecological Criteria Division'
Washington, DC.

U.S. EPA.   1993a.  Methods for  Measuring the Acute Toxicity of
Effluents and Receiving Waters  to Freshwater and Marine
Organisms.  Fourth Edition.  EPA/600/4-90/027F.  National
Technical Information Service,  Springfield, VA.

U.S. EPA.   1993b.  Short-term Metlrods  for Estimating the Chronic
Toxicity of Effluents and Receiving Waters to Freshwater
Organisms.  Third-Edition.  EPA/600/4-91/002.  National Technical
Information Service, Springfield, VA.           .

U.S. EPA.   1993c.  Short-Term Methods  for Estimating the Chronic
Toxicity of Effluents and Receiving Waters to Marine and
Estuarine Organisms.  Second Edition.  EPA/600/4-91/003.
National Technical Information  Service,  Springfield, VA.

U.S. EPA.  1993d.  Dilution Models  for Effluent Discharges.
Second Edition.  EPA/600/R-93/139.  National Technical
Information Service, Springfield, VA.
                               78

-------
Appendix A: Comparison of WERs Determined Using Upstream and
            Downstream Water

                     **      -"'       ' ' '       F •        *
The "Interim Guidance" concerning metals (U.S. EPA 1992) made a
fundamental change in the way WERs should be experimentally
determined because it changed the source of the site water.  The
earlier guidance  (U.S. EPA 1983,1984) required that upstream
water be used as the site water, whereas the newer guidance" (U.S.
EPA 1992) recommended that downstream water be used as the site
water.  The change in the source of the site water was merely an
acknowledgement that the WER that applies at a location in a body
of water should, when possible, be determined using the water
that occurs at that location.

Because the change in the source of the dilution water was
expected to result in an increase in the magnitude of many WERs,
interest in and concern about the determination and use of WERs
increased.  When upstream water was the required site water, it
was expected that WERs would generally be low and that the
determination and Use ,of WERs could be fairly simple.  After
downstream water became the recommended site water, the
determination and use of WERs was examined much more closely.   It
was then realized that the determination and use of upstream WERs
was more complex than originally thought.   It was also realized
that the us,e of downstream water greatly increased the complexity
and was likely to increase both the magnitude and the variability
of many WERs.  Concern about the fate of discharged metal also
increased because use of downstream water might allow the ,    :
discharge of large amounts of metal that has reduced or no
toxicity at the end of the pipe.  The probable increases in the
complexity, magnitude, and variability of WERs and the increased
concern about fate, increased the importance of understanding the
relevant issues as they apply to WERs determined using both
upstream water and downstream.water.  -,.	


A. Characteristics of the Site Water

   The idealized concept of an upstream water is a, pristine water
   that is relatively unaffected by people.  In the rea-1 world,
   however, many upstream waters contain naturally occurring
   ligands, one or more effluents, and materials from nonpoint
  • sources; all of these might impact a WER.  If the upstream
   water receives an effluent containing TOC and/or TSS that
   contributes to the WER> the WER will probably change whenever
   the quality or quantity of the TQC and/or TSS changes.  In
   such a case, the determination and use of the WER in upstream
   water will have some of the increased complexity associated
   with use of downstream water and some of the concerns
   associated with multiple-discharge situations (see Appendix
   F) .  The amount of complexity will.depend greatly on the

      •     '   :   . '" .   :  \  '"  79    *     .        ' :       .....

-------
  number and type of upstream point and nonpoint sources  the
  frS?™V?d' ^gnituje of fluctuations,  and whe?he? the 5IR
  3 SrSS^      " B*~ *» *** of complete mix
    Downstream water is a mixture of -'effluent and upstream water
    each of whxch can contribute to the WER, and so there^re twd
    components to a WER determined in 'downstream water- the
    effluent component and the upstream component . .  The existence
    ?  ™se,3two comP°nents has the following implications-
    -L. WERS determined using downstream water are likely to be
       larger and more variable than WERs determined usino
       upstream water.                                   u
    2. The effluent component should be applied only where the
       effluent occurs,  which has implications concerning
       implementation.                                '
    3. The magnitude of the effluent component of a WER will
       depend on the concentration of effluent in the downstream
       water.  (A consequence of this is that the effluent   *
       component will be. zero where the concentration of effluent
       is zero,- which is the point of, item 2 above  )      eicj-uent
    4. The magnitude of  the effluent component of a  WER is likelv
       to vary as the composition of the effluent varies
    5. Compared to upstream water,  many effluents contain higher
       concentrations of a wider variety of substances that  can
       impact the toxicity of metals in a wider variety o? way's,
       and so the effluent component of a WER can be due to  a
       SSJfS °   f^0?1 ef f ects  ih addition to such factors as
       hardness,  alkalinity,  pH,  and humic acid.

           UStot?Le^1USnt  comSone2t "^ht be  due,  in whole or  in
           Iv  r^e  dxscharse of  refractory metal  (see Appendix
           25 JJ* ?ann?t  be  Bought of simply as being Sused by
             ffZ**0* Water ^alitY on the  toxicity of  the metal.
           with  downstream WERs  is  so  much simpler if  the
             PS  (?WES ' a?d the uPstream WER (SwER) are  additive
      LS  x? desirable to understand  the concept of additivity
    ppe^dix GK  experxmental determination, and its use  (see


B. The Implications of Mixing Zones.

   When WERs are determined using upstream water, the presence or
   absence of mixing zones has no impact; the cmcWER and the
   cccWER will both be determined using site water that contains
   S?f? S^ST °? the effluent of conSem, i.e. , the bwS W^S
   will be determined using the same site water.

   When WERs are •determined using downstream water,  the magnitude
           ™  will probably depend on the concentration of
  f                                                   o
ettiuent in the downstream water used  (see Appendix D)   The
concentration of .effluent in the site water wi?? depend on
                             80

-------
where the sample is  taken, which Will  not be  the  same  for  the
cmcWER and the.cccWER if  there  are mixing zone(s).  Most,  if
not  all,  discharges  have  a chronic  (CCC) -mixing zone;  many,
but  not all,  also have an acute (CMC)  mixing  zone.  The CMC
.applies at all  points except  those inside a CMC mixing zone;
thus if there is no  CMC mixing  zone, the CMC  applies at the
end  of the pipe.  The CCC applies at all points outside the
CCC  mixing zone.  It is generally assumed that if permit
limits are based on  a point in  a stream at which  both  the  CMC
and  the CCC apply, the CCC will control the permit  limits,
although  the  CMC might control  if different averaging  periods
are  appropriately taken into  account.   For this discussion, it
will be assumed that the  same design flow  (e.g.,  7Q10) is  used
for  both  the  CMC and the  CCC.

If the cmcWER is to  be appropriate for use inside the  chronic
mixing zone,  but the cccWER is  to be appropriate  for use
.outside the chronic  mixing zone, the concentration  of  effluent
that is appropriate  for use in  the determination  of the two
WERs will not be the.same.  Thus even  if the  same,toxicity
test is,used  in the  determination of the cmcWER and the
dccWER, the two WERs will probably be  different because the
concentration of effluent will  be different in the  two site
waters in which the  WERs  are  determined.

If the CMC is only of concern within the CCC  mixing zone,  the
highest relevant concentration  of metal will  occur  at  the  edge
of the CMC mixing zone if there is.a CMC mixing zone;  the
highest concentration will occur at the end of the  pipe if
there is  no CMC mixing zone.  In contrast, within the  CCC
mixing zone,  the lowest cmcWER  will probably  occur  at  the
outer edge of the CCC mixing  zone.  Thus the  greatest  level of
protection would be  provided  if the cmcWER is determined using
water at  the  outer edge of the  CCC mixing zone, and then the
calculated site-specific  CMC  is applied -at the edge of the CMC
mixing zone or  at the end of  the pipe,  depending  on whether
there is  an acute mixing  zone.   The cmcWER is likely to be
lowest at the outer  edge  of the CCC mixing zone because of
dilution  of the effluent, but this dilution will  also  dilute
the  metal.  If  the cmcWER is  determined at the outer edge  of
the  CCC mixing  zone  but the resulting  site-specific CMC is
applied at the  end of the pipe  or at the edge of  the CMC
mixing zone,  dilution is  allowed to reduce the WER  but it  is
not  allowed to  reduce the concentration of the metal.  .This
approach  is environmentally conservative, but it  is probably
necessary given current implementation procedures.  (The
situation might be more complicated if the uWER is  higher  than
the  eWER  or if  the two WERs are less-than-additive.)

A comparable  situation applies.to the  CCC.  Outside the CCC
mixing zone,,  the CMC and  the  CCC both  apply,  but  it, is assumed
that the  CMC  can be  ignored because the CCC will  be more

• ' -. ''• :•-••••'''      '   .  .81   •':.••'..     .     •'.-.'

-------
    SJ2S?   ?'   ??S G9cWE\should probably be determined for the
    complete-mix situation,  but the site-specific CCC will have to
    £5 *S£naS the fdge °f the CCC mining zone.   Thus dilution of
    the  WER from the edge of the CCC mixing zone to the- point of
    complete mzx is  taken into account,  but dilution of the metal
    3.S HO                         '""'"
       ™o   *"* !^ithe,r an  acute nor a  chronic mixing  zone, both
   the CMC and the CCC apply at the end  of  the pipe,  but  the CCC
   should sto.ll be determined  for the  complete-mix  situation.


C. Definition of site.

   In the general context 'of site-specif ic  criteria,  a  "site" mav
   be a state, region, watershed, waterbody, segment  of a
   waterbody, category of water (e.g.,, ephemeral streams) , etc. ,
   but the site-specific criterion  is to be derived to provide
   adequate protection for the entire site, however the site is
   defined.  Thus, when a site-specific criterion is derived
   using the Recalculation Procedure, all species that "occur at
   the site- need to be taken. into account when deciding what
   species,  if any,  are to be deleted from the dataset.
   SS1 a£ yT'ra»hen a site-specific criterion is derived using a
   WER,  the WER is to be adequately protective of the entire
   sice,   if, for example,  a site-specific criterion is being
   derived for an estuary,  WERs could be determined, using samples
   2w £  surface water obtained from various sampling stations,
•   H ?i-to av°ld confus:i-on,  should not be called "sites".   If
•   Jll the WERs were sufficiently similar,  one site-specific
   S Te™=°n COUld 5e.deriv,ed to apply to the whole estuary.   If '
   the WERs  were  sufficiently different,  either the lowest WER
   could be  used  to. derive  a site-specific criterion for the .
                         data ^^t  indicate that the  estuary
                         two- or more sites, each with its own
                                             ~         •
                       ^hat  ^wuld'be applied when defining the
  SoSld hS^??!™38? in th*sit* is very simplistic:  The site
  should be neither too  small  nor too large.
  1. Small sites  are probably  appropriate for cmcWERs, but
     usually are  not appropriate  for cccWERs  because rnetals are
     persistent,  although some oxidation  states are riot
     persistent and some, metals are  not persistent in  the water
     column.  For _ cccWERs,  the smaller the defined site,  the
     more likely  it is that the permit limits will be  controlled
     by a criterion for  an  area that is outside the site,  but
     which could  have been  included  in the site without
     substantially  changing the .WER  or increasing the  cost of
     determining  the  WER.
  2. Too large an area might unnecessarily increase the cost of
     determining  the  WER.   As  the size of  the site increases,
                               82

-------
    the spatial and temporal variability is likely to increase,
    .which will probably increase the number'of water samples in
    which WERs will need to be determined-before a site-
    specific criterion can be derived.                 .
 3. Events that import or resuspend TSS and/or TOC are likely.
    to increase the total recoverable concentration of the
    metal and the total recoverable 
-------
    WheneWERs are determined Using downstream water, the followino
    defined^10™3 should be taken into account when the site is
       If a^site-specific criterion is derived using a WER that
       SiX1^6? fc? the.complete-mix situation, the upstream edge of
       the site to which this criterion applies should be the
       point at which complete mix actually occurs.  If the site
       to which_the complete-mix WER is applied starts at the end
       of the pipe and extends all the way across the stream,
       there will be an area beside the plume that will not be
       adequately protected by the site-specific criterion.
       Upstream of the point of complete mix, it will usually be
       protective to apply a site-specific criterion that was
       derived using a WER that was determined using upstream
    3. The plume might be an area in which the concentration of
       metal could exceed a site-specific criterion without
       causing toxicity because of simultaneous dilution of the
       metal and the eWER.  The fact that the plume is much larger
       than the mixing zone might not be important if there is  no
       toxicity within the plume.  As long as the concentration of
       metal in 100 % effluent does not exceed that allowed by  the
       additive portion of the eWER,  from a toxicological
       standpoint neither the size nor the definition of the.plume
       needs ^ to be of concern because the metal will not cause
       tK^L^S111 thS ?lume-  ff ^ere is np toxicity  within
       tne plume,  the area in the plume might be like a
       traditional mixing zone in that the concentration of metal
       exceeds  the site-specific  criterion,  but it  would be
       different from a traditional mixing zone in  that the level
       of  protection is not reduced.                            •

   Special  considerations  are likely to be necessary in order  to
   take into account the eWER when defining a site related to
   multiple discharges, (see Appendix F) .  ;       ,  ,   ,


D. The variability in the  experimental" determination of a WER.

   When a WER is determined using upstream water,  the two major
   sources of variation  in  the WER are  (a) variability  in the
   ?^/  y*?  the  !lte water' which might be related to season
   and/or flow, and  (b)  experimental variation.  Ordinary day-to-
   day variation will account for some of 'the variability,  but
   seasonal variation is likely to be more important.

   As explained in Appendix D, variability in the concentration   -
   °J £°?£OXXC dissolved metal.will contribute to the variability
   of both total recoverable WERs and dissolved WERs; variability
   in the concentration of nontoxic particulate metal will
   contribute to the variability in a total recoverable WER, but
   not to the variability in a dissolved WER.  Thus, dissolved
                               84

-------
WERs are expected  to be  less variable  than  total  recoverable
JWERs, especially where events commonly increase1 TSS  and/or
TOC.  In some cases, therefore, appropriate use of analytical
chemistry can greatly increase the usefulness of  the
experimental determination of WERs.  The .concerns regarding
variability are'increased if an up'stream effluent contributes
to the _ WER.           •

When a WER is determined in downstream water, the four major
sourqes of variability in the WER are  (a) variability in the
quality of the upstream  water, which might  be related to
season and/or flow,  (b)  experimental variation,  (c)
variability in the composition of the  effluent, and  (d)
variability in the ratio of the flows  of the upstream water
and the effluent.  The considerations  regarding the  first two
are the same as for WERs determined using upstream water;
because of the additional sources of variability, WERs
determined using downstream water are  likely to be more
variable than WERs determined using upstream water.

It would be desirable if a sufficient  number of WERs could be
determined to define the variable factors in the  effluent and
in the upstream water that contribute  to the variability in
WERs that' are determined using downstream water.  Not only is
this likely to be  very difficult inmost cases, but  it is also
possible that the  WER will, be dependent on  interactions
between constituents of  the effluent and the upstream water,
i.e., the eWER and uWER  might be additive,  more-than-additive,
or less-than-additive (see Appendix G).  When interaction
occurs, in order to completely understand the variability of
WERs determined using downstream water, sufficient tests would
have to be conducted to  determine the  means and variances of:
   a. the effluent component of the WER.
   b. the upstream component of the WER.
   c. any interaction between the two  components.
An interaction might occur, for example, if the toxicity of a
metal is affected  by pH, and the pH and/or  the  buffering
capacity of the effluent and/or the upstream water vary
considerably.

An increase in the variability of WERs decreases  the     ^
usefulness of any  one WER.  Compensation for this decrease in
usefulness can be  attempted by determining  WERs at more times;
although this will provide more data,  it will not necessarily
provide a proportionate  increase in understanding.  Rather
than determining. WERs at more times, a better use of resources
might be to obtain more  information concerning  a  smaller
number of specially selected occasions. ,

It is likely that  some cases will be so complex that achieving
even a reasonable  understanding will require unreasonable
resources.  In contrast, some WERs determined using  the

      :      ' ...  .    .       :  85-'         .    '    '••.'.       -  -

-------
     methods presented herein might be relatively easy to

     S£!r^sdaj!             °hemiCal «*««-*.  «,  performed
     1.  If  the variation of  the total  recoverable WER is
        substantially greater than  the variation  of the comparable
        dissolved  WER,  there is probably a variable and substantial
        concentration of particulate nontoxic metal.   It  might be
        advantageous  to use  a dissolved WER just  because  it will
        have less  variability than  a total recoverable WER
     2.  If  the total  recoverable and/or dissolved WER correlates
        with the total  recoverable  and/or dissolved concentration
        of metal in the site water,  it  is likely  that  a substantial
        percentage of the metal  is  nontoxic.  In  this  case the WER
        will probably also depend on the concentration of effluent
        in  the site water and on the concentration of metal in the
        e£x -Luent .                                       .          •
    These approaches are more likely , to be useful when WERs are
    determined using downstream water, rather than upstream water,
    unless both the magnitude of the WER and the concentration7 of '
    the metal in the upstream water are elevated by an upstream
    effluent and/or events that increase TSS and/or TOG.

    Both of these Approaches can be applied to WERs that  are
    °!!t2rm?-ned us^ng actual downstream water,  but the second can
    E£2 *S y Pr?vlde much better -information if it is used with
    WERs determined using simulated downstream water that is
    prepared by mixing a sample of the effluent with a sample of
    the upstream water.  In this way the  composition and
    ri^?5S;tl?* 2f b°? Jhe effluent  and the  upstream water  .
    can be  determined,  and the exact ratio  in the downstream water
    xs
    Use of simulated, downstream water is 'also  a way to  study  the
    ESS"? between Jhe WER and the. ratio  of  effluent  to upstream
    ££? r. at pnf .??int -in tine,- which is  the most  direct wa?  to
    test for additivity  of the  eWER and the uWER  (see Appendix G)
    This can be  viewed as. a test of the assumption that WERs
    determined using downstream water will  decrease as  the
    ?™en~a£i°Vf e?fluent ^creases.  If this  assumption  is
    true,  as the flow increases,  the concentration of effluent in
    ObLS^ J^ ™a*er Wi^ decrease and  the WER, will decrease.
    Obtaining ^ such information  at one point in time is useful, but
    confirmation at  one  or more other times would  be much more .
   usenuj. .                                             .         •
E.
The fate of 'metal that has reduced or no toxicity.
              haS reduced or no toxicity at the end of the pipe
              ^6 ^°X1^ at Some time in the future.  For example,
     ™   £at x? ^ the water col™i and is not toxic now might
   become more toxic in the water column later or might move into

                                86             '.'./;

-------
the sediment and become toxic.  If a WER allows a surface
water to contain as much toxic metal as is acceptable, the WER
would not be adequately protective if metal that was nontoxic
when the WER was determined became toxic in the water .column,
unless a,compensating change occurred.  Studies of the fate of
metals need to address hot only the changes that take place,
but also the rates of the changes.

Concern about the fate of discharged metal justifiably raises
concern about the possibility,that metals might contaminate
sediments.  The'possibility of contamination of sediment by
toxic and/or nontoxic metal in, the water column was one of the
concerns that led to the establishment of EPA's sediment
quality criteria program, which is developing guidelines and
criteria to. protect sediment.  A separate program was
necessary because ambient water quality criteria are not
designed to protect sediment.  Insofar as technology-based
controls and water quality criteria reduce the discharge of
metals, they tend to reduce the possibility of contamination
of sediment.  Conversely, insofar as WERs allow an increase in
the discharge of metals, they tend to increase the possibility
of contamination of sediment.'

When WERs are determined in upstream water, the concern about
the fate of metal with reduced or no toxicity is usually 'small
because the WERs are usually small.  In addition, the factors,
that result in upstream WERs being greater than 1.0 usually
are (a) natural organic materials such as humic acids and (b)
water quality characteristics such as hardness, alkalinity,
and pH.  It is easy to assume that natural organic materials
will not degrade rapidly, and it is easy to monitor changes-in
hardness, alkalinity, and. pH.  Thus there is usually little
concern about the fate;of the metal when WERs are determined
in upstream water, especially if the WER is small.  If the WER
is large and possibly due at :least.\in part to an upstream
effluent, there is more concern about the fate of metal that
has,reduced or no toxicity.

When WERs are determined in downstream water, effluents are
allowed to contain virtually unlimited amounts of nontoxic
particulate metal and nontoxic dissolved metal.  It would-seem
prudent to obtain some data concerning whether the nontoxic
metal might become toxic at some time in the future whenever
(1) the concentration of nontoxic metal is large, (2) the
concentration of dissolved metal is below the dissolved
national criterion but the concentration of total recoverable
metal is substantially above the.total recoverable, national
criterion, or r(3) the site-specific criterion is substantially
above the national criterion.  It would seem appropriate to:
a. Generate some data concerning whether "fate"  (i.e.,
   environmental processes) will cause any of the nontoxic
   metal to become toxic due to oxidation of organic matter,
              ,              , -                        !
    • .  •   •   ••'••;   ....        s?   '   • •'    .  •         .   .   '  •

-------
       oxidation of sulfides",  etc.  For example,  a WER could be
       determined using a sample of actual or simulated downstream
       water,  the sample .aerated for a period of  time (e.g.,  two
       weeks),  the pH.. adjusted if necessary,  and  another WER
       determined.  If aeration reduced the WER,  shorter and
       longer periods of aeration could be used to study the rate
       of change.
    b.  Determine the effect of a change in water  quality
       characteristics on the  WER; for example, determine the
       effect  of lowering the  pH on the WER if influent lowers  the
       pH of the downstream water within the  area to which the
       site-specific criterion is to apply.
    c.  Determine'a WER in actual downstream water to demonstrate
       whether downstream conditions change sufficiently (possibly
       due to  degradation of organic matter,  multiple dischargers,
       etc.)  to lower the WER  more than the concentration of  the
       metal is lowered.                       .
    If  environmental processes cause nontoxic metal  to become
    toxic,  it  is important to  determine whether the  time scale
    involves days,  weeks,  or years.              ,


 Summary                                                     .

 When WERs  are  determined using downstream water,  the site water
 contains  effluent and the WER will  take into account hot only  the
 constituents of the upstream  water,  but also the  toxic  and
 nontoxic metal and other constituents  of the effluent as they
 exist  after mixing .with  upstream water.   The determination of  the
 WER automatically takes  into  account any additivity,  synergism,
 or  antagonism  between  the metal and components of the effluent
 and/or the upstream water.  The effect  of calcium, magnesium, and
 various.heavy  metals on  competitive binding  by such  organic
 materials  as humic acid  is also taken  into account.   Therefore, a
 site-specific  criterion  derived .using.,a,WER  is likely to be more
 appropriate for a  site than a national, -state, or recalculated
 criterion  not  only because it takes  into account the water
 quality characteristics  of the site  water but also because it
 takes  into account other constituents.in the effluent and
 upstream water.

 Determination  of WERs  using downstream water causes  a, general
 increase in the  complexity, magnitude, and variability of WERs,
and an  increase  in concern about  the fate of metal that has
reduced or no  toxicity at  the end of the,pipe.  In addition,
there are  some other drawbacks with  the  use  of downstream water
in the determination of a WER:
1. It might serve  as a.disincentive  for  some dischargers to
   remove any more organic  carbon and/or particulate matter than
   required, although WERs  for some metals will not be related to
   the concentration of TOG or TSS.


                                88         .

-------
                                     *'
 2.  If conditions change, a WER might decrease in the future.
 ,   This is not a problem if the decrease is due to a reduction in
    nontoxic metal,  but it might be a problem if the decrease is
    due to a decrease in TOG or TSS or an increase in competitive
    binding.          •
 3.  If a WER is determined when the. effluent contains refractory
    metal but a change in operations results in the discharge of
    toxic metal in place of refractory metal, the site-specific
    criterion and the permit limits will not provide adequate
    protection.  In  most cases chemical monitoring probably will
    not detect such  a change,  but toxicological monitoring
    probably will.

 Use of WERs that are determined using downstream water rather
 than upstream water increases:
 1.  The importance of understanding the various issues involved '-In
    the determination and use of WERs.
 2.  The importance of obtaining data that will provide
  . understanding rather than obtaining data that will result in
    the highest or lowest WER.
 3.  The.appropriateness  of site-specific criteria.
 4.  The resources needed to determine a WER.
 5.  The resources needed'to use a WER.
 6. .The resources needed; to monitor the acceptability of the
    downstream water.
 A WER determined using, upstream water will  usually be smaller,
 less  variable,  and  simpler to implement than a WER determined
 using downstream water.   Although in some situations a downstream
 WER might be smaller than an  upstream .WER,  the important
 consideration is that a WER should be determined using the water
 to which it is to apply.  .
References    -  ."''••'.'.."'  '.-..-.--	••-.•••'• .'...--•/--• - •-'•	..'.. ..v ,.„'... ...

U.S. EPA.  1983.  Water Quality  Standards Handbook.   Office of
Water Regulations and Standards, .Washington, DC.

U.S. EPA.  1984.  Guidelines  for Deriving Numerical Aquatic Site-
Specific Water Quality  Criteria  by Modifying National Criteria.
EPA-600/3-84-099 ' or  PB85-121101.  National Technical
Information Service,-Springfield, VA.
          ....-.'•   •      *  '. •   •  •       ~ ,
U.S. EPA.  1992.  Interim Guidance on .Interpretation  and
Implementation of Aquatic Life Criteria  for Metals.   Office of
Science and Technology, Health and Ecological Criteria Division,
Washington, DC.
                                89

-------
 Appendix B: The Recalculation Procedure


 NOTE: The National Toxics Rule (NTR) does not allow use of the
       Recalculation Procedure in the derivation of a'. site-
       specific criterion.  Thus nothing in this appendix applies
       to jurisdictions that are subject to the NTR.


 The Recalculation Procedure is intended to cause a site-specific
 criterion to appropriately differ from a national aquatic life
 criterion if justified by demonstrated pertinent toxicological
 differences between the aquatic species that occur at  the site
 and those that were used in the derivation of the national
 criterion.  There are at least three reasons why such  differences
 might exist between the two sets  of species.  First, the national
 dataset contains aquatic species  that/are sensitive to many
 pollutants, but these and comparably sensitive species .might  not
 occur at the site.   Second,  a species that is critical at the
 site might be sensitive to the pollutant and require a lower
 criterion.  (A critical species is  a species that is commercially
 or recreationally important at the  .site,  a species that  exists at
 the site and is listed as threatened or endangered under section
 4 of the Endangered Species Act,  or a species for .which there is
 evidence that the loss of the species from the site is likely to
 cause an unacceptable impact on a commercially or recreationally
 important species,  a threatened or  endangered species,  the '
 abundances of a variety of other  species,  or the structure or
 function of the community.)   Third,  the species that occur at the
 site might represent a narrower mix of  species than those in  the
 national dataset due to a limited range of natural environmental
 conditions.   The procedure presented here  is structured  so that
 corrections and additions can be  made to the national  dataset
 without  the deletion process being  used to take into account  taxa
 that do  and do not  occur at  the site; in effect,  this  procedure
 makes it possible to update  the national aquatic life  criterion.

 The  phrase -occur at the site"  includes the species, crenera,
 families,  orders, classes, and phyla that:
 a. are usually present  at the site.
 b. are present at the site only seasonally due to migration.
 c. are present intermittently because they periodically  return to
   or extend their  ranges into  the  site.
 d. were present at'the  site  in  the past, are not currently
   present at  the site  due to degraded  conditions, and are
   expected  to return to  the site when  conditions improve.
 e. are present in nearby  bodies5 of water,  are not currently
   present at  the site  due to degraded  conditions, and are
   expected  to be present at  the site when conditions improve.
The taxa that  -occur  at the  site" cannot be'determined merely by
sampling downstream and/or upstream  of  the site at one point  in
time.  "Occur  at the  site" does not  include taxa that were once

                                90

-------
present at the site but cannot exist at the site now due to
permanent physical alteration of the habitat at the site
resulting from dams, etc.           "        •
      ,;        '       , •           .         '• .               «
The definition of the "site" can be extremely important when
using the Recalculation Procedure.  For example, the number of
taxa that occur at the site will generally decrease as the size
of the site decreases. .Also., if the site is defined to be very
small, the permit limit might be, controlled by a criterion that
applies outside (e.g., downstream of)  the site.

   Note: If the variety of aquatic invertebrates, amphibians, and
         fishes is so limited that species in fewer than eight
         families occur at the site, the general Recalculation
         Procedure is not applicable and the following special
         version of the Recalculation Procedure must be used:
         1. Data must be available for at least one species in
            each of the families that occur at the site.,.
         2. The lowest Species Mean Acute Value that is available
            for a species that occurs at the site must be used as
            the FAV.
         3. The site-specific CMC and CCC must be calculated as
            described below in part 2 of step E, which is titled
            "Determination of the CMC and/or CCC".

The concept of,the.Recalculation Procedure is to create a dataset
that is appropriate for deriving a site-specific criterion by
modifying the national dataset in some, or all of three ways:
   a. Correction, of data that are in the national dataset.
   b. Addition of data to the national dataset.
   c. Deletion of data that are  in the national dataset.
All corrections and additions that have been approved by U.S. EPA
are required, whereas use of the deletion process is optional.
The Recalculation Procedure is more likely to result in lowering
a criterion if the net result of addition and deletion is to
decrease the number of genera in the dataset, .whereas, the
procedure, is more likely to result in raising a criterion if the
net result of addition and deletion is to increase the number of
genera in the dataset.      '                   -

The Recalculation Procedure .consists of the following steps:
A. Corrections are made in the national dataset.         . ' „     .
B. Additions are made to the national dataset.
C: The deletion process may be applied if desired.   :
D. If the new dataset does not satisfy the applicable Minimum
   Data Requirements  (MDRs), additional pertinent data must be
   generated; if the new data are approved, by the IKS. EPA, the
   Recalculation Procedure must  be started again at  step B with
   the addition of the new data.
E. The new CMC or CCC or both are determined.
F. A report is written.                     v      ••"...
Each step is discussed in more detail below.

                   '"•.-.-    .91.   "•••".'  '      ••'".'•

-------
 A, Corrections         .           f                    .

 1. Only corrections approved by the U.S. EPA may be made.
 2. The concept of "correction" includes removal of data that
    should not have been in the national dataset in the first
    place.  The concept of "correction" does not include removal
    of a datum frbm the national dataset just because the quality
    of the datum is claimed to be suspect.  If additional data are
    avaxlable -for the same species, the U.S. EPA will decide which
    data should.be used, based on the available guidance (U.S. EPA
    1985); also, data based on measured concentrations are usually
    preferable to those based on nominal -concentrations.
 3. Two kinds of corrections are possible:
    a. The first includes those corrections that are known to and
       have been- approved by the U.S. EPA; a list of these will be
       available from the U.S.  EPA.
    b. The second includes those corrections that are submitted to
       the U.S.  EPA for approval.   If approved,  these will be
       added to EPA's list of approved corrections.
 4. Selective corrections are not  allowed.  All  corrections on
    EPA's newest list must be made.
                                         ,'    .    '

 B. Additions

 1. Only additions approved by  the U.S.  EPA may  be made.
 2. Two kinds of additions are  possible:   ,  '                    .
    a.  The first includes those additions that are known  to and
      .have been approved by the" U.S.  EPA;  a list of these  will be
       available from the U.S.  EPA.
    b.  The second includes those additions that  are  submitted to
       the U.S.  EPA for approval.   If approved,  these will  be
       added to  EPA's  list of approved additions.
 3.  Selective additions are not allowed.   All additions on  EPA's
    newest list  must be made*,   "    .


 C,  The Deletion Process

 The basic principles  are:
 1.  Additions  and corrections must be made as per steps A and B
    above, before  the  deletion  process is performed.,
 2.  Selective  deletions  are no't allowed.   If any  species is  to be
    deleted, the deletion process described below must be applied
    to  all species in  the national dataset, after any necessary
    corrections and additions have been made to the national
    dataset.   The deletion  process specifies which species must be
    deleted and which species must not be deleted.  Use of the
    deletion process is  optional, but no deletions are optional
   when the deletion process is used.                        '
3. Comprehensive information must be available concerning what
    species occur at the sit'e; a species cannot be deleted based

                                92

-------
   on incomplete information concerning the  species  that do and
   do_.not satisfy the definition of  "occur at the  site".
4. Data might have to be generated before the deletion process is
   begun:             -,                       '-•..;•
   a. Acceptable pertinent toxicological data must,be available
      for at least /one species in each class of aquatic plants,
      invertebrates,-amphibians, and fish that contains a -species
L   .   that is a critical species at  the site.
   b. For each aquatic plant, invertebrate,  amphibian, and fish
      .species that occurs at the site and is listed  as threatened
      or endangered under section 4  of the Endangered Species
      Act, data must be available or be generated  for an
      acceptable surrogate species.  Data for each surrogate
      species must be used as if they are data for species that
      occur at the site.
   If additional data are generated  using acceptable procedures
'-.  (U.S. EPA 1985) and they are approved by  the U.S. EPA, the
   Recalculation Procedure must be started again at  step B with
   the addition of the new data.  .                             :  ,
5. Data might have to be generated after the deletion process is
   completed.  Even if one or more species are deleted, there
   still are MDRs. (see step P below) that must be  satisfied.  If ,
   the data remaining after deletion, do not  satisfy  the
   applicable MDRs, additional toxicity tests must be conducted
   using acceptable procedures (U.S. EPA 1985) so  that all,MDRs
   .are satisfied.  If the new data are approved by the U.S. EPA,
   the Recalculation Procedure must  be started again at step B
   with the addition of new data.                     ,
6. Chronic tests,, do not have to be conducted^ because the national
.   Final Acute-Chronic Ratio (FACR)  may be used in the derivation
   of the site-specific Final Chronic Value  (FCV) .   If acute-"
   chronic ratios (ACRs): are available or are generated so that
   the chronic MDRs are satisfied using only species that occur
   at the site, a site-specific FACR may be  derived  and used in
   place of the national FACR. .Because..a.FACR was not used in
   the derivation of the freshwater  CCC for  cadmium, this CCC can
   only be modified the same way as  a FAV; what is acceptable
   will depend on which species are  deleted.
               "~ . :  ' V       .         '  »' •

If any species are to be deleted, the following deletion process
must be applied:
   a. Obtain a copy of the national  dataset,i.e., tables 1, 2,
      and,3 in the national criteria document (see Appendix E).
   b. Make corrections in and/or additions to the  national
      dataset as described in,steps  A and B  above.
   c. Group all the species in the dataset taxonomically by
      phylum, class,  order, family,  genus, and species.
   d. Circle each species that satisfies the definition of "occur
      at the site" as presented on the first .page  of this
      appendix, and including any data for species that are
      surrogates of threatened or endangered species that occur
      at the site.

-------
e. Use the following step-wise process to determine
   which of the uncircled species must be deleted and
   which must not be deleted: '          .

   1.  Does the genus occur at the site?
        If
        If
             "No", go to step 2.
             "Yes", are there one or more species in the genus
                   that occur at the site but are not in the
                   dataset?
                      If"No", go to step 2.
                      If "Yes", retain the uncircled species.*

    2. Does the family occur at the site?
         If "No", go to step 3.                 ,
         If "Yes", are there one or more genera in the family
                   that occur at the site but are not in the
                   dataset?                  ,
                     ^If "No", go to step 3.
                      If "Yes", retain the uncircled species.*

    3. Does the.order occur at the site?
         If "No", go to step 4.       ,     .
         If "Yes", does the dataset contain a circled species
                   that is  in the same order?
                      If "No",  -retain the uncircled species.*
                      If "Yes", delete the uncircled species.*

    4.  Does the class occur at the site?
         If "No",  go to step 5.
         If "Yes",  does the dataset contain a circled species
                   that is  in the same class?
                      If "No",  retain the uncircled species.*
                      If "Yes",  delete the uncircled species.*

    5.  Does the phylum occur, at ..the site?
         If "No",  delete the uncircled species.*
         If "Yes",  does  the dataset contain a circled species
                   that  is  in the same phylum?
                     If "No",  retain the uncircled species.*
                     If "Yes",  delete the  uncircled species.*

    * = Continue  the  deletion process, by starting at  step 1 for
       another uncircled species unless  all  uncircled species
       in  the  dataset have been considered.

The species that are circled and those that  are retained  >
constitute the site-specific dataset.   (An example  of the
deletion process is given  in Figure  Bl.)

This deletion process is designed  to ensure  that:
a. Each species that occurs both in  the national dataset and
   at the  site also occurs  in the  siter-specific dataset.
                           94

-------
   b. Each  species  that occurs.at  the  site but  does  not. occur in
      the national  dataset  is  represented in  the  site-specific
      dataset bv all  species in the national  dataset that  are in
      the same genus..-
   c. Each  genus that occurs at the site but  does not occur  in
      the national  dataset  is  represented in  the  site-specific
      dataset by all  genera in the national dataset  that are in
   .   the same family.
   d. Each  order, class, and phylum that occurs/ both in the
      national dataset and  at  the  site is represented in the
      site-specific dataset by the one or more  species in  the
      national dataset that are most closely  related to a  species
 /..-•  that  occurs at  the site.


D. Checking the Minimum Data Recruirements
                               •';'•'    •         .  -             .
The initial, MDRs for  the Recalculation Procedure  are the same as
those for the derivation of a  national .criterion.  If a specific
requirement cannot  be satisfied after  deletion  because that  kind
of species  does not occur at the site, a taxonomically similar
species must be substituted in order to meet  the  eight MDRs:

   If no species of the kind required  occurs  at the  site, but a
   species  in the same order does, the MDR can  only  be satisfied
   by data  for a species that  occurs at the site  and is in that
   order; if no species in  the order occurs at  the site, but a
   species  in the class does,  the  MDR  can only  be satisfied  by
   data for a species that  occurs  at the site and is  in .that
   class.  "If no species in the same class occurs at  the site,
   but a species in the phylum does, the MDR  can  only be
   satisfied by data  for a  species that occurs,  at the site and is
  '• in that  phylum.  If no species  in the same phylum occurs  at
   the site, any species that  occurs at the site  and is not  used
   to satisfy a different .MDR  can  be used,-to  satisfy the MDR.  If
   additional data  are generated using .acceptable procedures
   (U.S.. EPA 1985)  and they are approved by the U.S.  EPA,-the
   Recalculation Procedure  must be started again  at  step B with
   the addition of  the new  data.

If fewer than eight families of aquatic invertebrates,
amphibians, and fishes occur at the site, a Species Mean Acute
Value must  be available for at  least one species  in  each-of  the
families and the special version of the Recalculation Procedure
described on the second page of this appendix must be used.


E. Determining the CMC and/or  CCG

1. Determining the FAV:     ^-.-.-''."
   a. If the eight  family MDRs  are satisfied, the site-specific
      FAV must be calculated from  Genus Mean  Acute Values using

   •':•'*''"'•    '''  •          95                 ••- '    ;  ':  ..  • •

-------
       the procedure described in the national aquatic life
      , guidelines  (U.S. EPA 1985).
    b. If fewer than eight families of aquatic invertebrates,
      •amphibians, and fishes occur at the site, the lowest
       Species Mean Acute Value that is available for a species
       that occurs at .the site must be used as the FAV, as per the
       special version of the Recalculation Procedure described on
       the second page of this appendix.              .
 2. The site-specific CMC must be calculated by dividing the site-
    specific FAV by 2.  The site-specific FCV must be calculated
    by dividing the site-specific FAV by the1 national FACR (or by
    a site-specific FACR if one is derived).   (Because a FACR was
    not used to derive the national CCC for cadmium in fresh
    water, the site-specific CCC .equals the site-specific FCV.)
 3. The calculated FAV, CMC,  and/or CCC must  be lowered,  if
    necessary, to (1)  protect an aquatic  plant,  invertebrate,
    amphibian, or fish species that is a  critical species at the
    site,  and (2)  ensure that the criterion is.not likely to
    jeopardize the continued existence of any endangered or
    threatened species listed under section 4 of the Endangered
    Species Act or result in the destruction  or adverse
    modification of such species'  critical habitat.


 FT Writing the Report

 The report of the results of use of the  Recalculation  Procedure
 must include:            ,
 1. A list of all  species of  aquatic invertebrates,,  amphibians,
    and fishes that are known to "occur at the site", along with
    the source of"the  information.
 2. A list of all  aquatic plant,  invertebrate,  amphibian,  and fish
    species that are critical species  at  the  site, including all
    species that occur at the site and are listed as threatened  or
    endangered under section  4 of. the  Endangered Species  Act.
 3.  A site-specific version of Table 1 from a criteria  document
    produced by the U.S.  EPA  after 1984.
 4.  A site-specific version of Table 3 from a criteria  document
    produced by the U.S.  EPA  after 1984.           •
 5.  A list of all  species that were deleted.
 6.  The new calculated FAV, CMC, and/or CCC.
 7.  The lowered FAV, CMC,  and/or CCC,  if  one  or more were lowered
    to protect a specific species.


Reference

U.S.  EPA.   1985.  Guidelines  for Deriving Numerical National
Water Quality Criteria  for the  Protection of Aquatic Organisms
and Their Uses.   PB85-227049.  National  Technical Information
Service,  Springfield, VA.            ...",'             •


                                96

-------
Figure Bl: An Example of the Deletion Process Using Three Phyla
SPECIES THAT ARE IN THE THREE PHYLA AND OCCUR AT THE SITE.
Phylum    Glass     Order      Family     Species
Annelida  Hirudiri.  Rhynchob.
Bryozoa   {No species in this
Chordata  Osteich.. Cyprinif.
Chordata  Osteich.  Cyprinif.,,
Chordata  Osteich.  Cyprinif.
Chordata  Osteich.  Cyprinif.
Chordata  Osteich.  Salmonif.
Chordata  Osteich.  Percifor.
Chordata  Osteich., Percifor.
Chordata  Amphibia  Caudata
 Glossiph. ... Glossip. complanata
phylum occur at the site.) ,
 Cyprinid.  Carassius auratus
 Cyprinid.  Notropis anogenus
 Cyprinid.  Phpxinus eos
 Catostom.  Carpiodes carpio
 Osmerida.  Osmerus mordax
 Centrarc.  Lepomis cyanellus
 Centrarc.  Lepomis humilis
 Ambystom.  Ambystoma gracile-
SPECIES THAT ARE IN THE THREE PHYLA AND IN THE NATIONAL DATASET
Phylum    Class     Order      Family   ... Species   ,         Code

                                          Tubifex tubifex      P
                                          Lophopod. carteri    D
                                          Petromyzon marinus   D
                                          Carassius auratus    S
                                          Notropis hudsonius   G
                                          Notropis stramineus  G
                                          Phoxinus eos         S
                                          Phoxirius oreas       D
                                          Tinea tinea          D
                                          Ictiobus bubalus     F
                                          Oncorhynchus mykiss  O
                                          Lepomis cyanellus    S
                                          Lepomis macrochirus  G
                                          ...Perca flavescens     D
                                          Xenopus laeyis       C
Annelida
Bryozoa
Chordata,
Chordata
Chordata
Chordata
Chordata
Chordata
Chordata
,Chordata
Chordata
Chordata
Chordata
Chordata
Chordata
Oligoch.
Phylact.
Cephala .
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Osteich.
Amphibia
Haplotax .
- — - : ' - " .
Petromyz «
Cyprinif.
Cyprinif.
Cyprinif.
Cyprinif .
Cypririif .
Cyprinif .
Cyprinif .
Salmonif.
Percifor.
Percifor.
Percifor.
Anura
Tubifici.
Lophopod.
Petromyz .
Cyprinid .
Cyprinid,
Cyprinid,
Cyprinid ,
Cyprinid.
Cyprinid.
Catostom,
Salmonid.
Centrarc .
Centrarc .
Percidae
Pipidae
Explanations of Codes:
  S = retained because  this Species occurs at the site.
  G = retained because  there  is a species in this Genus that
      occurs at the  site but  not in the national dataset.
  F = retained because  there  is a genus in this Family that
      occurs at the  site but  not in the national dataset.
  O = retained because  this. Order occurs at the site and  is not
      represented by a  lower  taxon.
  C = retained because  this Class occurs at the site and  is not
      represented by a  lower  taxon.
  P = retained because  this Phylum occurs at the site and is  not
      represented by a  lower  taxon.
  D = deleted because this  species does not satisfy any of the
      requirements for  retaining species.
                                97

-------
  Appendix C: guidance Concerning the Use of "Clean Techniques" and
              QA/QC when Measuring Trace Metals


     Note: This version of this appendix contains more information
           than the version that was Appendix B of. Prothro  (1993°?
        -             (shiller and Boyle_1987; Windom et al. 1991)
  has raised questions concerning the quality of .reported
  concentrations of trace metals ,in both fresh and salt (estuarin-
  and marine  surface waters,  A lack- of awareness of true ambient
  concentrations of metals in fresh and salt surface waters^an £e
  both a cause and a result of the problem.   The ranges of
  dissolved metals that are typical in surface waters of the Unit^r
  ?QKe%a™a? fro™the immediate influence of discharges m?uXnd
  af ? 19^ ale:     ** * 1985'1987*'  Trefry et al.  1986; wlndom et
            Metal
                  Salt water
           Cadmium
           Copper
           Lead
           Nickel
           Silver
           Zinc
                0.01
                0.1
                0.01
                0.3
to
to
to
to
            metals
    p
 some metals
                              Fresh water
                                (u.cr/LV	

                             0.002  to  0.08
                             0.4    to  4.  ,
                             0.01   to  0-.19
                             1.     to: 2.

                             0.03   to  5.

          has published analytical methods for
      waters and wastewaters, but  these methods
      (termination of ambient concentrations of
some surface waters.  Accurate and precise
T™ rl^^C' ^i i <•*».» «••«•* wk ^t* A.« J_ __.	 A_ • _ .       •
          0.2
          3.
          1.
          5.
0.005 to  0.2
0.1   to 15.
°"Clean
                              during, collecting, handling,
                                              to avoid
 2*  J|ed°f ,«* -"ifled. reference ,
6. Use of replicates  to assess precision.
Z_! P3?, °? certif ied standards.    .
           :,J!e?se/Jthe  term "clean techniques"  refers to
           that reduce  contamination and enable the accurate .
^..^  measurement of  trace metals in fresh and salt surface
waters.  In a broader sense, the  term also refers to related
issues concerning  detection limits,  quality control? and quality

                                98

-------
  assurance.'  Documenting data.quality demonstrates ;the amount of
  confidence that can be placed in the data,  whereas  increasing the
  sensitivity of methods reduces the problem of deciding how to
  interpret results that: are reported to be below detection, limits.

  This appendix is written for those analytical laboratories that
  want guidance concerning ways to lower detection limits,  increase
  accuracy,  and/or increase precision.  The ways to achieve these
  goals are to increase the sensitivity of the analytical methods,
  decrease contamination,  and decrease interference.   Ideally,
  validation of a procedure for measuring concentrations of metals
  in surface water requires demonstration that agreement can be
  obtained using completely different procedures beginning with the
  sampling step and continuing through the. quantification step
  (Bruland et.al. 1979), but few laboratories have the resources to
  compare two different procedures.   Laboratories can*  however,  (a)
  use techniques that others have found useful for improving
  detection limits, accuracy,  and precision,  and (b)  document data
  quality through use of blanks,  spikes.,, CRMs,  replicates,  and
 •standards.                                 '

:  Nothing contained or not contained in this appendix adds to or
  subtracts from any regulatory requirement set forth in other EPA
  documents concerning analyses of metals.  A WER can be acceptably
  determined without the use of clean techniques as long as the
  detection limits,, accuracy/  and precision are acceptable.  No
  QA/QC requirements beyond those that apply to measuring metals in
  effluents are necessary for the determination of WERs.  The word
  "must" is not used in this appendix.  891116 items, however,  are
  "considered so important by analytical, cnemists who  have worked to
  increase accuracy and precision and lower detection limits in
  trace-metal analysis that "should" is in bold print to draw
; attention to the item.  Most such items are emphasized because
  they have been found to have received inadequate attention in
  some laboratories performing; trace-metal .analyses	     ,

  In general, in order to achieve accurate and precise measurement
•  of a particular concentration,  both the detection limit and the.
  blanks should be less .than one-tenth of that concentration.
  Therefore, the term "metal-free" can be interpreted to mean that
  the total amount of contamination that occurs during sample
  collection and processing (e.g., from gloves, sample containers,
  labware, sampling apparatus, cleaning solutions, air,  reagents,
  etc.)  is.sufficiently low that blanks are less than one-tenth of
  the lowest concentration that needs to be measured.

  Atmospheric particulates can be a major source of contamination
  (Moody 1982; .Adeloju and Bond 1985).  The term "class-100" refers
  to a specification concerning the amount of particulates in air
  (Moody 1982); although the specification says nothing about the
  composition of the particulates, generic control, of particulates
  can greatly reduce trace-metal blanks.  Except during collection

         • .  '.-:.-..-.   " •  ..:'  ..     •  99  "": •••-'       •. •'    '"..-' •.

-------
 of samples,  initial cleaning of equipment,  and handling of
 samples  containing high concentrations  of metals,  all  handling  of
 samples,  sample containers,  labware,  and sampling  apparatus
 should be performed in a class-100  bench, room,  or glove .box.

 Neither  the  "ultraclean techniques »" that might be  necessary when
 trace  analyses  of  mercury are performed nor safety in  analytical
 laboratories is addressed herein.   Other documents should be
 consulted if one or both of  these topics are of  concern.
Avoiding contamination by use of "clean techniques"

Measurement of trace metals in surface waters should take into
account the potential for contamination during each step in the
process.  Regardless of the specific procedures used for
collection, handling, storage, preparation (digestion,
fj*?S£i?nV£nd/°r fraction)  and quantification (instrumental
analysis), the general principles of contamination control should
£>e applied.  Some specific recommendations are:
   Powder-free (non-talc,  class-iaO) latex, polyethylene,  or
   polyvinyl chloride (PVC,  vinyl)  gloves should be worn during
   all steps ^ from sample collection to analysis.  (Talc seems to
   be a particular problem with zinc;  gloves made -with  talc
   cannot be decontaminated sufficiently.)   Gloves should only
   contact surfaces that are metal-free;  gloves should  be changed
   if even suspected of contamination.                      ««««*
   The acid used to acidify .samples for preservation and
   digestion and to acidify water for  final cleaning of labware,
   sampling apparatus,  and sample containers  should be  metal-
   free.   The  quality of the acid used should be better than
   reagent-grade.   Each lot  of acid should be analyzed  for the
   metal (s)  of interest before use.
   The water used to prepare acidic cleaning. solutions  and to
   rinse  labware,  sample containers, and  sampling apparatus  may
   ^/^Pa™d^/istillation' deionization" or reverse osmosis,
   and should  be  demonstrated to  be metal-free.
   The work  area,  including bench tops and  hoods,  should be
   cleaned  (e.g.,  washed and wiped  dry with lint-free,  class-100
   wipes) frequently to  remove  contamination.
a
  Srf *Sdi=*g °f^Sa^1^S ±n the ^oratory, including filtering
  and analysis, should be performed in a class-100 clean bench
  or a glove box fed by particle-free air or nitrogen; ideally
  100 clean ro    °r gl°Ve box should be located within a class -

  «£?a£e/ ^sents, sampling apparatus, and sample containers
  2?SSS ?^Ver ?S   ,L0p.en to the atmosphere; they should be
  stored in a class-100 bench, covered with plastic wrap, stored
  in a plastic box, or turned upside down on a clean surface.
  Minimizing the time between cleaning and using will help
  minimize contamination.

                              100

-------
 Separate sets of sample containers,  labware,  and sampling
 apparatus should be dedicated, for different kinds, of samples;
 e.g.,  surface water samples,  effluent samples,  «tc.
 To avoid contamination of clean rooms,  samples  that  contain
 very high concentrations of metals and do not require use of
 "clean techniques"  should not be brought into, clean  rooms.
 Acid-cleaned plastic/.such as high-density polyethylene
 (HDPE),  low-density polyethylene {LDPE),  or a fluoroplastic,
 should be the only  material that ever contacts  a sample,
 except possibly during digestion for the total  recoverable
 measurement.
 1. Total recoverable samples  can be digested in some plastic
    containers. '
 2. HDPE .and LDPE might not be acceptable for mercury.
 3. Even if acidified,  samples and standards containing silver
    should be in amber containers.
 All labware,  sample containers,  and sampling apparatus should
 be acid-cleaned before use or reuse.
 1. Sample containers,  sampling apparatus,  tubing,  membrane
    filters,  filter  assemblies,  and other labware should be
    soaked in acid until metal-free.   The amount of cleaning
    necessary might  depend on  the amount of contamination  and
    the length of time ,the item will  be in contact with ,  .
    samples.   For example,  if  an acidified sample will  be
    stored in a sample container for three weeks,  ideally  the
    container should have been soaked in an acidified metal-
    free solution for at least three weeks.
 2. It  might  be desirable to perform initial cleaning,  for
    which reagent-grade acid may be used,  before the  items are
    taken into a clean, room.   For most metals, items  should  be
    either (a)  soaked in 10 percent concentrated nitric acid at
    50°C for  at least one hour,  or  (b)  soaked, in 50 percent
    concentrated;nitric acid at room temperature for  at least
    two days;  for arsenic and  mercury,  soaking for up to two
    weeks at- 50°C in 10 percent.concentrated nitric acid might
    be  required.   For plastics,that .might  be damaged  by strong
    nitric acid,  such as polycarbonate and possibly HDPE and
    LDPE,  soaking in 10 percent concentrated hydrochloric  acid,
    either in place  of or before soaking in a nitric  acid
    solution,  might  be desirable.
.3. Chromic acid should not be used to clean items that will be
    used in analysis of metals.
 4. Final soaking and cleaning of sample containers,  labware,
    and sampling apparatus should be  performed in a class-100
    clean room using metal-free acid and water.   The  solution
    in  an acid bath  should be  analyzed periodically to*
    demonstrate that it is metal-free.
 Labware,  sampling apparatus,  and sample containers should be
 stored appropriately after cleaning:
 1. After the  labware and sampling 'apparatus are cleaned,  they
    may be stored in a clean room in  a weak acid bath prepared
    using metal-free acid and  water.   Before use,  the items

                            101

-------
        should be rinsed at least three times with metal-free
        water.  After the final rinse,  the items should be moved
        immediately,  with the open end pointed down,  to a  class-100
        clean bench.   Items may be dried on a class-100 clean
        bench; items  should not be dried in an oven or with
        laboratory towels.   The sampling apparatus should  be
        assembled in  a class-100 clean room or bench  and double-
        bagged in metal-free polyethylene zip-type bags for
        transport to  the field;  new bags are usually  metal-free
    2.  After sample  containers  are cleaned,  they should be filled
        with  metal-free water that has  been acidified to a pH of 2
        with  metal-free nitric acid (about  0.5  mL per liter)  for
        storage until  use.               ,       '   •    r
 1. Labware,  sampling apparatus,  and sample containers  should be
    rinsed and not rinsed with  sample  as necessary to prevent hiah
    and low  bias  of analytical  results  because acid-cleaned
    plastic  will  sorb some metals  from unacidified solutions
    1.-Because samples  for the  dissolved measurement are not'
        acidified  until  after  filtration,, all sampling apparatus,
        sample  containers, labware, filter holders, membrane
        filters, etc., that contact the  sample before or during
        filtration should be rinsed with a portion of the solution
        and then that portion discarded.
    2. For the total recoverable measurement, labware, etc.,  that
       contact the sample only before it is acidified should be
       rinsed^ith sample, whereas items that contact the sample
      .after it is acidified should not be rinsed.  For example,
       the sampling apparatus should be rinsed because the sample
       ri <.?0t be aci3ified until it is in a sample container^
       but the sample container should not be rinsed if the sample
       •will be acidified in the sample container.
    3. If the total recoverable and dissolved measurements are  to
       be performed on the same sample  (rather than on two samples
       obtained at the same time and place), all the  apparatus and
       labware, including the sample container, should be  rinsed
       before the sample is placed in the sample container; then
       an unacidified aliquot should be removed for the total
       recoverable measurement (and acidified,  digested, etc.) and
       an unacidified aliquot should be removed for the dissolved
    •  measurement (and filtered,  acidified, etc.)   (If a
       container is rinsed and filled with sample and an
      unacidified aliquot  is removed for the dissolved
      measurement and then the  solution in .the container  is
      acidified before removal  of an aliquot for the total
      recoverable measurement,  the resulting measured total
      recoverable concentration' might  be biased high because the
      acidification.might  desorb  metal that had been sorbed  onto
      the walls of the sample container; the amount  of bias  will
      depend on the relative volumes j.nvt>lved  and on the  amount
      of sorption and desorption.)               •
m. Field samples  should be  collected in a  manner that  eliminates
   the potential  for  contamination from sampling platforms,
                               102

-------
    probes,  etc.   Exhaust from boats and the direction of wind and
    water currents should be taken into account.   The people who
    collect  the samples should be specifically trained on how to
 •-,. collect  field samples.  After collection,  all handling of
    samples  in the field that will expose the sample to air should
    be performed in a portable class-I'O.O clean bench or glove box.
 n.  Samples  should be acidified (after filtration if dissolved
  -  metal is to be measured)  to a pH of less than 2,  except that
    the pH should be less than 1 for mercury.   Acidification
    should be done in a clean room or bench,  and.so it might be
    desirable to wait and acidify samples in a laboratory rather
    than in  the Afield.   If samples are acidified  in the field,
 -   metal-free acid can be transported in plastic  bottles and
    poured into.a plastic container from which acid can be removed
  ,  and added to samples using plastic pipettes.   Alternatively,
  ^plastic  automatic dispensers can be used.
 o.  Such things as probes and thermometers should  not be put in
    samples  that  are to be analyzed for metals.  In particular,  pH
    electrodes and mercury-in-glass thermometers should not be'
    used if  mercury is  to be  measured.   If pH  is measured,  it
    should be done on a separate aliquot.
 p.  Sample handling should be minimized.   For  example,  instead  of
    pouring  a sample into a graduated cylinder to  measure the
    volume,  the sample  can be weighed after being  poured into a
    tared container,  which is less likely to be subject to error
    than weighing the container from which the sample is poured.
    (For saltwater samples, the salinity or density should be
    taken into account  if weight is converted  to volume.)
 q.  Each reagent  used should  be verified to be metal-free.   If
   metal-free reagents are not commercially available,  removal of.
   metals will probably be necessary.
 r.  For the  total recoverable measurement,  samples should be
 .  digested in a class-100 bench,  not  in a metallic  hood.   If
    feasible,  digestion should be done  in the  sample, container  by
   acidification and heating.      '.                   .
 s. The longer the time between collection and analysis  of
    samples,  the  greater the  chance of  contamination,  loss,  etc.
 t. Samples  should be stored  in the dark,.preferably  between 0  and
   4°C with no air space in  the sample container.
Achieving low detection limits

a. Extraction of the metal from the sample can be extremely
   useful if it simultaneously concentrates the metal and
   eliminates potential matrix interferences.  For example,
   ammonium 1-pyrrolidinedithiocarbamate and/or diethylammonium
   diethyldithiocarbamate can extract cadmium, copper, lead,
   nickel, and zinc  (Bruland et al. 1979; Nriagu et al. 1993).
b. The detection limit should be less than ten percent of the
   lowest concentration that is to be measured.

   ,            ,                103

-------
          -                   ,           .             •,_,-
  Avoiding interferences

  a.  Potential interferences should be assessed for the specific
     mealSeT    analysis  technique used  and for each me?S to be

  b.  If  direct analysis  is  used,  the salt  present in high-salinitv
     saltwater samples is likely  to cause  interference in most
     instrumental  techniques.                    «ience in most
  c.  As  stated above, extraction  of the metal  from the sample is
     particularly  useful because  it simultaneously concentrates the
     metal and eliminates potential matrix- interferences
 Using blanks to assess contamination
                     ^^^^^^^"^^^^^^™"*^™""^^""'^^^^~'^~
                .  (Prob?dura1' method) blank consists of filling a
              tain^-?^h analvzfd metal-free water and procesIiSg
              ,, acidifying, etc.) the water through the laboratnrS
    procedure in .exactly the same way as a sample   A laboratSrS^
    blank should be included in each set of Sn or f ewe? sSp?2
    to check for contamination in the laboratory,  anl shouS
    S S3^       a2 ten percent of the lowest concentration that
    is to be measured.  Separate laboratory blanks should be
    processed for  the total recoverable and dissolved
    measurements ,  if both measurements are performed .
                                                    .
            etc   coll-                         rugtu
            etc..^  collecting the. water in a sample container
   blSSj So2l!hS WatSr the,same  as  a field SmpL?  A SSi
   blank should be processed for  each sampling  trip    Senar£i-
   field blanks should be processed  for  the ?otal ?ecovSable
   measurement and for the  dissolved measurement,  if

   irtErfaSStSrv'th SitS*  Fi?ld. blanks Sh0uid b
   in tne laboratory the same as  laboratory blanks.
Assessing accuracy
                       should be determined for each analytical
                                  stsrs-
   ol                          9 S°»    e  -l— •*. each g.oup
      tdSitions^Pike 
-------
    2. A CRM, if one is available in a matrix that closely
       :approximates that of the samples.   Values obtained for the
       CRM should be within the published values.
 The concentrations in .-blind standards and solutions,  spikes, and
 CRMs should not be more than 5 times the median concentration
 expected to be present in the samples.
 Assessing precision

 a.  A sampling replicate should be included with each set of
    samples collected at each sampling location.            '
 b.  If the volume of. the sample is large enough, replicate
    analysis of at least one sample should be performed along with
    each group of about ten samples.



.Special considerations concerning the dissolved measurement

 Whereas total recoverable measurements are especially subject to
 contamination during digestion,  dissolved measurements are
 subject to both loss and contamination during filtration.
 a.  Because acid-cleaned plastic sorbs metal from unacidified
    solutions and because samples for  the dissolved measurement
    are not acidified before filtration,  all sampling apparatus,
    sample containers,  labware,  filter holders,  and membrane
    filter.s that contact the sample before-or during  filtration
    should be conditioned by rinsing with a portion of the
    solution and discarding that portion.    ,
 b.  Filtrations should be performed using acid-cleaned plastic
    filter holders and acid-cleaned membrane filters.   Samples
    should not be' filtered through glass  fiber filters,  even if
    the filters, have been cleaned with acid,. Mf positive-pressure
    filtration is used,  the air or gas should be passed through a
    0.2-nm in-line  filter;  if vacuum filtration  is used, it  should
    be  performed on a class-100  bench.
 c.  Plastic filter holders should be rinsed and/or dipped between
  :-. filtrations, ^buf they do not rhave  to  be soaked between
    filtrations if all  the samples contain about the  same
    concentrations  of metal.  It  is best  to filter samples from
    low to high concentrations.   A membrane filter should not  be
    used for more than  one filtration.  After each filtration, the
 .   membrane filter should be removed  and discarded,  and the
    filter holder should be either rinsed with metal-free water or
    dilute acid and dipped in a metal-free acid  bath  or rinsed at
    least  twice with metal-free dilute acid;  finally,, the filter
    holder should be rinsed at  least twice with  metal-free water
 d.  For each sample to  be filtered, the filter holder and membrane
    filter should be conditioned  with  the .sample, i.e.,  an initial
    portion of  the  sample should  be filtered and discarded.

                               105              '-,":'•

-------
 The accuracy_and precision of the dissolved measurement should be
 assessed periodically.  A large volume of a buffered solution
 (such as aerated_0.05 N sodium bicarbonate for analyses in fresh
 water and a combination of sodium bicarbonate and sodiumjchloride
 for analyses in salt water)  should be spiked so that the
 concentration of the metal of interest is in the range of the low
 concentrations that are to be measured.  Sufficient samples
 should be taken alternately for (a)  acidification in the same way
 as after_filtration in the dissolved method and (b) filtration
 and acidification using the procedures specified in the dissolved
 method until ten samples have been processed in each way.  The
 concentration of metal in each of~ the twenty samples should then
 be determined using the same analytical procedure.   The means of
 the two groups of ten measurements should be within 10 percent,
 and the coefficient of variation for each group of  ten should be
 less than 20 percent.   Any values deleted as outliers should be
 acknowledged.
Reporting results   •                                             ,

To  indicate  the  quality of  the  data/  reports of results of
measurements of  the concentrations of metals should include a
description  of the  blanks,  .spikes, CRMs, replicates, and
standards  that were run,  the number run, and the results
obtained.  All values deleted as outliers should be acknowledged.
Additional information                     »      •      ,

The items presented above'are some of the important aspects of
"clean techniques11; some aspects of quality.assurance and quality
control are also presented.  This is not -a definitive treatment
of these topics; additional information that might be useful is
available in such publications as Patterson and Settle  (1976)
Zief and Mitchell  (1976), Bruland et al. (1979), Moody and Beary
(1982), Moody  (1982), Bruland (1983), Adeloju and Bond  (1985),
Berman and Yeats (1985), Byrd and Andreae (1986), Taylor (1987),
Sakamoto-Arnold (1987), Tramontane et al. (1987), Puls and
Barcelona (1989),  Windom et al. (1991), U.S. EPA (1992), Horowitz
et al. (1992), and Nriagu et al. (1993).
                               106

-------
 References                           '•'..

 Adeloju,  S.B.'v  and A.M.  Bond.   1985.   Influence of Laboratory
 Environment  on  the Precision and Accuracy of Trace Element
 Analysis.  Anal.  Chem.  57:1728-1733.

 Berman,  S.S., and P.A.  Yeats.   1985.   Sampling of Seawater for
 Trace Metals.   CRC Reviews  in  Analytical Chemistry 16:1-14.

 Bruland,  K.W.,  R.P. Franks,  G.A.  Knauer,  and J.H. Martin.   1979.
 Sampling and Analytical  Methods for the Determination of Copper,
 Cadmium-,  Zinc,  and Nickel at the Nanogram per Liter Level  in Sea
 Water.  ,Anal. Chim. Acta 105:233-245.

 Bruland,.K.W.   1983.  Trace Elements.in Sea-water.   In:  Chemical
 Oceanography, Vol'.  8.   (J.P. Riley and R.  Chester,  eds.)
 Academic Press, .New York, NY.   pp.  157-220.

 Byrd, J.T.,  and M.O. Andreae.   1986.   Dissolved and Particulate
 Tin  in North Atlantic Seawater.   Marine Chem.  19:193-200.

 Horowitz, A.J., K.A. Elrick, and M.R.  Colberg.   1992.  The Effect
 of Membrane  Filtration Artifacts  on Dissolved Trace Element
 Concentrations.   Water Res.  26:753-763."

 Moody, J..R.  ,1982.  NBS  Clean  Laboratories for Trace Element
 Analysis.. Anal.  Chem.  54:1358A-1376A.
   •:     ,   '    '  "'    '*'''*•..   '- -     ..'•:,•'        .  "'   '  '/
 Moody, J.R., and  E.S. Beary.   1982.  Purified Reagents for Trace
 Metal Analysis.   Talanta 29:1003-1010.

 Nriagu, J.O.,, G.  Lawson, H.K.T. Wong,  and  J.M.  Azcue.  1993.   A
 Protocol  for Minimizing  Contamination  in the Analysis of Trace
 Metals in Great Lakes Waters.   J.  Great Lakes  Res.  19:175-182.

 Patterson,.C.C.,  and D.M. Settle.   1976.   The Reduction  in Orders
 of Magnitude Errors in Lead Analysis of Biological  Materials and
 Natural Waters by Evaluating and  Controlling the Extent  and
 Sources of Industrial Lead  Contamination Introduced during Sample
 Collection and  Processing.   In: Accuracy in  Trace Analysis:
 Sampling, Sample  Handling, Analysis.   (P.D.  LaFleur,  ed.)
 National Bureau of  Standards Spec.  Publ. 422^  U.S.  Government  ,
 Printing Office,  Washington, DC.

 Prothro, M.G.   1993.  Memorandum  titled "Office of  Water Policy
 and Technical Guidance on Interpretation and Implementation  of
 Aquatic Life Metals Criteria".  October 1.

 Puls, R.W.,  and M.J. Barcelona.,  1989.  Ground Water Sampling for
Metals Analyses.  EPA/540/4-89/001.  National  Technical
 Information  Service, Springfield, VA.


         .    •                  107

-------
 Sakamoto-Arnold, C.M-, A.K. Hanson, Jr., D.L. Huizenga, and D R
 Kester.  1987.  Spatial and Temporal Variability of Cadmium in
 Gulf Stream Warm-core Rings and Associated Waters.  J. Mar  Res
 45:201-230.          .
                      •           '                         «

 Shiller, A.M., and E. Boyle.  1985. - Dissolved Zinc in Rivers
 Nature 317:49-52.

 Shiller, A.M., and E.A. Boyle.  1987.  Variability of Dissolved
 Trace Metals in the Mississippi River.  Geochim. Cosmochiiru Acta
 51:3273-3277. ,                                       •

 Taylor, J.K.  1987.  Quality Assurance of Chemical Measurements
 Lewis Publishers, Chelsea, MI.

 Tramontane, J.M., J.R.  Scudlark,  and T.M. Church.   1987.   A
 Method for the Collection, Handling, and Analysis  of Trace Metals
 in Precipitation.  Environ. Sci.  Technol. 21;749-753.

 Trefry, J.H., T.A.  Nelsen, R.P.  Trocine,  S.  Metz.,  and T.W.
 Vetter.  1986.  Trace Metal Fluxes through .the Mississippi River
 Delta System.  Rapp.  P.-v. Reun.  Cons.  int.  Explor.  Mer.  186:277-
 288.

 U.S.  EPA.   1983.  Methods  for Chemical Analysis of  Water  and
 Wastes. EPA-600/4-79-020.  National Technical Informal-ion
 Service, Springfield, VA.   Sections 4.1.1, 4.1.3, and  4.1.4

 U.S.  EPA.   1991.  Methods  for the Determination of  Metals  in
 Environmental Samples.   EPA-600/4-91-010.  ', National  Technical
 Information Service,  Springfield,  VA.

 U.S.  EPA.   1992.  Evaluation of Trace-Metal  Levels  in  Ambient
 Waters  and  Tributaries  to  New York/New Jersey Harbor for Waste
 Load  Allocation.  Prepared by Battelle  ,Ocean -Sciences  under
 Contract No.  68-C8-0105.

 Windom, H.L.,  J.T. ,Byrd, R.G.  Smith, and  F.  Huan.   1991.
 Inadequacy  of NASQAN Data  for Assessing Metals Trends  in the
 Nation's Rivers.  •Environ.  Sci. Technol.  25:1137-1142.  (Also see
 the comment and response:  Environ.  Sci. Technol. 25:1940-1941.)

 Zief, M., and J.W. Mitchell.   1976.  Contamination Control in
 Trace Element  Analysis.  Chemical Analysis Series, Vol. 47
Wiley, New  York, NY.
                               108

-------
 Appendix D: Relationships between WERs and the Chemistry and
             Toxicology of Metals

.       '     <'      '     :  . '   ' '. •  '    '   '    .    •'"'     . '  •    ."
 The aquatic toxicology of metals is complex in part because the
 chemistry of. metals in water is complex.   Metals usually exist in
 surface water in various combinations of  particulate and
 dissolved forms,  some of which are toxic  and some of which are
 nontoxic.  In addition,  all toxic forms of a metal are not
 necessarily equally toxic,  and various water quality
 characteristics can affect the relative concentrations and/or
 toxicities of some  of the forms.             -

 The toxicity of a metal .has sometimes been reported to be
 proportional to the concentration or activity of a specific
 species of the metal.  For example,  Allen .and Hansen (1993)
 summarized reports,  by several investigators that the toxicity of
 copper is related to the free cupric ion,  but other data do not
 support a correlation (Erickson 1993a).   For example,  Borgmann
 (1983),  Chapman and McCrady' (1977),  arid French and Hunt (1986)
 found that toxicity expressed on the. basis  of cupric ion activity
 varied greatly with pH,  and Cowan et al.  (1986)  concluded that at
 least>one of the copper hydroxide species  is toxic.   Further,
 chloride and sulfate salts  of calcium,  magnesium,  potassium,  and
 sodium affect the toxicity of the cupric  ion (Nelson et al.
 1986).   Similarly for aluminum, Wilkinson  etal.  (1993)  concluded,
 that  "mortality was best predicted not  by  the free' A13+ activity
 but rather as a function of the sum Z( [A13+]  + [A1F2+]) " and that
 "no longer can the  reduction of Al toxicity in the presence of
 organic acids be interpreted simply as  a consequence of the
 decrease in the free A13+ concentration".        -

 Until a model has been demonstrated to  explain the quantitative
 relationship  between .chemical and toxicological measurements,
 aquatic  life  criteria should be established in an environmentally
 conservative  manner with provision for  site-specific adjustment.
 Criteria should be  expressed in terms of feasible analytical
measurements  that provide the necessary conservatism without
 substantially increasing the cost of implementation and site-
 specific adjustment.   Thus  current aquatic  life criteria for
metals are  expressed in  terms of  the total  recoverable
measurement and/or  the dissolved  measurement,  rather than a
measurement that  would be more  difficult to perform and would
 still.require empirical  adjustment.   The WER is operationally
 defined  in  terms  of chemical and  toxicological measurements to
allow site-specific adjustments that account for  differences
between  the toxicity of  a metal in laboratory dilution water and
 in site  water.'
                               109

-------
  Forms  of Metals
 Even if  the relationship of  toxici'ty to the -forms of metals is
 not understood well enough to allow setting site-specific water
 quality  criteria without using empirical adjustments, appropriate
 use and  interpretation of WERs-requires an understanding of how
 changes  in the relative concentrations of different forms of a
 metal might affect toxicity.  Because WERs are defined on the
 basis of relationships between measurements of toxicity and
 measurements of total recoverable and/or dissolved metal  the
 toxicologically relevant distinction is between the forms of the
 metal that are toxic and nontoxic whereas the chemically relevant
 distinction is between the forms that are dissolved and
 particulate.  "Dissolved metal" is defined here as "metal that
 passes through either a 0.45-fim or a 0.40-pm membrane, filter"  and
 -particulate metal" is defined as "total recoverable metal minus
 dissolved metal".  Metal that is in or on particles that pass
 tnrough the filter is operationally defined as ."dissolved".

 In addition,  some species of metal can be converted from one form
 to another.  Some conversions are the result of reequilibration
 in response to changes in water quality characteristics  whereas
 others are due to such fate processes as oxidation of sulfides
 and/or organic matter.  Reequilibration usually occurs faster
 than fate processes and probably results in any rapid changes
 that are duetto effluent mixing with receiving water or  changes
 in pH at a gill surface.   To account for rapid changes due  to
 reequilibration,  the terms "labile"  and "refractory"  will be used
 herein to denote metal species  that  do and do  not  readily convert
 to other species when in a nonequilibrium condition,  with
 ?readily"referring to substantial progression toward equilibrium
 in less than about  an hour.   Although the  toxicity and lability
 of a form of a metal  are not  merely yes/no properties, but  rather
 a?vo*ve gradations, a simPle  classification scheme such  as  this
 rs,2£uld  be sufficient  to  establish, the principles regarding  how
 WEKs are  related to various operationally  defined  forms  of  metal
 and how this affects  the  determination  and use of  WERs.
                                          i (     .                •  -
 Figure Dl presents  the classification scheme that  results from
 distinguishing forms of -metal based on  analytical  methodology,
 toxicity  tests, and lability, as described  above.  Metal that  is
 not measured,by the total recoverable measurement  is assumed to
 be sufficiently nontoxic and  refractory that it will not be
 further considered  here.  Allowance is made for toxicity due to
 particulate metal because some data indicate that particulate
metal might_contribute to toxicity and bioaccumulation, although
 other data imply that little  or no toxicity can be ascribed to
particulate metal (Erickson 1993b).  Even if the toxicity of
particulate metal is not negligible in a particular situation,  a
 dissolved criterion will not  be underprotective if the dissolved
 criterion was derived using a dissolved WER  (see below) or if
there are sufficient compensating factors.
                               110

-------
Figure Dl: A Scheme for Classifying Forms  of Metal in Water
     Total recoverable metal
          Dissolved                  -•    ;•
               Nontoxic
                    Labile
                    Refractory
      .'  ,    . ', Toxic' • '.-••  ;          -"..."....;  .•..'"  ..
                    Labile '
          Particulate                   ,
               Nontoxic
                    Labile
                    Refractory     -               ,
      ''.  '    .• Toxic                    . ' '  -
                    Labile       •       ,
     Metal.not measured by the total recoverable measurement
Not only can some changes in water quality characteristics shift
the relative concentrations of toxic and nontoxic labile species
of a metal, some changes in water quality can also increase or
decrease the toxicities of.the toxic species of a metal and/or
the sensitivities of aquatic organisms.  Such changes might be
caused by  (a), a change in ionic' strength that affects the
activity of toxic species of the metal in water,"  (b) a
physiological .effect whereby an ion affects the permeability of a
membrane and thereby alters both uptake and apparent toxicity,
and (c) toxicological additivity, synergism, or antagonism due to
effects within the organism.

Another-possible complication is that a.form of metal that is
toxic to one aquatic organism might not be toxic to another.
Although such differences between organisms have not been
demonstrated, the possibility cannot be ruled out.
The Importance of Lability

The only common metal measurement that can be validly
extrapolated from the effluent and the upstream water to the
downstream water merely by taking dilution into account is the
total recoverable measurement.  A major reason this measurement .
is so useful is because it is the only measurement that obeys the
law of mass balance  (i.e., it is the only measurement that is
conservative) .  Other met;al measurements usually do not obey the
law of mass balance because they measure some, but not all, of
the labile species of metals.  A measurement of refractory metal

-------
  would be conservative in terms  of  changes  in water  quality
  characteristics/  but  not necessarily  in  regards  to  fate
  processes;  such a measurement has  not been developed, however.

  Permit limits apply to effluents,  whereas  water  quality criteria
  apply to surface  waters.  If permit -'limits and water quality
  criteria are both expressed in  terms  of  total recoverable metal
  extrapolations  from effluent to surface  water only need to take
  dilution into'account  and can be performed as mass balance
  calculations.   If either  permit limits or  water  quality criteria
  ?r. .?^h are expressed  in  terms  of  any other metal measurement,
  lability needs  to be taken into account, even if both are
  expressed in terms of  the same measurement.             '

  Extrapolations  concerning labile species of metals from effluent
  to surface water  depend to a large extent on the differences
  between  the water quality characteristics of the effluent and
  those of  the surface water.   Although equilibrium models of the
  speciation of metals can provide insight, the interactions are
  too complex to be able to make useful nonempirical extrapolations
  from-a wide variety of effluents to a wide variety of surface
 waters of either  (a)  the speciation of the metal or (b)  a metal
 measurement other than total recoverable.

 Empirical extrapolations can be performed fairly easily  and the
 most common case will  probably occur when permit limits  are based
 on the total recoverable measurement but water  quality criteria
 are based on the dissolved measurement.   The  empirical
 extrapolation is intended to answer the  question "What percent of
 the total recoverable  metal  in the  effluent becomes  dissolved in
 the downstream water?".  This question can be  answered by
 a.  Collecting samples  of effluent and upstream water
 b.  Measuring total recoverable metal and dissolved metal  in both
    samples.
 c.  Combining aliquots  of the .two .samples  in the  ratio  of  the
    flows when the  samples were obtained  and mixing for an
    appropriate period  of time  under appropriate  conditions.
 d.  Measuring total recoverable metal and  dissolved metal  in the
    mixture.             .
 An  example is presented in Figure D2.  This percentage cannot  be
 extrapolated from  one metal  to another or from one effluent to
 another.   The data needed to calculate the  percentage will be
 obtained each time a WER is  determined using simulated downstream
water  if both dissolved and  total recoverable metal are measured
 in  the effluent, upstream water, and simulated downstream water.

The interpretation of the  percentage is not necessarily as
straightforward  as might be assumed.   For example, some of the
metal  that is dissolved in the upstream water might sorb onto
particulate matter in the  effluent,  which can be viewed as a
detoxification of  the upstream water by the effluent.  Regardless
of the interpretation,   the described procedure provides a simple
                               112

-------
 way of relating the total recoverable concentration in the
 effluent to the concentration of concern in the downstream water.
 Because this empirical extrapolation can be -used with any
 analytical measurement that is chosen as the basis for expression
 of aquatic life criteria, use of the total recoverable
 measurement to express permit limits-~on effluents does not place
 any restrictions on which analytical measurement can be used to
 express criteria.  Further,?even if both criteria and permit
 limits are expressed in terms of a measurement such as dissolved
 metal, an empirical extrapolation would still be necessary
 because dissolved metal is not likely to be conservative from
 effluent to downstream water.  . • .
 Merits of Total Recoverable and Dissolved WERs and Criteria

 A WER is operationally defined as the value of an. endpoint
 obtained with a toxicity test using site water divided by the
 value of the same endpoint obtained with the same toxicity test
' using a laboratory dilution water.  Therefore, just as aquatic
 life criteria can be expressed in terms of either the total
 recoverable measurement or the dissolved measurement, so can
 WERs.  A pair of side-by-side toxicity tests can produce both a
 total recoverable WER and a dissolved WER if the metal in the
 test solutions in both of the tests is measured using both
 methods. , A total recoverable WER is obtained by dividing
 endpoints that were calculated on the basis of total recoverable
 metal, whereas a dissolved WER is obtained by dividing endpoints.
 that were calculated on, the basis of dissolved metal.  Because of
 the way they are determined, a total recoverable WER is used to
 calculate a total recoverable site-spepific criterion from a
 national, state, or recalculated aquatic life criterion that is
 expressed using the total recoverable measurement, whereas a
 dissolved WER is used to calculate a...dissolved ..site^-specific
 criterion from a national, state, or recalculated criterion that
.is expressed in terms of the dissolved measurement.              '

 In terms of the classification"scheme given in Figure Dl, the
 basic relationship between a total recoverable national water
 quality criterion and a total recoverable WER is:
 • A total recoverable criterion treats all the toxic and
       nontoxic metal in the site water as if its average
       toxicity were the same as the average toxicity of all
       the toxic and nontoxic metal in,the toxicity tests in
       laboratory dilution'water on which the criterion is
   • '   based."  '    ,-    •"''    .       -,.'"'"..  "   - -' -       '  •.-
 • A total recoverable WER is a measurement of the actual
       ratio of the average toxicities of the total
       recoverable metal and replaces the assumption that
       the ratio is 1.                        ,
                                113

-------
 Similarly, the basic relationship between a dissolved national
 criterion and a dissolved WER is:                   .
 • A dissolved criterion treats all the toxic and nontoxic
       dissolved metal in the site water as if its average.
       toxicity were the same as the average toxicity of all
       theAtoxic and nontoxic dissolved metal in the
       toxicity tests in laboratory dilution water on which
       the criterion is based,
 • A dissolved .WER is a measurement of the actual ratio of
       the average toxicities of the dissolved metal and
       replaces the assumption that the ratio is 1.
 In both cases/ use of a criterion without a WER involves
 measurement of toxicity in laboratory dilution water but only
 prediction of toxicity,in site water,  whereas use of a criterion
 with a WER involves measurement of toxicity in both laboratory
 dilution water and site water.

.When WERs are used to derive site-specific criteria,  the total
 recoverable and dissolved approaches are inherently consistent.
 They are consistent because the toxic effects caused by the metal
 xn the toxicity tests do not depend on what chemical measurements
 are performed; the same number of organisms are killed in the
 acute lethality tests regardless of what,  if any,  measurements of
 the concentration of the metal  are made.   The only difference is
 the chemical measurement to which the toxicity is  referenced
 Dissolved WERs can be derived from the same pairs  of toxicity
 tests  from which total recoverable WERs are derived,  if the metal
 in the tests is  measured using  both the total recoverable and
 dissolved measurements.   Both approaches  start at  the  same  place
 (x.e.,  the  amount of toxicity observed in laboratory dilution
 water)  and  end at the same  place (i.e.,  the amount of  toxicity
 observed in site water).  The combination of a total recoverable
 criterion and WER accomplish the same  thing as the combination of
 a  dissolved criterion and WER.   By extension,  whenever, a
 criterion and a  WER based on the .same measurement  of the metal
 are used together,  they will end up at  the same place.   Because
 use of  a total recoverable  criterion with a  total  recoverable WER
 ends up at  exactly the same place as use  of  a  dissolved criterion
with a  dissolved WER.  whenever  one WER  is  determined, both  should
be determined to .allow (a)  a- check on the  analytical chemistry,
 (b) use  of  the inherent internal  consistency-to check that  the
data are used correctly,  and (c)  the option  of  using either
approach in the•derivation  of permit limits.

An examination of how  the two approaches  (the  total recoverable
approach and  the dissolved  approach) address the four relevant
forms of metal (toxic  and nontoxic particulate metal and toxic
and nontoxic .dissolved metal) in  laboratory  dilution water  and in
site_water further explains why .the  two approaches are  inherently
consistent.  Here, only the way in which the two approaches
address  each of the four forms of metal in site water will  be
considered:
                               114

-------
 a. Toxic dissolved metal: .
      .This form contributes  to the tdxicity of the site water and
       is measured by both chemical measurements.   If this is the
     ,  only form of metal present,  the two WERs will be the same,
 b. Nontoxic dissolved metal:'
       This form does not contribute to the toxicity of the site
       water,  but it is measured by both chemical  measurements.
       If this is the only form of  metal present,  the two WERs
       will be the same.  '(Nontoxic dissolved metal can be the
       only form present, however,  only if all of  the nontoxic
       dissolved metal present  is refractory.  If  any labile
       nontoxic dissolved metal is  present,  equilibrium will
       require that some toxic  dissolved metal also be present.)
 c. Toxic particulate metal:.
       This form contributes  to the toxicological  measurement in
       both approaches; it is measured by the total recoverable
       measurement, but not by  the  dissolved measurement.   Even
/      though it is not measured by the dissolved  measurement,  its
       presence is accounted  for in the dissolved  approach because
       it increases the toxicity of the site water and thereby
       decreases the dissolved  WER.  It is accounted for because
      .it makes the dissolved metal appear to be more .toxic than
       it is.   Most toxic particulate metal  is probably not toxic
       when it is particulate;  it becomes toxic when it is
       dissolved at the gill  surface or in the digestive system;,
       in the surface water,  however,  it is  measured as
       particulate metal.
 d. Nontoxic particulate metal:                              .
       This form does not contribute to the  toxicity of the site
      water;  it is measured  by the total recoverable measurement,
   • .  but not .by the dissolved measurement.   Because it is
      measured by the total  recoverable measurement,  but  not by
       the dissolved measurement, it causes  the total recoverable
      WER to be higher than  .the dissolved WER.                   ' :.
 In addition to dealing with  the four forms  of metal similarly,
 the WERs used in the -two approaches comparably take synergism,
 antagonism,  and additivity into account.   Synergism and   '.  ,  '
 additivity in the site water increase its toxicity and therefore
 decrease the  WER;  in contrast, antagonism in the  site water
 decreases toxicity and increases the WER.

 Each of  the four forms of metal is appropriately  taken into
 account  because use of the WERs makes the two approaches
 internally consistent.   In addition,  although experimentalx
 variation will cause the measured  WERs to deviate from the actual
 WERs, the measured WERs will be internally  consistent with the
 data from which they were generated.   If the percent dissolved is
 the same at the test endpoint  in the two waters,  the two  WERs
 will be  the same.   If the percent  of the total recoverable metal
 that is  dissolved in laboratory dilution water is less than 100
 percent,  changing from the total recoverable measurement  to the
 dissolved measurement will lower the criterion but it will

                    .'• .   .  .     us    .'••.•'•'.-••

-------
  comparably lower the  denominator in  the WER, thus increasing the
  WER.   If  the percent  of  the  total .recoverable metal that is
  dissolved in the site water  is  less  than 100 percent, changing
  from the  total recoverable measurement to the dissolved  .
  measurement will lower the concentration in the site water that
  is to  be  compared with the criterion, but it also lowers the
  numerator in the WER, thus lowering  the WER.  Thus when WERs are
  used to adjust criteria, the total recoverable approach and the
  dissolved approach result in the  same interpretations of
  concentrations in the site water  (see Figure D3) and in the same
 maximum acceptable concentrations in effluents  (see Figure D4).

  Thus,  if WERs are based on toxicity  tests whose endpoints equal
  the CMC or CCC and if both approaches are used correctly, the two
 measurements will produce the same results because each WER is
 based  on measurements on the site water and then the WER is used
  to calculate the  site-specific criterion that applies to the site
 water when the same chemical measurement is used to express the
  site-specific criterion.  The equivalency of the two approaches
 applies if  they are based on the same sample of site water.   When
 they are applied to multiple samples, the approaches can differ
 depending  on how the results from replicate samples  are used:
 a. If an appropriate averaging process is used,  the  two will"be
    equivalent.            '
 b. If the lowest value is used,  the two approaches will probably
    be equivalent only if the lowest dissolved WER and the.lowest
    total recoverable WER were obtained using the same sample  of
    site water.                            ,

 There are several advantages to using a dissolved criterion even
 when a dissolved WER is  not  used.'  In some  situations  use of  a
 dissolved criterion to interpret results  of  measurements of the
 concentration of dissolved metal in site  water might demonstrate
 that there is no need to determine either a  total  recoverable WER
 or a dissolved WER.   This would occur when  so much of  the total
 recoverable metal was  nontoxic particulate metal that  even though
 the total  recoverable  criterion  was  exceeded, the  corresponding
 dissolved  criterion was  not  exceeded.-  The.particulate metal
 might  come from  an effluent,  a resuspension .event, or  runoff that
.washed particulates into the  body of  water.  In  such a situation
 the total  recoverable  WER would  also  show that the site-specific
 criterion  was not exceeded, but  there would be no need to
 determine  a WER  if the criterion were expressed  on the basis of  ,
 the dissolved measurement.  If the variation over time in the
 concentration of  particulate metal is much greater than  the
 variation  in the  concentration of dissolved metal, both  the total
 recoverable concentration and the total recoverable WER are  •
 likely  to  vary so much over time that a dissolved criterion would
 be much more useful than a total recoverable criterion.
                               116

-------
 Use of a dissolved criterion without a dissolved ,WER has three
 disadvantages,  hpwever:
 1.. Nontoxic dissolved metal in the site water is treated as if it
    is toxic.
 2.  Any toxicity due to particulate metal in the site water is
    ignored.     ,                    -"
 3.  Synergism,  antagonism,  and additivity in the site water are
    not taken into account.
 Use of a dissolved criterion with a dissolved WER overcomes all
 three problems.   For example,  if (a)  the total recoverable
 concentration greatly exceeds the total recoverable criterion,
 (b)  the dissolved concentration is below the dissolved criterion,
 and (c)  there is concern about the possibility; of toxicity of
 particulate metal,  the determination of a dissolved WER would
 demonstrate whether toxicity due to particulate metal is
 measurable.

 Similarly,  use of a total  recoverable criterion without a total
 recoverable WER has three  comparable disadvantages:
 1.  Nontoxic dissolved metal in site water is treated as if it is
    toxic-. • .  • -  -    '  •''.-:".  .  :••••'•""   • ,  .   •    '  .- ' -        '
 2.  Nontoxic particulate metal in site water is treated as if it
    is toxic.                                 .
 3.  Synergism,  antagonism,  and additivity in site water are not
    taken into account.
 Use of a total  recoverable criterion with a total recoverable WER
 overcomes  all  three problems.   For example,  determination of a
 total recoverable WER would prevent nontoxic particulate metal
 (as  well as nontoxic dissolved metal)  in the site water from
"being treated as if it is .toxic.
Relationships between WERs  and the  Forms  of Metals

Probably the best way to understand what  WERs  can and cannot  do
is to understand the relationships  between WERs and the  forms of
metals i,  A WER  is calculated by dividing  the concentration  of a
metal that corresponds  to a toxicity  endpoint  in  a  site  water by
the ^ concentration of the same  metal that  corresponds to  the same
toxicity endpoint in a  laboratory dilution water.   Therefore,
using the classification scheme given in  Figure Dl:
The subscripts  "S"  and "L" denote site water and laboratory
dilution water, respectively, and:

S   = the _ concentration of Refractory metal  in a water.   (By
      definition, all  refractory metal  is nontoxic metal.)
                               117'

-------
 N   3 the concentration of Nontoxic labile metal in a water.

 T   = the concentration of Toxic labile metal in a water.

 A2\T  = the concentration of metal added during a WER determination
    •  that is Nontoxic labile metal-'after it is added.

 AT  * the concentration of metal added during a WER determination
       that is Toxic labile metal after it is added.

 For a total recoverable WER,  each of these five concentrations
 includes both particulate and dissolved metal,  if both are
 present; for a dissolved WER only dissolved metal is included.


 Because the two side-by-side tests use the same endpoint and are
 conducted under identical,conditions  with comparable test
 organisms,  TS + *TS  ^  TL + *TL when the.toxic species of the metal
 are equally toxic  in the two waters.   If a difference in water
 quality causes one or more of the toxic species of the metal to
 be more toxic in one water than the other,  or causes a shift in
 the ratios of various toxic species,  we can define
                             _ TB
Thus H is  a multiplier that accounts for a proportional increase
or decrease in the toxicity of  thfe  toxic forms  in  site  water  as
compared to their toxicities in laboratory^ dilution water.
Therefore,  the- general WER  equation is:
                            NS
                         RL + NL + *NL + (TL

Several things are obvious  from this equation:
1. A WER should not be thought  of  as a simple ratio such as H.
   H is the ratio of  the toxicities of the toxic  species of the
   metal, whereas the WER is the ratio of  the sum of the toxic
   and the nontoxic species of  the metal.   Only under a very
   specific set of conditions will WER = H.  If these conditions
   are satisfied and if,  in addition, #=i, then  WER ~ l.
   Although it might seem that  all of these conditions will
   rarely be satisfied,  it is not  all that rare to find that an
   experimentally determined WER is close  to 1.
2. When the concentration of metal in laboratory  dilution-water
   is negligible, RL = NL = TL = 0 and

                    rrrm   Rs + Ns
                    WER = —£	?~
                               118

-------
    Even though laboratory dilution water is low in TOC and TSS,
    when metals are added to laboratory dilution water in toxicity
    tests,  ions such as hydroxide, carbonate; and chloride react
    with some metals to form some particulate species and .some
   •dissolved species, both of which might be toxic or nontoxic.
    The metal species that are nontoxic contribute to &NL,  whereas
    those that are toxic contribute to *TL.   Hydroxide,  carbonate,
    chloride,. TOC, and TSS can increase *NS.   Anything that, causes
    Aflk to  differ  from &NL will cause  the WER to differ  from 1.
 3.  Refractory metal and nontoxic labile metal in the site water
    above that in  the laboratory dilution water will increase the
    WER.  Therefore, if the WER is determined in downstream water,
    rather than in upstream water, the ,WER will be increased by
    refractory metal and nontoxic labile metal in the effluent.
 Thus( there are three major reasons why WERs  might be larger or
 smaller than 1:                    •
 a.  The toxic species of the metal might be more toxic in one
    water than in  the'other,  i.e., tf*l.
 b.  Atf might be higher in one water than in the other.
 c.  R and/or N might be higher in one water than in the.other<

 The last reason"might have great practical importance in some
 situations.   When a WER is determined in downstream water,  if
 most of the metal in the effluent is  nontoxic,  the WER and the
 endpoint in site  water will  correlate with the concentration of
 metal in the site water.   In addition,  they  will  depend on the
 concentration of  metal in the effluent and the concentration of
 effluent in the site water.   This correlation will be best for
 refractory metal  because its toxicity cannot be affected by water
 Quality characteristics;  even if the  effluent and upstream water
 are  quite  different so that  the  water quality characteristics  of
 the  site water depend on the percent  effluent, the toxicity of
 the  refractory metal will  remain constant at zero and the  portion
 of the WER that is due to  refractory  metal will be additive.
The Dependence of WERs on the Sensitivity of Toxicitv Tests

It would be desirable if the magnitude of the WER for a site
water were independent of the toxicity test used in the -
determination of the WER, so that any convenient toxicity test
could be used!  It can be seen from the general WER equation that
the WER will be independent of the toxicity test only if :
which would require that Rs = NS = AWS = RL = NL - AW^ = o .   (It would
be easy to assume that TL = 0, but it can be misleading in some
situations to make more simplifications than are necessary.)
                               119

-------
This is the simplistic concept of a WER  that would be
advantageous if it were true, but which  is not  likely  to  be true
very often.  Any situation in which one  or more of the terms is
greater than zero can cause the WER to depend on the sensitivity
of the toxicity test, although the difference in the WERs might
be small.                  .         '"

Two situations that "might be common can  illustrate how the  WER
can depend on the sensitivity of the toxicity test.  For  these
illustrations ,. there is no advantage to  assuming that  ff=l, so.
H will be retained for generality.
1. The simplest situation is when Rs > o,  i.e.,  when a
   substantial concentration of refractory metal occurs in  the
   site water.  If, for simplification,  it is assumed  that
   Ns » &NS ** EL « NL = *NL * 0 , then:
                      .
   The quantity TL + &TL Obviously changes as the sensitivity of
   the toxicity test changes .  When Rs = o ,  then  WER = H and the.
   WER is independent of the sensitivity  of the  toxicity test .
   When Rs > 0 , then the WER will decrease as the sensitivity of
   the test decreases because  TL + *TL will increase.

2. More complicated situations occur when (Ns + AWS) > 0.  If, for
   simplification, it is assumed that Rs « RL '» NL = &NL .= 0,  then:

                 ' ur


 -  a. If  (Ns + Aflrs) > 0  because the site water contains a
      substantial concentration of a complexing  agent that has an
      affinity for the metal and if complexation converts  toxic
      metal- into nontoxic metal, the complexation reaction will
      control the toxicity of the solution  (Allen 199.3) .   A
      complexation curve can be. graphed in  several ways, but the
     . S-shaped curve presented in Figure  D5 is most  convenient
      here.   The vertical axis is "% uncomplexed" , which is
      assumed to correlate with "% toxic".  The  "% complexed"  is
      then the "% nontoxic " .  The ratio of  nontoxic  metal  to
      toxic metal is-:

                    ^nontoxic s  % complexed  ± v
                     % toxic     %uncoaplexed    '    .

      For the complexed nontoxic metal:                      ,  -   '

                v _ concentration of nontoxic metal
                     concentration of toxic metal


                               120

-------
 In the site water,  the concentration of complexed nontoxic
 metal  is (Ns + A#S) and the concentration of toxic metal  is
 (Ts + ATS) , so that: .                   •
                               (Ns.+
                   (2V+ Aiy    H(TL
    .„'  •  VJI(TL + AT,) + H(TL + AT,)
    WER =                        = tyr + J5T =
 If  the WER is  determined using^ a sensitive toxicity test so
 that  the  % uncomplexed (i.e.,  the % toxic)  is 10 %, then
 vs - (90 %)/{lO %) =9, whereas if  a  less sensitive test  is
 used  so that the %  uncomplexed is 50 %,  then
 Vs'- (50 %)/(50 %) = 1. , Therefore, if  a portion of the WER is
 due to a  complexing agent in the site water,  the magnitude
 of  the WER can decrease as the sensitivity of the toxicity
 test  decreases' because the % uncomplexed will decrease.   In
 these situations, the .largest  WER will be obtained with the
 most  sensitive toxicity test;  progressively smaller WERs
 will  be obtained with  less sensitive toxicity tests.   The
 magnitude of a WER  will depend not only  on > the sensitivity
 of  the toxicity test but also  on the concentration of  the
 complexing agent and on its binding constant  (complexation
 constant,  stability constant).   In addition,  the binding
 constants of most complexing agents  depend on pH.

 If  the laboratory dilution water contains a low
 concentration  of a  complexing  agent,
and       '   .                            .

                    + H(TL + ATi)  ' : VgH + H   H(VS
   WER
The binding constant of the complexing agent  in the
laboratory dilution water is probably different from that
of the complexing agent in the site water.  Although
changing from a' more sensitive test to a  less sensitive
test will decrease both vs and VL, the amount of effect is
not likely to be proportional.

If the change from a more sensitive test  to a less
sensitive test were to decrease  VL proportionately  more
than V^,  the change could result in a larger WER, rather

                         121

-------
       than a smaller WER, as resulted in the case above when It
       was assumed that the laboratory dilution water did not
       contain any complexing agent.  This is probably most likely
       to occur if H 5j 1 and if  vs < VL, which would mea.n £hat
       WER < l .  Although this is likely to be a rare situation,
       it does demonstrate again the" importance of determining   '
       WERs using toxicity tests that have endpoints in laboratory
       dilution water that are close to the CMC or CCC to which
       the WER is to be applied.
    b. If (NS + jtJfg) > 0  because the site water contains  a
       substantial concentration of an ion that will precipitate
       the metal of concern and if precipitation converts toxic
       metal into nontoxic metal, the precipitation reaction will
       control the toxicity of the solution.  The "precipitation
       curve" given in Figure D6 is analogous to the " complexation
       curve" given in Figure D5; in the precipitation curve, the
       vertical axis is "% dissolved0, which is assumed to
       correlate with "% toxic".  If the endpoint for a toxicity
       test is below the solubility limit of the precipitate,
       (Ns + AATS) « 0 ,  whereas  if  the  endpoint  for  a toxicity test
       is above the solubility limit,   (Ns + Aifc) >  0 .  if WERs  are
       determined with a series of toxicity tests that  have
       increasing endpoints that are above the solubility limit,
       the WER will reach a maximum value and then decrease.   The
       magnitude of the WER will depend not only on the
       sensitivity of the toxicity test but also on the
       concentration of the precipitating agent,  the solubility
      ..limit,  and the solubility of the precipitate.

Thus,  depending on the composition of the site water,  a WER
obtained with an insensitive test might be larger,  smaller/ or
similar  to  a WER obtained with a sensitive test.  Because of .the
range of possibilities that, exist,  the best toxicity test to use
in  the experimental determination of  a WER is one whose endpoint
in  laboratory  dilution water is close to the CMC or CCC that is
to  be adjusted.  This  is the rationale that was  used in the
selection of the toxicity tests that  are suggested  in  Appendix I .

The available  data indicate  that a  less sensitive toxicity  test
usually  gives  a  smaller WER  than a  more sensitive test  (Hansen
1993a) .   Thus, use of  toxicity  tests  whose endpoints are higher
than  the CMC or. CCC probably will not result in  underprotection;
in  contrast, use of tests whose endpoints  are substantially below
the CMC  or  CCC might result  in  underprotection.

The factors that cause  Rs  and (NS + *NS)  to -be greater than zero
are all  external to the test organisms;  they are chemical effects
that affect the metal in the water.   The magnitude  of  the WER is
therefore expected  to depend on the toxicity test used  only  in
regard to the sensitivity of the test.   If  the endpoints for two
                                   '                      !
                                122                  .,

-------
different tests occur at the same concentration of.the metal, the
magnitude of the WERs obtained with the, two tests should be the
same; they should not depend on  (a) the7'duration of the test,  (b)
whether, the endpoint is based on a lethal or sublethal effect, or
ic) whether the species is a vertebrate or an invertebrate.

Another interesting consequence of the chemistry of complexation
is that the ,% uncomplexed will increase if the solution is
diluted (Allen and Hansen 1993).  The concentration of total
metal will decrease with dilution but the % uncomplexed will
increase.  The increase will not offset the decrease and so the
concentration of uncomplexed metal will decrease.  Thus the
portion of a WER that is due to complexation will not be strictly
additive (see Appendix G), but the amount of nonadditivity might
be difficult to detect in toxicity studies of additivity.  A
similar effect of dilutipn will occur for precipitation.

The illustrations presented above were simplified to make it
easier to understand the kinds of effects that can occur.  The
illustrations are qualitatively valid and demonstrate the '
direction of the effects^ but real-world situations will probably
be so much more complicated that the various effects cannot be
dealt with separately.
Other Properties of WERs                                         :

1. Because of the variety of factors that can affect WERs, no
   rationale exists at present for extrapolating WERs from one
   metal to another, from one effluent to another, or from one
   surface water to another.  Thus WERs should be individually
   determined for each metal at each site.,

2. The most important informat ion ..that the .determination of a WER
   provides is whether simulated and/or actual downstream water
   adversely affects test organisms that are sensitive to the
   metal.  A WER cannot indicate-how much metal needs to be
,_  removed from or how much metal can be. added to an effluent.
   a. If the_site water already contains sufficient metal -that it
      is toxic to the test organisms, a WER cannot be determined
      with a sensitive test and so an insensitive test will "have
      to be used.  Even if a WER could be determined with a
      sensitive test, the WER cannot indicate how much metal has
    ', ,tp be removed..  For example,  if a WER indicated that there
    .. was 20 percent too much metal in an effluent, a 30 percent
      reduction by the discharger would not reduce toxicity if
      only nontoxic metal was removed.  The next WER
      determination would show that the effluent still contained
      too much metal.  Removing metal is useful only if the metal
      removed is toxic metal.  Reducing the total recoverable
      concentration does not necessarily reduce toxicity.

•   '  -    ".'',••       .    ;    123    ' .    ..           :' '   '.  ; '  ' '

-------
    b. If the simulated or actual downstream water is not toxic, a
       WER^can be determined and used to calculate how much
       additional' metal the effluent could contain and still be
       acceptable.  Because an unlimited amount of refractory
       metal can be added to the effluent without,affecting the
       organisms, what the WER actually determines is how much
       additional toxic metal pan.be added to the effluent.  ;

    The' effluent component of nearly all WERs is likely to be due
    mostly to either (a)  a reduction in toxicity of the metal by
    TSS or TOC,  or (b)  the presence of refractory metal.  For both
    of these, if the percentage of effluent in the downstream
    water decreases,  the magnitude of the ,WER will usually
    decrease.  If the water quality characteristics of the
    effluent and the upstream water are quite different, it is
    possible that the interaction will not be•additive;  this can '
    affect the portion of the WER that is due to reduced toxicity
    caused by sorption and/or binding,  but it cannot affect the
    portion  of the WER'that is due to refractory metal.
                     /
    Test organisms are fed during some toxicity tests, but not
    during others; it is  not clear whether a WER determined in  a
    fed test will differ  from a WER determined in an unfed test.
    Whether  there is  a difference is likely to depend on the
    metal, the type and amount of food,  and/whether a total
    recoverable  or dissolved WER is determined.   This can be
    evaluated by determining two WERs using a test in which the
    organisms usually are not fed - one WER with no food added  to
    the tests and one with food added to the tests.   Any effect of
    food is  probably  due  to an increase in TOC and/or TSS.   If
    food increases the  concentration of nontbxic metal in both  the
    laboratory dilution water and the site water,  the food will
    probably decrease the WER.  .Because complexes of metals are
    usually  soluble,  complexation is likely to lower both total
    recoverable  and dissolved WERs;. sorption to  solids will
    probably reduce only  total recoverable WERs.   The food might
    also  affect  the acute-chronic ratio.   Any feeding during a
    test  should  be limited to the minimum necessary.
Ranges of Actual Measured WERs.

The acceptable WERs found by Brungs et al.  (1992) were total
recoverable WERs that were determined in relatively clean fresh
water.  These WERs ranged from about 1 to 15 for both copper and
cadmium, whereas they ranged from about 0.7 to 3 for zinc.  The
few WERs that were available for chromium, lead, and nickel
ranged from about 1. to 6.  Both the total recoverable and
dissolved WERs for copper in New York harbor range from about 0.4
to 4 with most of the WERs being between 1 and 2 (Hansen 1993b).


                               124

-------
 Figure D2:  An Example of the Empirical Extrapolation Process


 Assume the  following hypothetical effluent  and upstream water:

 Effluent:                 -           '•-.-.
     Ts:      100 Ug/L
     Ds:       10 ug/L   (10 % dissolved)
     Q-: -     24 cfs
 Upstream water:
              40 ug/L
              38 ug/L   (95 % dissolved)
              48 cfs
7* •
J-V
D,,:
 Downstream water:        ,  :
    TD:       60  ug/L
    X>D:    .   36  ug/L   (60 %  dissolved)
    O,:       72  cfs
 where: .'.''.,'-     - -.   .'••-.'•       ..'"-."   '....-.    •'_        '.

 T  = concentration of  total recoverable metal.
 D  = concentration of  dissolved metal.
• 0  = flow.. '.   • .  '•"..- .:-  .'..,'...' •. • •'.   • • •  .'-•-.

 The subscripts E, U, and D  signify effluent,  upstream water, and
 downstream water, respectively.

 By conservation of flow:  QD = QE + Qy .

 By conservation of total recoverable metal: TjffD = TJ£E +• T&a .

 If P =  the percent of  the  total  recoverable metal in the '
         effluent .that becomes dissolved in. the downstream water,
 For the data given above, the percent  of the total recoverable
 metal in the effluent that becomes  dissolved in the downstream
 water is:

          P = 1001(36 ug/L) (72 cfs) - (38 ug/L) (48 cfs) ] _-,-«.
                        (100 ug/L) (24  cfs}  ~~~    ~ -- 32%,
      \     '      ;              -h     .       ''.'••,,-    = •    .   '
 which is greater than the 10 % dissolved in the effluent and less
 than the 60 % dissolved in the. downstream water.
                                125

-------
Figure D3: The Internal Consistency of the Two Approaches


The internal consistency of the total recoverable and dissolved
approaches can be illustrated by considering the use of WERs  to
interpret the total recoverable and-'dissolved concentrations  of a
metal in a site water.  .For this hypothetical example, it will be
assumed that the national CCCs for  the metal are:
      200 ug/L as total recoverable metal.
      160 ug/L as dissolved metal.
It will 'also- be assumed that the concentrations of the metal  in
the site water are:                                •     .      ' .
      300 ug/L as total recoverable metal.
      120 ug/L as dissolved metal.
The total recoverable concentration in the site water exceeds the
national CCC, but the dissolved concentration does not.


The-following results might be'obtained if WERs are determined:

   In Laboratory Dilution Water      .             ,
      Total recoverable LC50 = 400  ug/L.
         % of the total recoverable metal'that is dissolved = 80.
            (This is based on the ratio of the national CCCs,
            which were determined in laboratory dilution water.)
      "Dissolved LC50 = 320 ug/L.                 .      ,-•...

   In Site Water
      Total recoverable LC50 =620  ug/L."
         % of the total recoverable metal that is dissolved =40.
          (This is based on the data given above for site water).
      Dissolved LC50 = 248 ug/L.


      Total recoverable WER = (620 ug/L)/(400 ug/L)  =1.55
      Dissolved WER = (248 ug/L)/(320 ug/L)  = 0.775


   Checking the Calculations

    Total recoverable WER _  1.55  _ lab water % dissolved   80 _ 0
        Dissolved WER    ~ 0.775 ~ site water % dissolved ~ 40


   Site-specific CCCs (ssCCCs)

      Total recoverable ssCCC = (200 ug/L)(1.55)  =310 ug/L.
      Dissolved ssCCC = (160 ug/L)(0.775)  =124 ug/L.


   Both concentrations in site water are below the respective
   ssCCCs.

                               126

-------
In contrast,  the following results might have been obtained when
the WERs were determined:              .;•-..

   In Laboratory Dilution Water                          ;
      Total  recoverable LC50 = 400 ug/L.
         % of the total recoverable metal that is dissolved = 80.
      Dissolved LC50  =,320 ug/L.
                            - ' ' - -        „   -      ''•''•/
   In Site Water
      Total  recoverable LC50 = 580 ug/L.
         % of the total recoverable metal that is dissolved - 40.
      Dissolved LC50  = 232 ug/L.

   WERs      ,'    •''•.  ,  '    ;, • •',     ;'•'•'.'     .;"   •'.'•" '       '
      Total  recoverable WER = (580 ug/L) / (400 oag/L)  = 1.45
      Dissolved WER = (232 ug/L)/(320 ug/L) = 0.725


   Checking  the Calculations
            'v   '-   '    _,         •     .    '    "     ,        ':• '_ .
     Total recoverable  WER _  1.45 _ lab water % dissolved _ 80
         Dissolved WER     ~ 0.725 ~ site water % dissolved ~ 40


   Site-specific CCCs (ssCCCs)

      Total  recoverable ssCCC = (200 ug/L)(1.45)  = 290 ug/L.
      Dissolved ssCCC = (160 ug/L)(0.725)  =116 ug/L,


   In this case,  both concentrations in site water are above the
   respective ssCCCs.
In each case, both approaches  resulted in the same conclusion
concerning whether the  concentration in site water exceeds the
site-specific criterion.             .
 " "     -        .   .          '       .                   ' '

The two key assumptions are:
1. The ratio of total recoverable metal to dissolved metal in
   laboratory dilution  water when the WERs are determined equals
  • the ratio of the national CCCs.
2. The ratio of total recoverable metal to dissolved metal in
,   site water when the  WERs are determined equals the ratio of
   the concentrations reported in the site water.
Differences in the ratios  that are outside the range of
experimental variation  will cause problems for the derivation of
site-specific criteria  and, therefore, with the internal
consistency of the two  approaches.        ,       ,


, ' ;  '; .;;    ';  ','."    '  .  >ri  :' '•'•   127••' '"•   '. .' " - '" "  '•". ' • •   ': '

-------
 Figure D4r The Application of the Two Approaches


 Hypothetical upstream water and effluent will be used to .
 demonstrate the equivalence of the total recoverable and
 dissolved approaches.  The upstream''water and the effluent will
 be assumed to have specific properties in order to allow •
 calculation of the properties of the downstream water,  which will
 be assumed to be a 1:1 mixture of the upstream water and
 effluent.  It will also be assumed that the ratios of the forms
 of the metal in the upstream water and in the effluent  do not
 change when the total recoverable concentration changes.
 Upstream water  (Flow = 3 cfs)
    Total recoverable:              400 ug/L
       Refractory particulate:          200 ug/L
       Toxic dissolved:                200 ug/L  (50  % dissolved)
 Effluent  (Flow = 3  cfs)
    Total recoverable:               440  ug/L
       Refractory particulate:  ,       396  ug/L
       Labile nontoxic  particulate:      44  ug/L
       Toxic  dissolved:                  6  ug/L   (0  %  dissolved)
          (The labile nontoxic  particulate,  which is 10  % of  the
          total recoverable  in  the effluent, becomes toxic
          dissolved in  the downstream water.)
Downstream water   (Flow =  6  cfs)                    '
   Total recoverable:              420 ug/L
      Refractory particulate:         298 ug/L
      Toxic dissolved:                122 ug/L   (29 % dissolved)

   The values for the downstream water are calculated from the
   values  for the upstream water and the effluent:
      Total recoverable:       [3(400) + 3(440)]/6  = 420 ug/L
      Dissolved:               [3(200) + 3(44+0)]/6 = 122 ug/L
      Refractory particulate:  [3(200) -s- 3(396)]/6  = 298 ug/L
Assumed National CCC  InCCC)
   Total recoverable = 300 ug/L
   Dissolved =240 ug/L
                               128

-------
Upstream site-specific CCC (ussCCC)           ,.                  ,
      t'   , ! -         '   •                 -        -  ' - v   :      '
   Assume:  Dissolved cccWER = 1.2 /          •    ,
      Dissolved ussCCG = (1.2)(240 ug/L) =288 ug/L      .
   By calculation:  TR ussCCC = (288  ug/L)/(0.5) = 576 ug/L
      Total recoverable cccWER = (576~ug/L)/(300 ug/L) =1.92

                           nCCC     cccWER    ussCCC      Cone.
   Total recoverable:    300 ug/L    1.92    576 ug/L    400 ug/L
   Dissolved:    .- ,       240 ug/L    1.2     288 ug/L    200 ug/L
      .    %  dissolved        80 %	  ,      50 %        50 %
      Neither  concentration exceeds  its respective ussCCC.

     Total recoverable WER _ 1.92 _  lab water % dissolved   80
         Dissolved WER     ~ 1.2 "•" site water % dissolved ~ ~50~ = 1
Downstream site-specific  CCC  (dssCCC)

   Assume: Dissolved cccWER =1.8
      Dissolved dssCCC  =  (1.8)(240  ug/L)  = 432 ug/L
   By calculation: TR dssCCC  =
      {(432 ug/L-[(200  ug/L)/2])/0.1}+{(400 ug/L)/2} = 3520 ug/L
            This calculation  determines  the amount of dissolved
            metal contributed by the effluent, accounts for the
          .  fact that ten percent of the total recoverable metal
            in the effluent becomes dissolved, and adds the total
            recoverable metal contributed by the upstream flow.
	•"  "Total recoverable cccWER  = (3520 ug/L)/(300 ug/L) = 11.73

                      '.   nCCC    -  cccWER   - dssCCC      Cone.
   Total recoverable:   300 ug/L    11.73   3520 ug/L   420 ug/L
   Dissolved:           240 ug/L     1.80    432 ug/L   122 ug/L
        , % dissolved       80  %   ——         12.27 %    29  %
      Neither concentration exceeds its  respective dssCCC.

  Total  recoverable WER _ 11.73  _  lab water~% dissolved     BO  _ K  K,
      Dissolved WER       1.80  ~  site water % dissolved   12.27 ~
Calculating the Maximum Acceptable Concentration in the Effluent

   Because neither the total recoverable  concentration nor the
   dissolved concentration in the downstream water exceeds its
   respective site-specific CCC, the  concentration of metal in
   the effluent could be increased.   Under  the  assumption that
   the ratios of the two forms of the metal in  the effluent do
   not change when the total recoverable  concentration changes,
   the maximum acceptable concentration of  total recoverable
   metal in the effluent can be calculated  as follows:

•-.-';•    -    •'..  '      '  .   129      '  '   -•'•'.-•'

-------
    Starting with the total recoverable  dssCCC of 3520 ug/L

            (6 cfs) (3520 ug/L) - (3 cfs) (400 ug/L)  _
                      .     3        ~ - •" - -
    Starting with the dissolved dssCCC of 432 ug/L

          (6 cfs) (4.32 ug/L) - (3 cfs) (400 ug/L) (0.5)   --,'
                       (3          -   - ~ -- 664°
 Checking the Calculations

    Total recoverable :

        (3 cfs) (66AO ug/L) + (3 cfs) (400 ug/L) _ ,_0ft
         ;             e cfs      ~~~   ^~~ ~ 35ZO

    Dissolved:

        (3 cfs) (6640 ug/L) (0.10)  + (3 cfg) (400 ug/L) (0.50)   .^   /T
                            6 cfs      : - - "~ — ~~ = 4'32 uflr/I, .

    The  value of 0.10 is used  because this is the percent of the
    total recoverable metal  in the effluent that becomes dissolved
    xn the downstream water.

    The  values of 3520 ug/L  and 432 ug/L equal the downstream
    site-specific CCCs derived above.
Another Way to Calculate the Maximum Acceptable Concentration

   The maximum acceptable concentration of total recoverable
   metal in the effluent can also be calculated from the
   dissolved dssCCC of  432 ug/L using a partition coefficient to
   convert from the dissolved dssCCC of 432 ug/L to the total
   recoverable dssCCC of 3520 ug/L:
           16 cf£r]          ~ (3
                                               ^*.*~.
                                    — '• - '—• = 664°
   Note that the value used  for  the partition coefficient in this
   calculation is 0.1227  (the  one  that  applies to the downstream
   water when the total recoverable concentration of metal in the
   effluent is 6640 ug/L) , not 0.29  (the one that applies when
   the concentration of metal  in the  effluent is only 420 ug/L) .
   The three ways of calculating the  maximum acceptable
   concentration give the same result if each is used correctly.
                               130

-------
Figure D5: A Generalized Complexation Curve
The curve is for a constant concentration- of the complexing

ligand and an increasing concentration of the metal.
    TOO
 O
 in

 8
 8
          LOG  OF  CONCENTRATION OF  METAL
                          131

-------
Figure D6: A Generalized Precipitation Curve

                         •          ; " '     •    ' "   ,  I


The curve is for a constant concentration of the precipitatino

ligand and an increasing concentration of the metal.
    100
 Q
 IU
 O
 CO
 CO
 5
         LOG OF CONCENTRATION OF METAL
                         132

-------
 References .
         * •    •   -      -..    '•    ' •  . •        •',•',   -   '•
 Allen, H.E.  1993.  Importance of Metal Speciation to Toxicity.
 Proceedings of the Water Environment Federation Workshop on
 Aquatic Life Criteria for Metals.  Anaheim, CA.  pp. 55-62.

 Allen, H.E.,  and D.J. Hansen.  1993.   The Importance of Trace
 Metal; Speciation to Water Quality Criteria.  Paper presented at
 Society for Environmental Toxicology and Chemistry.  Houston, TX
 November 15.                 ,                                    '

 Borgmann,  U.   1983.  Metal Speciation and Toxicity of Free Metal
 Ions to Aquatic Biota.  IN: Aquatic Toxicology.  (JiO.  Nriagu
 ed.)   Wiley,  New York, NY..".

 Brungs,  W.A.,  T.S. Holderman,  and M.T.  Southerland.  1992.
 Synopsis of Water-Effect Ratios for Heavy Metals as Derived for
 Site-Specific Water Quality Criteria.   U.S. EPA Contract 68-CO-
 0070.

 Chapman,  G.A.,  and J.K.  McCrady.   1977.   Copper Toxicity:  A
 Question of Form.   In:. Recent Advances  in Fish Toxicology.   (R A
 Tubb,  ed.)  EPA-600/3-77-085  or  PB-273  500.   National Technical
 Information Service,  Springfield,  VA.   pp.  132-151.

 Erickson,  R.   1993a.   Memorandum to C.  Stephan.   July 14i

 Erickson,  R.   1993b.   Memorandum to C.  Stephan.   November  12.

 French,  P., and D.T.E.  Hunt.   1986. The  Effects of Inorganic
 Complexing upon the Toxicity of Copper  to Aquatic Organisms
 (Principally  Fish)..  IN:  Trace Metal Speciation and Toxicity to
 Aquatic  Organisms  - A Review.   (D.T.E.  Hunt, e'd.)   Report TR 247
 Water Research Centre,  United Kingdom.

 Hansen, D.J.   1993a.   Memorandum to G..E.  Stephan.   April 29.

 Hansen,.D.J.   1993b.  Memorandum to C.E.  Stephan.   October  6.

 Nelson, H., D.  Benoit, R.  Erickson, V. Mattson,  and J.  Lindberg.
 1986.  The  Effects  of Variable Hardness,  pH, Alkalinity,
 Suspended Clay,  and Humics on  the  Chemical  Speciation and Aquatic
 Toxicity of Copper.   PB86-171444.   National Technical Information
 Service, Springfield, VA.

Wilkinson, K.J., P.M. Bertsch,  C.H. Jagoe,  and P.G.C. Campbell.
 1993.  Surface  Complexation of Aluminum on  Isolated Fish Gill
Cells.  Environ. Sci. Technol,  27:1132-1138.
                               133

-------
 Appendix E:  U.S.  EPA Aquatic Life Criteria Documents for Metals
  Metal

 Aluminum
 Antimony
 Arsenic
 Beryllium
 Cadmium
   •
 Chromium
 Copper
 Lead
 Mercury
 Nickel
 Selenium
 Silver
 Thallium
 Zinc
  .-  EPA Number
 EPA 440/5
 EPA 440/5
 EPA 440/5
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
 EPA 440/5-
EPA 440/5-
EPA 440/5-
-86-008
-80-020
-84-033
-80-024
-84-032
-84-^029
-84-031
-84-027
•84-026
•86-004
87-006
80-071
80-074
87-003
 NTIS Numbeir

 PB88-245998
 PB81-117319
 PB85-227445
 PB81-117350
. PB85-227031
 PB85-227478
 PB85-227023
 PB85-227437
 PB85-227452
 PB87-105359
 PB88-142237
 PB81-117822
 PB81-117848
 PB87-153581
All are available from:
          National Technical ^Information  Service  (NTIS)
          5285 Port Royal Road
          Springfield, VA 22161
             TEL: 703-487-4650
                               134

-------
  Appendix F:  Considerations Concerning Multiple-Metal,  Multiple-
              Discharge,  and Special Flowing-Water Situations
  Multiple-Metal Situations
  Both Method 1  and Method 2  work well in multiple-metal
  situations,  although the amount of testing required increases as
  the number of  metals increases.  The major problem is the same
  for both methods:  even when addition of two or more metals
  individually is  acceptable,  simultaneous addition of the two or
  more metals, each at its respective maximum acceptable
  concentration, might be unacceptable for at least two reasons:
  1.  Additivity  or synergism  might occur between metals.
  2.  More than one of  the metals  might be detoxified by the same
     complexing  agent  in the  site water.   When WERs are determined
     individually,  each metal can utilize all of the complexing
     capacity; when the metals are added together,  however,  they
   •  cannot  simultaneously utilize all, of the complexing capacity.
  Thus a  discharger might feel that it is cost-effective to try to
  justify the lowest site-specific criterion that is acceptable to
  the discharger rather than  trying to justify the  highest site-
  specific criterion that the appropriate regulatory authority
  might approve.

  There are  two  options for dealing with the possibility' of
  additivity and synergism between metals:
  a.  WERs  could  be' developed  using a mixture of  the metals but it
     might be necessary to use several primary toxicity tests
 '    depending on  the  specific metals that  are of interest.  Also,
     it might not  be clear what ratio of  the metals should be  used
     in the  mixture.        ^                                 •
  b.  If a  WER is determined for each metal  individually, one or
     more  additional toxicity tests must  be conducted at the end to
     show  that the combination of all metals.at  their proposed new
     site-specific criteria is acceptable.   Acceptability  must be
     demonstrated-with each toxicity test that was  used as a
     primary toxicity  test in the determination  of  the WERs  for the
     individual metals.   Thus  if  a different primary test  was  used
     for each metal, the  number of acceptability tests  needed  would
     equal the number  of  metals.   It is possible that  a toxicity
     test  used as  the  primary test for one  metal might  be  more
     sensitive than  the CMC (or CCC)  for  another metal  and thus
    might not be usable  in the combination test unless antagonism
     occurs.  When a primary  test  cannot  be used, an acceptable
     alternative test  must be  used.
 The second option  is  preferred because  it  is more definitive; it
 provides data for  each  metal  individually and  for the mixture.
 The first  option leaves the  possibility that one  of  the  metals is
 antagonistic towards  another so  that  the  toxicity of  the mixture
 would increase if  the metal  causing the antagonism were  not
 present.

'            l    -  .' "  -'   ••'    135  ;'   '  " "' ', •..=  ''•-

-------
  Multiple-Discharge Situations
              National Toxics Rule (NTR). incorporated WERs into the
          life criteria for some metals,  it might be envisioned-
  that more than one criterion could apply to a metal at a si?e if
         *'
 has permit
                 whenever a_site-specific criterion is to be
                          site at which more than one discharger
                   for the same metal,  it is important that all
   IiI—T^	• together with the appropriate recrulat-orv
 authority to develop a workplan that is designed to derive a
 site-specific criterion that adequately protects the entire site.

 discharger? ide?"y ^ted'for taking .into  account  more  than one


 Method 1 is straightforward  if the dischargers  are  sufficiently
  ar downstream of each other that  the stream can be divided into
                                 jer.  Method 1  can  also be  fairly
                             are additive, but it will be complex
                            2.   Deciding whether to  use a
                    . water or an actual downstream  water can be
   »^V~T *"  "  flowin£r-water multiple-discharge situation.  Use
   actual downstream water can be  complicated by the existence of
multiple mixing zones and plumes and by the possibility  '
varying discharge schedules,  <->iaaa o=™l ~-~iJi___ _-7_^y
if effluents  from
                             .sras

 ol f^° a=count synergism,  antagonism, -and additivi!?   if one
 of the discharges stops or is modified substantially,  however  it
 if tha ™l% be necessary to  determine a new WER,  eicepriossibly
 if the metal being discharged is refractory.   Situations  iJ°sslD-Ly
                                              need  to be^andled '
Special Flowing-Water Situations
            jntended Jo apply not only to ordinary rivers, and
            S^°J° ftreams ^at some people might consider
«™ ««.*      as streams whose design flows are zero and
streams that some state and/or federal agencies might refer to as
etc   ?Sue SeSdSfc"' "^tat-creating", «effluen?-doBlnIted°,
^oo,:^(?^     differences between agencies, some streams whose
design flows are zero are not considered . »ef fluent-dlpSdent^?

                               136                     .

-------
 etc./and some "effluent-dependent"  streams Have design flows
 that  are greater than zero.)   The application of Method 1 to
.these kinds of streams has tHe following implications:
 1.  If the design flow .-is zero,  at least some WERs ought to be
    determined in 100% effluent.
 2.  If thunderstorms,  etc.,  occasionally dilute the effluent
    substantially,  at  least one WER should be determined in
    diluted effluent to assess whether dilution by rainwater might
    result in underprotection  by decreasing the WER faster than it
    decreases the concentration of the metal.  This might occur,
   ;for example,  if rainfall reduces  hardness,  alkalinity,  and pH
    substantially.   This might not be a concern if the WER
    demonstrates  a substantial margin of safety.
 3.  If the site-specific criterion is substantially higher than
    the national  criterion,  there should be increased concern
    about the fate of  the metal that  has reduced or no toxicity.
    Even.if the WER demonstrates  a substantial .margin of safety
    (e.g.,  if the site-specific criterion is three times the
    national criterion,  but the experimentally determined WER is
    11),  it might be desirable to study the fate of the  metal.
 4.  If the stream merges with  another body of water and  a site-
    specific criterion is desired for the merged waters,  another
    WER needs to  be determined for the mixture of the waters.
 5.  Whether WET testing is  required is not a WER issue,  although
    WET testing.might  be a  condition  for determining and/or using
    a  WER.    ,    .                                   .          -
 6.  A  concern about what species  should be present and/or
    protected in  a stream is a beneficial-use issue,  not a WER
    issue,  although resolution of this issue might affect what
    species should be  used  if  a WER is determined.   (If  the
    Recalculation Procedure is used,  determining what species
    should ,be present  and/or protected is obviously important.)
 7.  Human health  and wildlife  criteria and other issues  might
    restrict an effluent more  than an aquatic life criterion.
 Although there are ho scientific reasons why ."effluent--.
 dependent",  etc.,  streams  and streams whose design flows are  zero
 should be subject  to  different guidance than other streams, a
 regulatory decision (for example,  see 40 CFR 131)  might require
 or  allow some or all  such  streams to be subject  to different
 guidance.   For example,  it might be  decided on the basis/of a use
 attainability analysis  that one  or,more constructed streams do
 not have to comply with usual aquatic life criteria because it is
 decided  that the water  quality in such streams does not need  to
 protect  sensitive  aquatic  species.   Such a decision might
 eliminate  any further concern for site-specifid  aquatic life
.criteria and/or  for WET testing  for  such streams.   The  water-
 quality  might be unacceptable for other reasons,  however.

 In addition to its use  with rivers and streams,  Method  1 is also
 appropriate for  determining cmcWERs  that are applicable to near-
 field effects, of discharges into large bodies  of fresh  or salt
 water, such as an  ocean or a  large lake,  reservoir,  or  estuary:

                               137           :." '.

-------
   * S^ne?r;fieid effe?ts of a PiPe that extends far into a large
     body of fresh or salt water that has a current,  such as an
     ocean, can probably best be treated the saml as  a sinSe
     discharge into a flowing stream.  For example,  if a mixing
     zone is Defined, the concentration of effluent at the edge of
     the mixing zone might be used to'- define how to prepare a
     ?^at?d Slte water.  A dye dispersion study (Kilpatrick
     1992  might be useful,  but a dilution model (U.S. EPA ^3)  is
     JSSS F bS a more.cost-effective way of obtaining       '   S
     information concerning the amount  of dilution at the edge of
     une mixing zone.      .           .    •                •
     -Hie near-field effects of a single discharge" that is near a

     be«tebftL^^%S0dy °f frSSh °r  Salt Wat*r can also probably
     best be treated- the same as a single discharge into a flowing
     stream, especially  if there is a definite plume  and a defined
     !S£ng zone;   J^ Potential point  of Impac? .of near-?ieJd
     effects will  often  be an embayment,  bayou,  or estuary that is
     a nursery for fish  and  invertebrates and/or contain?
     commercially  important  shellfish beds.  Because  of  their
     i^?SaX?;J?ese areas  should receive  special consideration
     **Sje ^termination  and use of  a WER,  taking into account
     ?^S ^°f ^ater and  disqharges' mining patterns, and currents
     (and tides in  coastal areas).  The current  and flushing
    So^  ^S J? estuar:Les can result .in  increased pollutant
    concentrations in confined embayments and at the terminal un-
    SSSSJ' P°f,tioS.°f the estuary^ue  to poor tidal fining Ld
    ^ ?g ^ 1_r^e.dlsPersion studies  (Kilpatrick 1992) can bl
    used to determine the spatial concentration of the effluent in
    the receiving^ater, but dilution models  (U.S. EpJ J^might
    not be sufficiently accurate to be useful.  Dye studies of
    discharges in near-shore tidal areas are especially complex
    Pye injection into the discharge should occur over a?lSaS"

    ^?^a,  P^^rably tw° °r three' c°mplete tidal  cycles?
    S^ff^ent dispersion patterns should be monitored in the
    SSS??^^ °n consecutive^ tidal  cycles using an intensive
    sampling regime over time, location,  and depth   Information
    de??;?1^ disPersf°n and the community at ?Isk can Sensed to
    dSSSS E? appropriate mixing zone(s), which might be used to
    define how to prepare simulated site water.
References
                         S:Lm^ation  of  Soluble Waste Transport
  -        Sro        cerS US±ng  Tracers.  Open-File Report
25425  P^S;?;in  °giCai Survey' Books  and Open-File Reports, Box
-it»4^5, Federal Center, Denver, CO 80225.


?;?* 5P^-^^993-  Dilution Models for Effluent Discharges
Second Edition. .EPA/600/R-93/139.   National Technical
Information Service, Springfield, VA.
                               138

-------
 Appendix 6: Additivity and the Two Components of a WER Determined
             Using Downstream Water


 The Concept of Additivitv of WERs                             ^  .

 In theory, whenever samples of effluent and upstream water are
 taken, determination of a WER in 100 % effluent would quantify
 the effluent WER (eWER) and determination of a WER in 100 .%
 upstream water would quantity the upstream WER  (uWER);          \
 determination of WERs in known mixtures of the two samples would
 demonstrate whether the eWER and the uWER are additive.  For
 example, if eWER =40, uWER = 5, and the two WERs are additive, a
 mixture of 20 % effluent and 80,% upstream water would give a WER
 of 12, except possibly for experimental variation, because: -

       2Q(eWER) * BO(uWER) _ 20(40)  + 80(5)  _  800 + 400   1200 _ . _
              100        ~      100      ~    100   ~  100  ~   '

 Strict additivity of an eWER and an uWER will probably be rare
 because one or both WERs will probably consist of a portion that
 is additive and a portion that is not.  The portions of the eWER
 and uWER that are due to refractory metal will be strictly
' additive, because a change in water quality will not make the
 metal more or less toxic.  In contrast, metal that is nontoxic
 because it is complexed by a complexing agent such as EDTA will
 not be strictly additive because the % uncomplexed will decrease
 as the solution is diluted; the amount of change in the %
 uncomplexed will usually be small and will depend on the
 concentration and the binding constant of .the complexing agent
 (see Appendix D).  Whether the nonrefractory portions of the uWER
 and eWER are additive will probably also depend on the;
 differences between the water quality characteristics of the
 effluent and the upstream water, because these will determine the
 water quality characteristics of the downstream water. .If,  for
 •example, 85. % of the eWER and 30 % of the uWER are due to
 refractory metal, the WER obtained in the mixture of 20 %
 effluent and 80  % upstream water could range from 8 to 12.  The
 WER of 8 would be obtained if the only portions of the eWER and
 uWER that are,additive are those due to refractory metal,
 because:                                                .

    20(0.85) (eWER)  + 80(0;30) (uWER) _ 20(0.85) (40) +80(0.30) (5)
" '  .             100              ~  ~       100~       ~~ ~

 The WER could be as high as' 12 depending on the percentages of
 the other, portions  of the WERs that are also additive.  Even if
 the eWER and uWER are not strictly additive,  the concept of
 additivity of WERs  can be useful insofar as the eWER and uWER are
 partially additive,  i.e.,  insofar as  a portion of at least one of
.the WERs is additive.   In the example given above,  the WER
 determined using downstream water that consisted of 20 % effluent

-------
  and 80  %•' upstream water would be 12  if the eWER and uWER were
  strictly 'additive;  the downstream- WER would be less than 12  if
  the eWER and uWER were partially additive.  - -
                      •*              •                       *

  The Importance  of Additivity   .

  The major advantage of additivity of  WERs  can  be demonstrated
  S?™?*J?e«.£* fluent  and upstream  water that  were used above.   To
  simplify this illustration, the  acute-chronic  ratio will  be
  assumed  to be large, and the  eWER of  40 and the uWER of 5 will be
  assumed  to be cccWERs  that will  be assumed  to  be due to
  refractory metal  and will therefore be strictly addit/ve   In
  addition, the complete-mix downstream water at  design-flow
  conditions will be  assumed to be 20 .% effluent  and  80 % upstream

                                       '
 Because the eWER and the uWER are cccWERs and are strictly
 additive, this metal will cause neither acute nor chronic
 toxicity in downstream water if (a) the concentration of metal in
 the effluent is less than 40 times the CCC and (b)  the
 f£nc^tration of : metal in the upstream water is less than 5 times
 the CCC.  As the effluent is diluted by mixing with upstream
 water,  both the eWER and the concentration of metal will b?
 ?£ lutf^ simultane°usly; proportional dilution of the metal and
 S£i2?5 ^   Prevent the metal from causing acute  or chronic"
 SSSi^fi  fc *!£ ^lution.  When the upstream flow equals the
 design  flow/ .the WER in the plume  will .decrease from 40 at the
 5Sd,°f*-the P13?e t0 P at comPlete  mix as the effluent is diluted
 by upstream water;  because this WER is due to refractory metal,
 neither fate processes nor changes in water quality

 Sa?nS^r^tiCl W^1:L ^ffSCt the WER-   ^eri stream fl°w is higher
 or lower than  design flow,  the complete-mix WER will be lower or
 higher,  respectively,  than 12,  but toxicity ..will  not occur
 because the concentration of metal will also be lower or higher.

 If the  eWER and the uWER are strictly additive and  if 1-he
 ^So1^ C?° J? 1 ""I711,!  the foll°wing conclusions are valid when
                          metal  in  10° % ef f^ent  is  less than 40
                            °£ ^  metal   >1              water
   This metal will not  cause  acute  or  chronic  toxicity  in  the
   upstream water, in 100  % effluent,,  in  the plume, or  in
   downstream water.
   There is no need for an acute or a  chronic  mixing zone  where a
   lesser degree of protection is provided.
   If no mixing zone exists,  there  is  no  discontinuity  at  the
   edge of a mixing zone where the  allowed concentration of metal
   decreases instantaneously.
These results also apply to partial additivity as long  as  the"
concentration of metal  does not exceed that allowed by  the amount
3
                               140

-------
 of additivity that exists.  It would be more difficult to take
 into, account the portions of the eWER and uWER that are not
 .additive.                   .         r       •
                '..'••'          '  :     . •' '' •    •     .     .
 The  concept of additivity becomes unimportant when the ratios,
 concentrations of the metals, or WERs are very different.  For
 example, if eWER = 40, uWER .=. 5, and they are additive,, a mixture
 of 1 % effluent and 99 % upstream water would have a WER of 5.35.
 Given the reproducibility of toxicity tests and WERs, it would be
 extremely difficult to distinguish a:WER, of 5 from a WER of 5.35.
 In cases of extreme dilution, rather than experimentally
 determining a WER, it is probably acceptable to use the limiting
 WER of 5 or to calculate a WER if additivity has been
.demonstrated.

 Traditionally it has been believed that it is environmentally
 conservative to use a WER determined in upstream water (i.e., the
 uWER) to derive a site-specific criterion that applies downstream
 (i.e., that applies to areas that contain effluent).  This belief
 is probably based on the assumption that a larger WER would be
 obtained in downstream water that contains effluent, but the
 belief could also be based on the assumption that the uWER is
 additive.  It is possible that in some cases neither assumption
 is true, which means that using.a uWER to derive a downstream
 site-specific criterion might result in underprotection.   It
 seems likely, however, that WERs determined using downstream
 water will usually be at least as large as the uWER.,

 Several kinds of concerns about the use of WERs are actually
 concerns about additivity:
 1. Do WERs need to be determined at higher flows in addition to
    being determined at design flow?
 2. Do WERs need to be determined when two bodies of water mix?
 3.; Do WERs need to be determined for each additional effluent in
    a multiple-discharge situation...  <,...	  .,:
 In each case, the best use of resources might be to test for
 additivity of WERs.                  "


 Mixing Zones

 In the example presented above,  there would be no need for a
 regulatory mixing zone with a reduced level  of protection if:
 1. The eWER is always 40 and the concentration of the metal in
    100 % effluent is always less than 40 mg/L.
 2. The uWER is always 5 and the concentration of the metal in 100
    % upstream water is always less than 5 mg/L.
 3. The WERs are strictly additive.
 If,  however,  the concentration exceeded 40 mg/L in 100 %
 effluent,  but there is some assimilative capacity in the upstream
 water,  a.regulatory mixing zone would be needed if the discharge
 were to be allowed to utilize some or all of the assimilative

    •  •    '     '   ,  •••'.'-.."   141"":. -   • •       ':      .-  •    ' ':•

-------
  capacity.   The concept of additivity of WERs  can be used to
  calculate the maximum allowed concentration of the metal in the
  effluent if the" eWER and the uWER: are strictly additive?
                       °f        in the
      ™/                                      water never exceeds
   n  Jr/r7 ;*     discharger might want' to determine how much above
  40  mg/L the concentration could 'be  in 100 % effluent   If  £br
  example,  the downstream water at the edge of the chronic mixino
  zone under  design-flow conditions consists of }0 % Sf^eS and
         zone would be:


         TQ(QWER) +30(uNER) _ 70(40)  +30(5)   2800+150
                100               100      =	loo	= 29 •5
offSr™/?he IS*****3* concentration allowed at  this point would
29.5 mg/L.  If the concentration of the metal in the uostream
                                                          osre
 water was 0.8 mg/L  the maximum concentration allowed  n 100  I
 effluent would be 41.8 mg/L because:


      70(41.8 ntcr/L) •»• 30 (0.8 mer/L) _ 2926 mg/L * 24 mall,         ,
                 lOO            ~ - "      - •JU— =29.5 mg/L .
 %flui     «    r    .    conce*tration of the metal in 100
 * ®f?iuen^ « 41.8 mg/L,  there would be chronic toxicity inside
 the chronic mixing zone.   If the concentration in  100  % effluent

 th/SB^o^S11 4J'8 ^g/L^  ^here would be Chronic  toxicity pal?
 the uraS aL ??vf r
-------
 that starts at the edge of the chronic mixing zone and extends
 all the way across the stream, there^wpuld be overprotection at
 the edge of the chronic mixing zone (because the maximum allowed
 concentration is 12 mg/L,  but a concentration of 29.5 mg/L will
 not cause chronic toxicity),  whereas there would be
 underprotection 'on the other side of-'the stream (because the
 maximum allowed concentration is 12 mg/L,  but concentrations
 above 5 mg/L can cause chronic toxicity.)


 The Experimental Determination of Additivitv

 Experimental variation makes  it difficult  to quantify additivity
 without determining a large  number of WERs,  but the advantages of
 demonstrating additivity might be sufficient to make it worth the
 effort.  It should be possible to decide whether the eWER and
 uWER are strictly additive based on determination of the eWER in
 100 % effluent,  determination of the uWER  in 100 % upstream
 water,  and determination of WERs in 1:3, 1:1,  and 3:1 mixtures of
 the effluent and upstream  water,  i ./e.,  determination of WERs in
 100,  75,  50,  25,  and 0 % effluent.   Validating models of partial
 additivity and/or interactions will probably require
 determination of more WERs and more sophisticated data analysis
 (see,  for example,  Broderius  1991).
        • '        "         *           "  '                •       NV--.
 In  some cases chemical measurements or manipulations might help
 demonstrate that at least  some portion of  the eWER and/or the
 uWER is additive:
 1.  If.the difference between  the dissolved WER and the total
    recoverable WER is explained by the difference  between the
  •  dissolved and tota^. recoverable concentrations, .the difference
    is probably due to particulate refractory metal.
 2.  If the WERs in  different samples of the effluent  correlate
    with the concentration  of  metal  in  theieffluent,  all,  or
    nearly all, of  the metal in the  effluent ...is probably nontoxic.
 3.  A> WER  that  remains constant as the  pH is  lowered  to 6.5 and
    raised to  9.0 is probably  additive.
 The concentration  of refractory metal  is likely to be low in   ,
 upstream  water except during  events  that increase  TSS and/or TOC;
 the concentration  of refractory metal  is more  likely to be
 substantial in effluents.  -Chemical  measurements might help
 identify  the percentages of the eWER and the uWER  that are due to
 refractory metal,, but again experimental variation will limit the
 usefulness of  chemical measurements  when concentrations are low.


 Summary

The distinction between the two components of  a WER  determined
using downstream water has the  following implications:
 1. The magnitude of  a WER determined using downstream water will
   usually depend on the percent  effluent  in the sample.

                                143     •    •

-------
    Insofar as the eWER and uWER are additive,  the magnitude  of  a
    downstream WER can be calculated from the eWER1,  the uWER,  and
    the ratio of effluent and upstream water -in the  downstream
    water.           .-
    The derivation and implementation of site-specific criteria
    should ensure that each component is applied only 'where it
    OCCU3T£> •                                          *
                            °CCUr if' f0^ example, any portion of
       the eWER is applied to an area of a stream where the
       effluent does not occur.

    b- ^??rSte?i0n i^J °CTCUr ^f ' for example, an unnecessarily
       small portion of the eWER is applied to an area of .a stream
    .   where the effluent occurs.                           »i-f«euii
    Even though the cpncentration of metal might 'be higher than a
    criterion in both a regulatory mixing zone and a plume, a
    reduced level of protection, is allowed in a mixing zone
    whereas a reduced level of protection is not allowed in 'the
   •portion of a plume that is not inside a mixing zone
    Regulatory mixing zones are necessary if,  and only if  a
    discharger wants to make use of .the assimilative capacity of
    the upstream water.  .                                      •   .
    It might be cost-effective to quantify the eWER .and uWER
    determine the extent of additivity,  study variability over
    time,  and then decide how to regulate the metal in the
Reference
Broderius, S.J,.  1991.  Modeling the Joint Toxicity  of
T^°ii°=i°S m° ^*tic Organisms: Basic Concepts and Approaches
/2 V^S   C To^xcol°9y and Risk Assessment: Fourteenth Volume
(M.A  Mayes and M.G. Barren, eels.. )  ASTM STP 1124.   American-
Society for Testing and Materials, Philadelphia, PA.  pp. 107-
JL^i / .            .                             .
                               144

-------
Appendix H: Special Considerations Concerning the Determination
            of WERs with Saltwater Species

                  '   •                .'.,'*•.        «   - . -
1. The test organisms should be compatible with the salinity of
   the site water/ and the salinity of the laboratory dilution
   water should match that of the site water:  Low-salinity
   stenohaiine organisms should not be tested in high-salinity
   water, whereas high-salinity stenohaiine organisms should not
   be tested in low-salinity water; it is not known, however,
 t  whether an incompatibility will affect the WER.  If the
   community to be protected principally consists of euryhaline
   species, the primary and secondary toxicity tests should  use
   the euryhaline species suggested in Appendix I  (or
   taxonomically related species) whenever possible, although the
   range of tolerance'of the organisms should be checked.
   a. When Method 1 is used to determine cmcWERs at saltwater
      sites, the selection of test organisms is complicated  by ,
      the fact that most effluents -are freshwater and they are
      discharged into salt waters having a wide range of
    ,  salinities.  Some state water quality standards require a
      permittee to meet an LC50 or other toxicity limit at the
      end of the pipe using a freshwater species.  However,  the
      intent of the:site-specific and national water quality
      criteria program is to protect the communities that are at
      risk.  Therefore, freshwater species should not be used
      when .WERs are determined for saltwater sites unless such
      freshwater species (or closely related^species) are in the
      community at risk.  The addition of a small amount of  brine
      and the use of salt-tolerant freshwater species is
      inappropriate for the same reason.  The addition of a  large
      amount of brine and .the use of saltwater species that
      require high salinity should also be avoided when salinity
      is likely to affect the toxicity of the metal.  Salinities
      that are acceptable: for testing euryhaline species can be
   .   produced by dilution of effluent with sea water and/or
      addition of a commercial sea salt or a brine that is
     ; prepared by evaporating site water; small increases -in
      salinity are acceptable because the effluent will be
      diluted with salt water wherever the communities at risk
      are exposed in the real world.  Only as a last resort
      should freshwater species that tolerate low levels of
      salinity and are sensitive to metals, such as Daphnia maqna
      and Hvalella azteca, be used.
   b. When Method 2 is .used to determine cccWERs at saltwater
: ' ••,    sites:         ' - . :-  •.;•_ •'•••.;'"  •".•--/'
      1)  If the site water is low-salinity but all the sensitive
         test organisms are high-salinity stenohaiine organisms,
         a commercial sea salt or a brine that is prepared by
         evaporating site water may be added in order to increase
         the salinity to the minimum level that is acceptable to
         the test organisms; it should.be determined whether the

 .             "—      • ,     ,  145; . .   •       '          "•'

-------
        • f£lt: £.r ^rine Deduces  the toxicity of the metal  and
         thereby increases  the  WER.                mecaj.  ana
      2) If the site water  is high-salinity/selecting test
         organxsms should not < be difficult because mSny of
                                                with
2. It is especially important -to consider the availabilitv of
   test organxsms when saltwater species are to be used  becfii^
   many of the commonly used saltwater species art not culSred
   and are only available seasonally.                  cultured

3. Many standard published methodologies for tests with sal tmt- «r-
   specxes recommend, filtration of dilution  wa^er? effluISt
   and/or test solutxons through a 3 7 -pm sieve or screen to
   remove predators.  Site water should^be fi?terel on!y ±f
   predators are observed in the sample of the water because
                                   .,
water gualxty characteristic, such as salinity
                                                or      is
                                              win

              ^'^s^sss^s^.
                            146

-------
 Appendix I: Suggested Toxicity Tests for Determining WERs for
             Metals  -''''.'  .     •"."••.-"    .   •'.'.•'     . .  .

      •....         '  '. '•   . '•-  •    -.'.-•      "...   . -    ,   .
 Selecting primary and secondary toxicity tests for determining
 WERs for metals should take into account the following:
 1. WERs determined with more sensitive tests are likely to be
    larger than WERs determined with less sensitive tests . (see
    Appendix D).  Criteria are derived to protect sensitive
   .species and so WERs should be derived to be appropriate for
    sensitive species.'  The appropriate regulatory authority will
    probably accept WERs derived with less sensitive tests because
    such WERs are likely to provide at least as much protection as,
    WERs determined with more sensitive tests.,
 2. The species used in the primary and secondary tests must be in
    different orders and should include a vertebrate and an
    invertebrate.                                       ,
-.3. The test organism (i.e., species and life stage) should be
    readily available throughout the testing period.
 4. The. chances of the test being successful should be high.
 5. The relative sensitivities of test organisms vary
    substantially from,metal to metal.
 6. The sensitivity of a species to a metal usually depends on
    both the life stage and kind of test used.
 7. Water quality characteristics might affect chronic toxicity
    differently than they affect acute toxicity (Spehar and
    Carlson 1984; Chapman,  unpublished; Voyer and McGovern 1991).
 8. The endpoint of the primary test in laboratory dilution water
    should be as close as possible (but must not be below)  the CMC
    or CCC to which the WER is to be applied; the endpoint of the
    secondary test should be as close as possible (and should not
    be below)  the CMC or CCC.
 9- Designation of tests as acute and chronic has no bearing on
    whether they may be used to determine a cmcWER or a cccWER.
 The suggested toxicity tests should be considered, but the .actual
 selection should depend on the specific circumstances that apply
 to a particular WER determination. ,                       .-.-',

 Regardless  of whether test solutions are renewed when tests  are
 conducted for other purposes,  if the concentrations of dissolved
 metal and dissolved oxygen remain acceptable when determining
 WERs, tests whose duration is not longer than 48 hours may be
 static tests,  whereas tests whose duration is longer than 48
 hours must  be.renewal tests.  If the concentration ,of dissolved
 metal and/or the concentration of dissolved oxygen does not
 remain acceptable,  the  test solutions  must be renewed every 24.
.hours.   If  one test in  a pair of side-by-side tests is a renewal
 test, both  of the tests must be renewed on the same schedule.

 Appendix H  should be read if.-WERs are  to be determined with
 saltwater species.                         .   .
       "•            '       ' ~             -  -                      f


   - .     '••:'-'    -"-   '     ' '•   147   :   .-•.'.      -' '•  ' ; '.

-------
     Suggested Tests1 for Determining cmcWERs and cccWERs2

        (Concentrations are to be.measured in all tests.)
 Metal
Water3
                             cmcWERs4
Aluminum
Arsenic (III)
Cadmium
Chrom(Ill)
Chrom(VI)

Copper

Lead
-
Mercury
"*
Nickel

Selenium

Silver

Zinc •

FW
. FW
SW
FW
SW
FW
FW
SW
FW
SW
FW
SW
FW
SW
FW
SW
FW
SW
FW
SW
FW
SW
DA
DA
BM
DA
MY,
GM
DA .
MY
- DA
BM
DA
BM
DA •
MY
DA
MY
Y
CR
DA
BM
DA
BM ,
X
GM
CR
SL5 or FM
CR
SL or DA
GM
NE
FM or GM
AR
GM
MYC
GM
BM
FX
BM
Y
MYC
FMC
CR
FM
MY
u(-:<_:w
CDC
CDC
MYC
CDC
MYC
FMC
CDC
MYC
CDC
BMC
CDC
MYC
Y
Y •
CDC
MYC
Y
MYC
CDC
MYC
CDC "
MYC
Eit\S
X
FMC
BM
FMC
X
CDC
GM
NEC
FM
AR
X
X
Y
Y
I
FMC
BMC
Y
X
FMC
BMC
FMC
BMC
                 of a test  specifies  not only the -test  species
and the  duration of the test  but  also the life stage of  the

                 adVSrSe effeGt(s)  on whi<* the endpoint  is  to
rtm      that are  sensitive and are used  in criteria
documents are not suggested here because the chances of  the

™?Xf°£ga?1SinS oei£g availabl* and .the test being successful
might be low.  Such  tests may be used if desired!
                            148

-------
    FW =  Fresh Water;  SW =  Salt Water.

    Two-letter codes are used for  acute  tests, whereas  codes  for
    chronic  tests  cbntain three letters  and  end  in  "C".   One-
    letter codes are used, for comments.

    In acute tests on  cadmium .with salmonids, substantial numbers
    of fish  usually die  after 72 hours.  Also, the  fish are
    sensitive  to disturbance,  and  it is  sometimes difficult,to
    determine  whether  a  fish  is dead or  immobilized.
ACUTE TESTS

AR. A 48-hr EC50 based oh mortality and abnormal development from
    a static test with embryos and larvae of sea urchins of a
    species in the genus Arbacia  (ASTM 1993a) or of the species
    Stronovlocentrotus purpuratus  (Chapman 1992).

BM. A 48-hr EC50 based on mortality and abnormal larval
    development from a static test with embryos and larvae of a
    species in one of four genera  (Crassostrea. Mulinia. Mytilus.
    Mercenaria.) of bivalve molluscs (ASTM 1993b).

CR. A 48-hr EC50 (or LC50 if there is no ,immobilization) from a
    static test with Acartia or larvae of a saltwater crustacean;
    if molting does not occur within the first 48 hours, renew at
    48 hours and continue the test to 96 hours .(ASTM 1993a).

DA. A 48-hr EC50 (or LC50 if there is no immobilization) from a
    /Static test with a species in one of three genera
    (Geriodaphnia.  Daphnia. Simocephalus) in the family Daphnidae
    (U.S. EPA 1993a; ASTM 1993a).

FM. A 48-hr LC50 from a static test at. 25°C with fathead minnow
    (Pimephales-promelas) larvae that are 1 to 24 hours old (ASTM
    1993a; U.S. EPA.1993a).  The embryos must be hatched in the
 .   laboratory dilution water, except that organisms to be used
    in the site water may be- hatched in the site water.  The
    larvae must not be fed before or during the test and at least
    90 percent must survive in laboratory dilution water for at
    least six days after hatch.
       Note:  The following 48-hr LCSOs were obtained at a
             hardness of 50 mg/L with fathead minnow larvae that
      .       were 1 to 24 hours old.   The metal was measured
             using the total recoverable procedure (Peltier
             1993) :
                          Metal              LG50 (ua/L)
                         Cadmium                13.87
                         Copper      ;            6.33
                         Zinc    -              100.95

                               149

-------
      ^-?£~hr«.JC52 from a renewal test (renew at 48 hours) at 25°c
      with fathead minnow (Pimephales £ronielas )  larvae that are 1
      to 24 hours' old (AS1M 1993a; U.S. EPA 1993at   The embryos
      must be hatched, in the laboratory dilution water/ except ?hat
      organisms to be used in the site water may be hatched in the

      ?£? WaSer; n^1^36 ffiust not be fed bef°re or during the
      test and at least  90 percent must survive  in laboratory"
      dilution water_for at least six days after hatch.    ^
         Note:  A 96-hr LC50 of 188.14 ng/L was obtained at a
               hardness  of 50 mg/L in a test 'on  nickel with fathead
               minnow larvae that were 1  to 24 hours old   The
               metal- was measured using the total recoverable
             •  Pro?edure (Peltier 1993).   A 96-hr LC50 is  used for
               nickel because substantial mortality occurred  after
               48 hours  in the test  on nickel, but not in  the tests
               on cadmium, copper, and zinc.                   tests
                   ,(°r LG5° if there is no immobilization) from a
                            48 h°UrS) With a S            fr
                                          a

      uva                    te?t (renew at 48 h°"«> using
     1993aK            Polychaetes in the genus Nereidae (ASTM
CERONIC TESTfi


BMC
                         I   • '     -      -•
A 7-day IC25  from a survival  and development renewal tecsi-
 (renew every  48'hours)  with a species of bivalve moilu-sf
•such as a  species, in the genus Mulinia.   One such test h4s
been described by Burgess  et  al. 1992.   [Note-  Wh^n

                                '              *n
                 test has not been widely used.]

                  based on reduction in survival and/or
                  in a renewal test with a species in the genus
                  in thejfamily Daphnidae (U.S. EPA 1993b).  The
                               150

-------
 MYC
      test, solutions must be renewed every 48 hours.   (A 21-day
   1   life-cycle test with Daphnia maana is also acceptable.)

 FMC.  A 7-day IC25  from a survival and growth renewal test, (renew
      every 48 hours)  with larvae (£ 48-hr old)  of the fathead
      minnow (Pimephales promelas)  (UVS.  EPA 1993b).   When
      determining WERs,  the fish must be fed four hours before
      each renewal  and minimally during the non-renewal days.

    .  A 7-day IC25  based on reduction in survival,  growth,  and/or
      reproduction  in a renewal  ^est with a species in one of  two
      genera (Mysidopsis.  Holmesimvsis [nee Acanthomvsisl)  in  the
.     family Mysidae (U.S.  EPA 1993c).  Mysids must be fed during
      all  acute  and chronic tests,-  when determining WERs,  they
      must be fed four hours before each renewal,  The test
      solutions  must be renewed  every 24  hours.

,NEC.  A 20-day IC25  from a survival and growth renewal test (renew
      every 48 hours)  with a"species in the genus Neanthes  (Johns
      et al.  1991),  -[Note:  When determining WERs, sediment must
      not  be in  the  test  chamber.]   [Note:  This test has not been
      widely used.]
 COMMENTS

 X. Another sensitive test cannot be identified at this time, and
    so other tests used in the criteria document should be
    considered.

 Y. Because neither the CCCs for mercury nor the freshwater
    criterion for selenium is based on laboratory data concerning
    toxicity to,aquatic life, they cannot be adjusted using a WER.
 REFERENCES                             .

 ASTM.  1993a.  Guide for Conducting Acute Toxicity Tests with
 Fishes, Macroinvertebrates, and Amphibians.  Standard E729.
 American Society for Testing and Materials, Philadelphia, PA.

rASTM. _  1993b.  Guide for ConductingStatic Acute Toxicity Tests
 Starting with Embryos of Four Species of Saltwater Bivalve
 Molluscs.  Standard E724.  American Society for Testing and
 Materials,  Philadelphia, PA.

 Burgess,  R.,  G; Morrison, and S. Rego.  1992.  Standard Operating
 Procedure for 7-day Static Sublethal Toxicity Tests for Mulinia
 lateralis.   U.S. EPA, Environmental Research Laboratory,
 Nairragansett, RX.

         '• .    '•  '"....'.-'     -151'.   .        ,

-------
 Chapman, G. A.  1992.  Sea Urchin  (Stronavlocentrotiis
 Fertilization Test Method.  U.S. EPA, Newport, OR.

 Johns, D.M., R.A. Pastorok, and T.G. Ginn.  1991.  A Sublethal
 Sediment Tbxicity Test using Juvenile Neanthes sp.
 (PolychaetarNereidae) .  In: Aquatic"' Toxicology and Risk
 Assessment: Fourteenth Volume.  ASTM STP 1124.   (M.A. Mayes and

                                      for
 Peltier,  W.H.  1993.  Memorandum. to C.E. Stephan.  .October 19.

 Spehar,  R.L., and A.R. Carlson.  1984.  Derivation of Site-
 Specific Water Quality Criteria for Cadmium and the St  Louis
 River Basin,  Duluth, Minnesota.  Environ. Toxicol. Chem. 3:651
 oo'5 .                                                .

 Hif;  EPA*  1993a-   Methods for Measuring the Acute Toxicity of
 Effluents and Receiving Waters to Freshwater and Marine
 Organisms.  Fourth Edition.   EPA/600/4-90/Q27F.   National
 Technical Information Service,  Springfield,  VA.
                    Short-term Methods  for Estimating the  Chronic
          of  Effluents  and Receiving Waters to Freshwater
Organisms.   Third Edition.   EPA/600/4-91/,002.   National Technical
Information  Service, Springfield, VA.

U.S. EPA.  1993c.   Short-term Methods  for Estimating the  Chronic
Toxicity  of  Effluents  and Receiving Waters to Marine and
Estuarine Organisms.   Second Edition.  EPA/600/4-91/003
National  Technical  Information Service, Springfield,  VAJ

Voyer, R.A., and D.G.  McGovern.  1991.  Influence of "constant and
Fluctuating  Salinity on Responses of Mvsidopsis bahia Exposed to
Cadmium in a Life-Cycle Test.  Aquatic Toxicol. 19:215-^230
                               152

-------
 Appendix J: Recommended Salts of Metals


 The following  salts  are recommended for use when determining a
 WER for the metal  listed.   If available, a salt that meets
 American Chemical  Society  .'(ACS)  specifications for reagent-grade
 should be used.                                        ,
        \             '".   '   .    ..'.-.••.       .."         -..•••

 Aluminum                                              , .
 *Aluminum chloride 6-hydrate:  A1C13-6H2O
  Aluminum sulfate  18-hydrate:  A12(SO4)3-18H2O .
  Aluirdnum potassium  sulf ate 12-hydrate: A1K(SO4)2-12H2O

 Arsenic(III)                               •
 *Sodium arsenite:  NaAsO2               "

 Arsenic(V)
  Sodium arsenate 7-hydrate,  dibasic:  Na2HAsO4 • 7H2O

 Cadmium                    ''-.'',   ',
  Cadmium chloride  2.5-hydrate: CdCl2.2.5H2Q
  Cadmium sulfate hydrate:  3CdSO4-8H2Q              .

 Chromium(III).                                           .
 *Chromic chlpride  6-hydrate  (Chromium chloride):  CrCl3-6H2O
 *Chromic nitrate 9-hydrate (Chromium nitrate):  CrJNO3)3-9H2O
  Chrpmium potassium  sulf ate  12-hydrate :.CrK(SO4) 2.12H2O

' Chromium (VT)
  Potassium chromate:  K2CrQ4
 - Potassium dichromate:   K2Cr2O7
 *Sodium chromate 4-hydrate:  Na2CrO4«4H2O     .          -
  Sodium dichromate 2-hydrate:  Na2Cr2O7• 2H2O ',"

 Copper ••'.".•  ,  .    .  " :' ,  •  -  ."' '   ''.'"'<      '' .'.   •:        : , '•
 *Cupric chloride 2-hydrate (Copper chloride) : C,uCl2.2H2O
  Cupric ,nitrate 2.5-hydrate  (Copper nitrate) : Cu(NO3)2-2.5H2O
.  Cupric sulfate 5-hydrate  (Copper sulfate):  CuSO4-5H2O

 Lead   • •:  '       '        •''-•"'  •'.'..   ". •  ,\  '-  •  •  .  -
 *Lead chloride: PbCl2
  Lead nitrate:  Pb(NO3)2

Mercury
  Mercuric chloride: HgCl2
  Mercuric nitrate monohydrate: Hg(NO3)2-H2O
  Mercuric sulfate: HgSO4
                                153

-------
 Nickel '
 *Nickelous chloride 6-hydrate  (Nickel  chloride)-  NiCl .-SH n
 *No.ckelous nitrate 6-hydrate  (Nickel nitrate) •  Ni (NO >  6H o
  Nxckelous sulfate 6-hydrate  (Nickel sulfate) i  NiSO?'6H2O *
 Selenium (TV)                       .-',,'
 *Sodium selenite 5-hydrate: Na2Se03.5H2O
.Selenium ( VI }                     "       .-•....
 *Sodium selenate 10-hydrate: Na2SeO4-10H2O

 Silver
  Silver nitrate:  AgNO3

                                                         -silver
Zinc                             '                            ;
 Zinc chloride : ZnCl2
*Zinc nitrate 6-hydrate:  Zn (NO, ) , . 6H,O
 Zinc sulfate 7-hydrate:  ZnSO4.7H2O                  •


*Note: ACS reagent -grade ^ specifications  might not be available

                              154

-------