8f.
nf       Integrated Approach to Assessing the Bioavailability
             and Toxicity of Metals in Surface Waters and
                                Sediments
                Presented to the EPA Science Advisory Board
                                               U S EPA Headquarters Library
                                                   Mail code 3201
                                               1200 Pennsylvania Avenue NW
                                                 Washington DC 20460
                               April 6-7,1999

                      U.S. Environmental Protection Agency
                               Office of Water
                      Office of Research and Development
                              Washington, D.C.

-------

-------
                  Principal Authors
                  In Alphabetical Order

                       Herbert E. Allen
           University of Delaware, Newark, DE 19716

                        Heidi E. Bell
   U.S. Environmental Protection Agency, Washington, D.C. 20460

                       Walter J. Berry
   U.S. Environmental Research Laboratory, Narragansett, RI02592

                     Dominic M. Di Toro
              HydroQual, Inc., Mahwah, NJ 07430
Environmental Engineering Department, Manhattan College, Bronx NY

                       David J. Hansen
     Great Lakes Environmental Center, Traverse City MI 496S6
              HydroQuaL Inc., Mahwah, NJ 07430

                       Joseph S. Meyer
           University of Wyoming, Laramie, WY 82071

                     Jennifer L. Mitchell
   U.S. Environmental Protection Agency, Washington, D.C. 20460

                       PaulR. Paquin
              HydroQual, Inc., Mahwah, NJ 07430

                       MaryC. Reiley
   U.S. Environmental Protection Agency, Washington, D.C. 20460

                      Robert C. Santore
              HydroQual, Inc., Mahwah, NJ 07430

                     Contributors
                  In Alphabetical Order

                     Harold L. Bergman
           University of Wyoming, Laramie, WY 82071

                     Warren S. Boothman
   U.S. Environmental Research Laboratory, Naragansett, RI 02592

                       Tyler K. Linton
      Great Lakes Environmental Center, Columbus, OH  43212

                       John D. Mahony
Environmental Engineering Department, Manhattan College, Bronx NY

                       Joy A. McGrath
               HydroQual, Inc., Mahwah, NJ 07430

                       Devanshi Trivedi
               HydroQual, Inc., Mahwah, NJ 07430

                       Kiuen-BingWu
               HydroQual, Inc., Mahwah, NJ 07430

-------
 se.
ir

-------
                                     Disclaimer
       This document is a compilation of data and analyses from scientific investigations into the
bioavailability of metals in surface waters and sediments to aquatic and benthic organisms.  The
intent of this document is to propose an integrated approach to assessing metals contamination of
surface waters and sediments for the protection of aquatic and benthic organisms.

       This document does not establish or affect legal rights or obligations. It does not establish
a binding norm and is not finally determinative of the issues addressed. Agency decisions in any
particular case will be made applying the law and regulations on the basis of specific facts when
permits are issued or regulations promulgated.

       The mention of trade names or specific products does not constitute endorsement.

-------

-------
                                    Contents

Introduction and Charge  		,	vii
Metals Mixtures Equilibrium Partitioning Sediment Guideline (ESG) Document ...	1-1
Sediment Assessment Presentation Materials	.2-1
      The Technical Basis of the Use of Acid Volatile Sulfide (AVS) and Interstitial
        Water Normaliziations in Sediment Guidelines	:......		2-2
      Predicting the Toxicity of Metals in Sediments		: 2-22
      Chromium Chemistry in Sediments	2-38
      The Addition of Chromium to the Metals Mixtures ESG	'.....	2-50
      The Addition of Silver to the Metals Mixtures ESG ...'.	 2-56
A Model of the Acute Toxicity of Metals	3-1
      Part 1: Technical Basis  for the Biotic Ligand Model
      Part 2:, Application of the Biotic Ligand Model for Copper to Fish and Invertebrates
      Part 3: Application of the Biotic Ligand Model for Silver to Fish and Invertebrates
                    Appendices - Background Material


Appendix A: Research Papers on the Bioavailability and Toxicity of Metals in Sediments

Key Papers (Roughly in order of importance)

Ankley, G.T., D.M. Di Toro, DJ; Hansen, J.D. Mahoney, W. J. Berry, R.C. Swartz, R. A. Hoke, N. A.
Thomas, A.W. Garrison, H.E. Allen and C.S. Zarba. 1994. Assessing potential bioavailability of
metals in sediments: A proposed approach. Environ. Management.  18:331-337.

Ankley, G.T., D.M. Di Toro, DJ. Hansen and WJ. Berry.  1996. Technical basis and proposal for
deriving sediment quality criteria for metals. Environ. Toxicol. Chem. 15:2056-2066.

Di Toro, D.M., J.D. Mahoney, DJ. Hansen, K.J. Scott, M.B. Hicks, S.M. Mayr and M.S. Redmond.
1990. Toxicity of cadmium in sediments: The role of acid volatile sulfide. Environ. Toxicol. Chem.
9:1487-1502.                                  .          .

Hansen, D.J., WJ. Berry, J.D. Mahoney, W.S. Boothman, D.M. Di Toro, D.L. Robson, G.T. Ankley,
D. Ma, Q. Yan and C.E. Pesch. 1996. Predicting the toxicity of metal-contaminated field sediments
using interstitial concentration of metals and acid-volatile sulfide normalizations. Environ. Toxicol.
Chem. 15:2080-2094.

Berry, W.J., D J. Hansen, J.D. Mahoney, D.L. Robson, D.M. Di Toro, B.P. Shipley, B. Rogers, J.M.
Corbin and W.S. Boothman.  1996. Predicting the toxicity of metal-spiked laboratory sediments
                                         in

-------
  using acid-volatile sulfide and interstitial water normalizations. Environ. Toxicol. Chem. 15:2067-
'2079..      '        .         ,                             .

  Call, DJ., C.N. Polkinghorne,  T.P. Markee, L.T. Brooke, D.L. Geiger, J.W. Gorsuch and K.A.
  Robillafd.  1999. Silver toxicity to Chironomus tentans in two .freshwater sediments. Environ.
  Toxicol. Chem. 18:30-39.

  Berry, WJ., M.G. Cantwell, P. A. Edwards, J.R. Serbst and DJ. Hansen. 1999. Predicting toxicity
  of sediments spiked with silver. Environ. Toxicol. Chem.  18:40-48.

  Di Toro, D.M., J.D. Mahoney, D J. Hansen, KJ. Scott, A.R. Carlson and G.T. Ankley.  1992. Acid
  Volatile Sulfide predicts the acute toxicity of cadmium and nickel in sediments.  Envrion. Sci.
.  Techno!. 26:96-101..
 Other Relevant Papers

 Ankley, G.T. 1996. Evaluation of metal/acid-volatile sulfide relationships in the prediction of metal
 bioaccumulation by benthic macroinvertebrates. Environ. Toxicol. Chem.  15:2138-2146.

 DeWitt, T.H., R.C. Swartz, D J. Hansen, D. McGovem and WJ. Berry. 1996. Bioavailability and
 chronic toxicity  of cadmium in sediment to the  estuarine amphipod Leptocheirus plumulosus.
 Environ. Toxicol. Chem. 15:2095-2101.

 Di Toro, D.M., J.D. Mahoney and A.M. Gonzalez.  1996. Particle oxidation model of synthetic FeS
 and sediment acid-volatile sulfide.  Environ. Toxicol. Chem. 15:2156-2167.

 Di Toro, D.M., J.D. Mahoney, D.J. Hansen and W J. Berry. 1996.  A model of the oxidation of iron
 and cadmium sulfide in sediments. Environ. Toxicol. Chem. 15:2168-2186.

 Gonzalez, A.M.  1996.  A laboratory-formulated sediment incorporating synthetic acid-volatile
 sulfide. Environ. Toxicol. Chem. 15:2209-2220.                                          \

 Hansen, D J., J.D. Mahoney, W J. Berry, S J.'Benyi, J.M. Corbin, S.D. Pratt, D.M. Di Toro andM.B.
 Abel. 1996. Chronic effect of cadmium in sediments on colonization by benthic marine organisms:
 An evaluation of the role of interstitial cadmium and acid-volatile sulfide in biological availability.
 Environ. Toxicol. Chem. 15:2126-2137.

 Hassan, S.M., A.W. Garrison,  H.E. Allen, D.M! Di Toro and G.T. Ankley.  1996.  Estimation of
 partition coefficients for five trace metals in sandy sediments and application to sediment quality
 criteria. Environ. Toxicol. Chem. 15:2198-2208.

 Leonard, E.N., G.T. Ankley and R.H. Hoke. Evaluation of metals in marine and freshwater surficial
 sediments from the environmental monitoring and assessment program relative to proposed sediment
 quality criteria for metals. Environ. Toxicol. Chem.  15:2221-2232.
                                           IV

-------
Liber, K., D.J. Call, T.P. Markee, K.L. Schmude, M.D. Balcer, F.W. Whiteman and G.T. Ankley.
1996.    Effects  of acid-volatile  sulfide on  zinc bioavailability and toxicity  to  benthic
.macroinvertebrates: A spiked-sediment field experiment. Environ. Toxicol. Chem.  15:2113-2125.

Mahoney, J.D., D.M. DiToro, A.M. Gonzalez, M. Curto, M. Dilg, L.D. De Rosa and L.A. Sparrow.
1996. Partitioning of metals to sediment organic carbon. Environ. Toxicol. Chem.  15:2187-2197.

Peterson, G.S., G.T. Ankley and E.N. Leonard.  1996. Effect  of bioturbation on metal-sulfide
oxidation in surficial freshwater sediments. Environ. Toxicol. Chem. 15:2147-2155.

Sibley, P.K., G.T. Ankley, A.M.  Cotter and E.N. Leonard. .1996. Predicting chronic toxicity of
sediments spiked with zinc: An evaluation of the acid-volatile sulfide model using a life-cycle test
with the midge Chironomus tentans. Environ. Toxicol. Chem. 15:2102-2112.
Appendix B. Research Papers on the Bioavailability and Toxicity of Metals in Surface
             Waters

Key Papers                                       '

Allen, H. E. and D. J. Hansen. 1996. The importance of trace metal speciatiori to water quality
criteria.  Water Environment Research. 68(l):42-54.

Campbell, P.G.C., 1995. "Interactions Between Trace Metals and Aquatic Organisms: A Critique
of the Free-ion Activity Model" Metal Speciatipn ana1 Bioavailability in Aquatic Sy^
and D.R. Turner, eds., IUPAC, John Wiley and Sons.

Janes, N. and R.C Playle. 1995. Modeling silver binding to gills of rainbow trout {Oncorhynchus
myMss).  Environ. Toxicol Chem.  14; 1847-1858.

Meyer, J.S., R;C. Santore, J.P. Bobbitt, L.D. DeBrey, CJ. Boese, P.R. Paquin, H.E. Allen, H.L.
Bergman and D.M. DiToro.  1999.  Binding of nickel and copper to fish gills predicts toxicity when
water hardness varies, but free-ion activity does not. Environ. Sci. Technol. (in Press).

Meyer, J.S. 1999. A mechanistic explanation for the ln(LC50) vs ln(Hardness) adjustment equation
for metals. Envrion. Sci. Technol.  (in Press).

Pagenkopf, G.K.  1983. Gill surface interaction model for trace-metal toxicity to fishes: Role of
complexation, pH, and water hardness. Envrion. Sci. Technol. 17:342-347.

Playle, R.C., D.G. Dixon and K. Burnison. 1993a.  Copper and cadmium binding to fish gills:
modification by dissolved organic carbon and synthetic ligands. Can. J. Fish. Aquat. Sci. 50:2667-
2677.                                                                  ,

-------
Playle, R.C., D.G. Dixon and K. Burnison. 1993b. Copper and cadmium binding to fish gills:
estimates of metal-gill stability constants and modeling of metal accumulation. Can. J. Fish. Aquat.
Sci.  50:2678-2687.

Tipping, E. 1994.   WHAM-A chemical equilibrium model  and computer code for waters,
sediments,  and soils incorporating a discrete site/electrostratic model of ion-binding by humic
Substances. Computers and Geosciences, 20:973-1023.                                 ,
Other Relevant Papers

Allison, J.D., D.S. Brown, KJ. Novo-Gradac.  March 1991.  MINTEQA2/PRODEFA2, A
Geochemical Assessment  Model for Environmental, Systems:  Version 3.0, Users Manual.
EPA/600/3-91/021, USEPA ERL ORD, Athens, GA.

Bury, N;R., F. Galvez and C.M. Wood. 1998. Effects of chloride, calcium and dissolved organic
carbon on silver toxicity: Comparison between rainbow trout and fathead minnows.  Environ.
Toxicol. Chem. 18:56-62.

Fehny, A.R., D.C. Girvin and E.A. Jenne, 1984. MINTEQ: A Computer Program for Calculating
Aqueous Geochemical Equilibria. USEPA Environmental Research Laboratory, Office of Research
and Development, Athens, Georgia.

Gorsuch, J. ans S. Klaine, Editors. 1999. Annual Review Issue: Silver Toxicity. Environ. Toxicol
Chem. 18:1-108.                                                     .

MacRae, R.K., D.E. Smith, N. Swoboda-Colberg, J.S. Meyer and H.L. Bergman. 1999. Copper
binding affinity of rainbow trout (Oncorhynchus myJass) and brook trout (Salvelinusfontinalis)&\\s.
Environ. Toxicol. Chem. (in Press).

Playle, R.C., R.W. Gensemer and D.G. Dixon, 1992. Copper accumulation on gills of fathead
minnows:  Influence of water hardness, complexation and pH on the gill micro-environment.
Environ. Toxicol. Chem. 11:381-391.                                         .

Playle, R.C. 1999.  Physiological and lexicological effects of metals at gills of freshwater fish.
Environ. Toxicol. Chem. (in Press).

Santore, R.C. and C.T. Driscoll. 1995. The CHESS Model for calculating chemical equilibria in
soils and solutions. Chemical Equilibrium and Reaction Models, SSSA Special Publication 42, The
Soil Society of America, American Society of Agronomy.

Schecher,  W.D. and D.C.  McAvoy. 1992.  MINEQL+: A software environment for chemical
equilibrium modeling. Computer Environ. Urban Systems.16:65-76.

Wood, C.M:, R.C. Playle and C. Hogstrahd.  1999. Physiology and modeling of the mechanisms
of silver uptake and toxicity in fish. Environ. Toxicol. Chem. 16:71-83.
                                         VI

-------
                                    SAB REVIEW:
            Integrated Approach to Assessing the Unavailability and Toxicity
                       of Metals in Surface Waters and Sediments

                                    April 6-7,1999
INTRODUCTION

       Persistent contaminants discharged into the waterbodies of the U.S. may remain in the water
column (in various chemical forms), partition into sediments or accumulate in aquatic and benthic
biota.  In particular, metals such as cadmium, copper, lead, nickel, silver and zinc are commonly
elevated  in aquatic sediments due to their widespread release and  persistent nature.  Several
environmental variables affect the partitioning of metals amongst media and the chemistry of the
metals within each medium. The partitioning behavior thereby strongly influences the chemicals'
bioavailability in aquatic systems.  Fundamental to assessing the potential for metals to adversely
affect aquatic biota is the capability to quantify metals bioavailability and toxicity.

       For the past several years the Office of Water (OW) and the Office of Research and
Development (ORD) have been investigating the bioavailability of metals in surface waters and
sediments. These efforts have led to a firm understanding of the environmental factors that influence
metals bioavailability and have allowed the Agency to produce more accurate guidance on the
quantification of potential metals toxicity to aquatic and benthic organisms.

       Presently the Agency has existing aquatic life ambient water quality criteria (AWQC) for 11
individual metals and a draft sediment guidance document based on the equilibrium partitioning
(EqP) approach for a mixture of five metals. Recent research developments have allowed OW to
develop guidance that more fully accounts for factors influencing the bioavailability of metals in
both surface waters and sediments. With these developments EPA will be providing users with an
integrated approach to assessing metals bioavailability and toxicity to both benthic and aquatic
organisms. EPA believes this integrated approach is a significant improvement over current EPA
guidance because it better addresses metals toxicity in the water column and in sediments. Likewise,
it provides users with complimentary tools designed to work together for aquatic system assessment
and protection.                       ,

The following question is charged to the SAB for its consideration:

       Does this integrated metals methodology improve our ability to make both protective
       and predictive assessments of toxicity due to copper, silver and other selected metals
       in the water column and sediment?

       The information presented below briefly summarizes the history of OW guidance on metals
toxicity, research developments leading to the new guidance, and the key issues currently in need
of peer review. The balance of the advance briefing materials provide the theoretical and scientific
foundation needed for the SAB to make its evaluation.                            ^
                                           vn ,

-------
Unavailability and Toxicity of Metals in Sediments

Background
                               f
       In January of 1995, the Sediment Quality Criteria Subcommittee of the Ecological Processes
and Effects Committee (EPEC) of the SAB met to review research and data developed to support
the use of the equilibrium partitioning (EqP) approach to predict the bioavailability of five metals
in sediment (i.e., cadmium, copper, lead, nickel and zinc).  The EqP  approach assumes that a
contaminant in sediment partitions between the sediment, interstitial (pore) Water and organisms.
In the case of metals,  a key factor controlling the partitioning between these phases is the acid
volatile sulfide (AVS) concentration in the sediment. Sulfide binds a number of cationic metals
(such as cadmium, copper, nickel, lead, silver and zinc) forming insoluble sulfide complexes which
are minimally bioavailable. The difference between the AVS concentration (i.e., [AVS]) and the
measured metals concentrations (simultaneously extracted metal or [SEM]) in the sediment indicates
the potential for toxicity. So long as the [AVS] * [SEM] the sediments are not expected to cause
acute or chronic toxicity to benthic organisms.

     ,  In addition to the solid phase Equilibrium Partitioning Sediment Guideline (ESG) described
above, a proposed interstitial water phase ESG was also presented to the SAB in 1995. Since the
sensitivities of benthic and water column organisms to metals are similar, the currently established
AWQC final chronic values (FCVs) can be used to define the acceptable effects  concentration of
freely dissolved metals in the interstitial water. .Thus, the interstitial water phase ESG is defined as:
the sum of the (metals concentrations in interstitial water/the metal-specific Final Chronic Value)
based on dissolved metal is less than or equal to one (SfMJ/FCVjS 1).,.

       Thus, the guidance reviewed by the SAB in 1995 stated that ESGs for all five metals could
collectively be derived using two  procedures:  (a) by comparing the sum  of  their  molar
concentrations,  measured as SEM, to the molar concentration of AVS in sediments, or  (b) by
comparing the  measured interstitial  water concentrations of the metals to FCVs.   A lack of
exceedence of the solid phase or interstitial water phase ESGs based upon either of the  two
procedures indicates that metal toxicity should not occur.  Exceedence of both values would be
indicative of a potential problem that would require further evaluation.

Recent Developments

       Since the 1995 SAB Review, OW  has conducted additional work to respond  to the
recommendations of the SAB, updated the document and developed a revised draft entitled
"Equilibrium Partitioning Sediment Guidelines (ESGs) for the Protection of Benthic Organisms:
Metals Mixtures (Cadmium, Copper, Lead, Nickel, Silver and Zinc)"; hereafter referred to as the
Metals Mixtures ESG. Research developments have enabled OW to incorporate both silver and
chromium into the Metals Mixtures ESG. Researchers have also identified a means to predict metals
toxicity by normalizing with the fraction organic carbon (foe) concentration present in the sediment.
We believe the addition of these new components to the Metals Mixtures ESG has greatly increased
our ability to be both protective and predictive of metals toxicity in sediments.
                                          Vlll

-------
The SAB Review

The sediment presentations will include:

•      An introduction to ESGs: their derivation, use, and application;
•      A summary of the technical basis on the use of AVS and interstitial water normalization in
       sediment guidelines;
•      A discussion on  the use of fraction organic carbon (foe) normalization as a method of
       increasing our ability to predict toxicity in contaminated sediments; and
•      A discussion of the bioavailability and toxicity of chromium and silver in sediments, and the
       incorporation of these two metals into the Metals Mixtures ESG.

Although invited to comment on any part of the Metals Mixtures ESG document and associated
presentations, the SAB is requested to respond to the following questions:

1) By incorporating the fraction organic carbon into the bioavailability equation, have we
   retained the protective features of the guidelines and improved its predictiveness of toxic
   effects?
2) If the BLM is used to derive or adjust a water quality criterion, is the revised criterion
   appropriate for use in the interstitial water component of the Metals Mixtures ESG?
3) Are the data presented from lab and field experiments with chromium and silver sufficient to
  ^support then- addition to the Metals Mixtures ESG?


Bioavailability and Toxicity of Metals in Surface Waters
           . ^\ •     '
Background

       The ambient water quality program has long been criticized that the numerical aquatic life
water quality criteria approach to assessing and controlling metals in aquatic systems is not adequate.
Complaints from the regulated community, States, Regions, environmental groups, and academics
are that the Agency's laboratory based metals criteria do not accurately reflect the toxicity of metals
in ambient waters. Therefore, they may be under or over protective. The issue was partially rectified
when the Agency recommended that ambient metals  criteria be expressed  as dissolved metal
concentrations rather than as total recoverable metal concentrations.  The change was an attempt to
more fully  account for the bioavailable concentration of metals in the water column, beyond the
existing hardness correction, to aquatic organisms. Though most considered the recommendation <
a welcome change in Agency policy, this change was only one of a number that three expert
workshops (Annapolis, 1/93, Pellston at Pensacola, 2/96, and Pellston at Gregson, 6/98) cited as
potential improvements to metals assessment and control. Three other significant changes advanced
by the workshops were to: (1) incorporate the "biotic ligand model" (BLM  - formerly the "gill
model") into the aquatic life water quality criteria for metals to more precisely account for the
bioavailability  of metals to aquatic life, (2)  develop  metals mixtures equilibrium partitioning
sediment guidelines (ESGs) to address  metals contamination in sediments and compliment the
                                           IX

-------
dissolved criterion in the water column, (3) include sediment as a source and a sink for metals in
TMDL development.

       Over the last year, these improvements have become even more critical to the program.
Requirements to develop TMDLs across the country and the need to resolve issues with the Fish and
Wildlife Service and National Marine Fisheries Service on how to ensure adequate protection of
threatened and endangered species are driving program activities and time lines: We believe these
two assessment tools will go a long way toward resolving the concerns of both the Services and the
regulated community.

Proposal     .                                                                        •
                                  -                     x

       Research on the acute toxicity of copper to fish has revealed that copper toxicity is
manifested by gill membrane disruption. Copper binds to metal binding sites on the gill causing the
gill membrane to lose osmoregulatory function and become "leaky."   Should enough copper
accumulate on the  gill, the organism dies.  From this information it  was theorized that the
concentration of copper on the gill rather than the concentration of copper in the water was the
critical determinant of copper toxicity.

       OW and ORD have  supported the development  of a Biotic Ligand Model (BLM) which
quantifies the capacity of metals to bind to the gills of aquatic organisms. This model can be used
to calculate the bioavailable portion of dissolved metals in the water column based on site-specific
water quality parameters such as alkalinity, pH and dissolved organic carbon.  Currently the BLM
can be applied to copper and silver, and future efforts should allow it to be applied to additional
metals. .                                                   .

       The short-term goal for OW is to provide States and Regions with a BLM computer program
and supporting documentation for use as an alternative to the current Water Effect Ratio (WER)
methodologies. This will provide a site-specific tool-that is less expensive and less time consuming
to use and .that limits regression analysis. Based on the results of this SAB review, we will conduct
any necessary additional investigations and revise the guidance.

       In the longer term OW plans to conduct the necessary work to incorporate the BLM directly
into metals criteria documents.  OW and ORD will perform the necessary investigations and
verification to incorporate the model directly into the copper and silver aquatic life water quality
criteria derivation.  At the same time, it is our intention to initiate an application of the model to a
number of cadmium and lead data sets to determine which, if any, water quality competition and
binding parameters need to be adjusted.

The SAB Review

The Biotic Ligand Model presentations will include:

•     A discussion of the regulatory need for a new approach to assessing metals toxicity in the
       water column;                                                        ,  .

-------
•      A summary of metals chemistry and bioavailability to water column organisms;
•      A description of the BLM and its application to copper and silver; and
•      A summary of the anticipated future directions of the BLM.

The SAB is requested to address the following questions pertaining to the BLM:
                          ..      '                          v,
1) Does the BLM improve our ability to predict toxicity to water column organisms due to
   metals (copper and silver) in comparison to the currently applied dissolved metal
   concentration criterion?
2) Is the scientific and theoretical foundation of the model sound?
3) In comparison to the current WER adjustment for aquatic life criteria^ will the application of
   the BLM as a site-specific adjustment reduce uncertainty associated with metals
   bioavailability and toxicity?
4) Are the data presented for the validation of the BLM sufficient to support the incorporation of
   the BLM directly into copper and silver criteria documents?
                                           XI

-------

-------
                 DRAFT
EQUILIBRIUM PARTITIONING GUIDELINES (ESG) FOR




   THE PROTECTION OF BENTHIC ORGANISMS:




              METALS MIXTURES




         CADMIUM, COPPER, LEAD, NICKEL,




               SILVER AND ZINC
        U.S. Environmental Protection Agency:




         Office of Science and Technology and




         Office of Research and Development

-------
                               CONTENTS
Section
1     INTRODUCTION	1-10
      1.1   GENERAL INFORMATION 	. . . .  .	1-10
      1.2   OVERVIEW OF DOCUMENT	1-16

2     PARTITIONING OF METALS IN SEDIMENTS		 . 1-17
      2.1   METAL TOXICITY IN WATER-ONLY AND IN INTERSTITIAL WATER
           OF SEDIMENT EXPOSURES  ...	 1-17
           2.1.1  Toxicity correlates to metal activity	 1-18
           2.1.2  Toxicity correlates to interstitial water concentration	1-20
      2.2   SOLID PHASE SULFIDE AS THE IMPORTANT BINDING
           COMPONENT	  .	 ,	1-28
           2.2.1  Metal Sorption Phases	1-28
           2.2.2  Titration Experiments	1-29
                 2.2.2.1 Amorphous FeS	1-30
                 2.2.2.2 Sediments		 1-30
           2.2.3  Correlation to Sediment AVS	1-33
           2.2.4  Solubility Relationships and Displacement Reactions	1-34
           2.2.5  Application to Mixtures of Metals	'.	1-37

3     TOXICITY OF METALS IN SEDIMENTS	1-39
      3.1   GENERAL INFORMATION	1-39
      3.2   PREDICTING THE TOXICITY OF METALS IN SEDIMENTS,	1-39
      3.3   PREDICTING METAL TOXICITY:  SHORT-TERM STUDIES ...... 1-53
           3.3.1  Spiked sediments: Individual experiments	1-53
           3.3.2  Spiked sediments: Chromium experiments	 1-55
                 3.3.2.1 Chromium Chemistry in Sediments	 1-57
           3.3.3  Spiked Sediments:  All experimental  results summarized . . :	1-68
           3.3.4  Field sediments	:	; . .	1-74
           3.3.5  Field Sites and Spiked Sediments Combined	 1-77
      3.4   PREDICTING METAL TOXICITY: LONG-TERM STUDIES  	1-81
           3.4.1  Life-cycle toxicity  tests	1-82
           3.4.2  Colonization tests  . . I	 1-84

4     DERIVATION OF ESG FOR METALS .	1-87
      4.1   GENERAL INFORMATION 	*	; ...	1 1-87
      4.2   SINGLE METAL SEDIMENT GUIDELINES 	1-88
           4.2.1  AVS Guidelines	1-89

                              Draft for SAB 1-2 \

-------
i
           4.2.2  Interstitial Water Guidelines	1-89
      4.3   MULTIPLE METALS GUIDELINES :	1-90
           4.3.1  AVS Guidelines	1-92
           4.3.2  Interstitial Water Guidelines	'.....'	1-93
      4.4   ESG FOR METALS VS. ENVIRONMENTAL MONITORING
           DATABASES	1-95
           4.4.1  Data Analysis	 1-95

5     IMPLEMENTATION		„......:.	1-100
      5.1   CONSIDERATIONS IN PREDICTING METAL TOXICITY  	1-100
      5.2   SAMPLING AND STORAGE	  . 1-100
           5.2.1  Sediments .	1-102
           5.2.2  Interstitial Water	  . 1-103
      5.3   ANALYTICAL MEASUREMENTS		1-104
           5.3.1  Acid Volatile Sulfide .		 ..	.....'.	1-105
           5.3.2  Simultaneously Extracted Metal  :	1-105
           5.3.3  Interstitial Water Metal		1-106
      5.4   ADDITIONAL BINDING PHASES	1-106
      5.5  . PREDICTION OF THE RISKS OF METALS IN SEDIMENTS
           BASED ON EqP	  . 1-107

6     GUIDELINES STATEMENT	1-108

7     REFERENCES			1-110
         Appendix A: Glossary of abbreviations and equations
         Appendix B: Solubility Relationships for metals sulfides
         Appendix C: Lake Michigan EMAP sediment monitoring database
         Appendix D: Saltwater sediment monitoring database
                                       Draft for SAB 1,3

-------
                                      TABLES
Table 2-1.    Cadmium binding capacity and A VS of sediments
             (from Di Toro et al., 1990)		.	, 1-33,

Table 2-2.    Metal sulfide solubility products .	 1-37

Table 3-1.    Results of a toxicity test with the amphipod Ampelisca abdita ,
             exposed to sediments spiked with chromium VI	1-58

Table 3-2.    Toxicity of sediments from saltwater (SW) and freshwater (FW)
             field locations, spiked-sediment tests and combined field and
             spiked-sediment tests as a function of the difference between the
             molar concentrations of SEM and AVS (SEM-AVS), interstitial
             water toxic units (IWTUs) and both SEM-AVS and IWTUs
             (modified from Hansen et al., 1996a)	 . .	1-72

Table 3-3.    Summary of the results of full life-cycle and colonization toxicity
             tests conducted in the laboratory and field using sediments spiked
             with individual metals and metal mixtures	1-83

Table 4-1.    Water quality criteria (WQC) criteria continuous concentrations
             (CCC) based on the dissolved concentration of metal. These
             WQC CCC values are for use in the; IWTU procedure for
             deriving sediment guidelines based on the dissolved metal
             concentrations in interstitial water  	1-91
                                   Draft for SAB 1-4

-------

                                      FIGURES

                                                                                  Page
Figure 2-1.   Acute toxicity to Palaemonetes of total cadmium (top) and
             cadmium activity (bottom) with different concentrations of the
             complexing ligands NTA (left) and chloride as salinity (right)
             (This figure is from Sunda et al., 1978)	1-19

Figure 2-2.   Acute toxicity of total copper (top left) and copper activity
             (bottom left) to a dinoflagelate, with and without the complexing
             ligand EDTA (this portion of the figure is from Anderson and
             Morel, 1978). Toxicity of zinc to Microcystis aeruginosa
             showing growth as cells/ml versus tune with different levels of
             the complexing ligands EDTA and NTA (top right) and number
             of cells at five days as a function of free zinc concentration
             (bottom right) (this portion of the figure is from Allen et al.,
             1980)	.		1-21

Figure 2-3.   Specific growth rate of a diatom (left) (this portion of the figure
             is from Sunda and Guillard, 1976) and Microchrysis lutheri
             (right) versus total copper (top) and copper activity (bottom) for a
             range of concentrations of the complexing ligands Tris and
             natural DOC (this portion of the figure is from Sunda and Lewis,
             1978)		•..:..'	'. .  1-22

Figure 2-4.   Tissue accumulation of copper in oysters (Crassostrea virginica)
             versus total copper (top) and copper activity (bottom) with
             different levels of the complexing ligand NTA (this figure is from
             Zamuda  and Sunda, 1982)  .'	,	1-23

Figure 2-5.   Mean survival of Rhepoxynius abronis versus dissolved cadmium
             concentration after 4-day in seawater (symbols) and 0 and 4 days
             (bars) in interstitial water (this figure is from Swartz et al., 1985)	1-25
         \
Figure 2-6.   Mortality versus interstitial water cadmium activity for sediments
             from Long Island Sound, Ninigret Pond and a mixture of these
             two sediments (this figure is from Di Toro et al.,  1990). Water-
             only exposure data for Ampetisca dbdita and Rhepoxynius
             hudsoni.  The line is a joint fit to both water-only data sets   	1-26
• v.--.'
                                    Draft for SAB 1-5

-------
                                FIGURES (Continued)
                                                                                 Page
Figure 2-7.   Toxicity of copper to Hyalella azteca versus copper
             concentrations in a water-only exposure (open symbols) and
             interstitial water copper concentrations in sediment exposures
             (closed symbols) using Keweenaw Waterway sediments (this
             figure is from Ankley et al., 1993)  .....	1-27

Figure 2-8.   Cadmium titrations of amorphous FeS. X-axis is cadmium added
             normalized by FeS initially present (this figure is from Di Toro et
             al., 1990).  Y-axis is total dissolved cadmium.  The lines
             connecting the data points are an aid to visualizing the data	. . 1-31

Figure 2-9.   Concentrations of Fe2* and Cd2+ in supernatant from titration of .
             FeS by Cd2+ (personal communication with Di Toro 1992)   	.1-32

Figure 2-10.  Cadmium titration of sediments from Black Rock Harbor, Long
             Island Sound, Hudson River and Ninigret Pond (this figure is
             from Di Toro etal., 1990).  Cadmium added per unit dry weight
             of sediment versus dissolved cadmium	,	1-35

Figure 3-1.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) on a sum
 ;            total SEM basis	. . 1-43

Figure 3-2.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for
             CtVJmii'TP   	,	t	 1-45

Figure 3-3.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for
             copper  .	 1-46

Figure 3-4.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for nickel	1-47
                 '                                     ' '                  '
Figure 3-5.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for lead  	1-48

Figure 3-6.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for zinc  ....... 1-49
                                   Draft for SAB 1-6

-------
                                FIGURES (Continued)
Figure 3-7.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for
             amphipod bioassay test spiked with a four metal mixture	 1-50

Figure 3-8.   Mortality versus [SEM]-[AVS] difference (top panel) and this
             difference normalized to organic carbon (bottom panel) for silver  	1-51

Figure 3-9.   Mortality versus the average Total Metal/ERM quotient for field
             data (o) and laboratory spiked data (V)	1-52

Figure 3-10.  Concentrations of individual metals in interstitial water of
             sediments from Long Island Sound (top) and Ninigret Pond
             (bottom) in the mixed metals experiment as a function of
             SEM/AVS ratio (this figure is from Berry et al., 1996).
          ,   Concentrations below the IW detection limits, indicated by
             arrows, are plotted at one half the detection limit.  K^ is the
             sulfide solubility product constant	'.	 . 1-56

Figure 3-11.  Stability fields for chromium species (Richard and Bourg,  1991)	1-60

Figure 3-12.  (A) Stability fields for chromium species.  (B) Solubility of
             chromium hydroxide (Rai et al., 1987)	 1-61

Figure 3-13.  Titration of AVS with chromate.  The final chromate
             concentration in solution is plotted vs. the initial chromate
             concentration. The line  corresponds to a 1:1 stoichiometry for
             chromate reduction by FeS	:.....-...	1-62

Figure 3-14.  Oxidation of Cr(m) by MnO2.  Initial concentrations of MnO2
             indicated in the legend.  Open symbols are Cr(VI)
             concentrations, closed symbols are 2/3 Mn(H) concentrations.
             The lines are model computations	1-64

Figure 3-15.  Model of Cr(III) sorption to suspended solids and DOC. Varying
             initial concentrations of Cr(lH)T (see the stars plotted on the y-
             axis).  Plots of total and  dissolved chromium vs. total suspended
             solids.  Alternating filled and hatched symbols represent the data
             for each Cr(in)r. (A) Doc=9.4 mg/L. (B) DOC=24 mg/L.
             (C) An example of the computed speciation for the DOC=24
             mg/L and Cr(IH)T= 1.0 mg/L,case	1-66
                                   Draft for SAB 1-7

-------
                                FIGURES (Continued)
Figure 3-16.  Chromium cycling in the aquatic environment (Richard and
             Bourg, 1991)  . . .  ,	.  . ..	1 . .  .	.	
 Page

, 1-67
Figure 3-17.  Percentage mortality of saltwater and freshwater benthic species
             in 10-day toxicity tests in sediments spiked with individual metals
             (Ag, Cd, Gi, Ni, Pb, or Zn) or a metal mixture (Cd, Cu,  Ni and
             Zn).  Mortality is plotted as a function of: (a) the sum of the
             concentrations of cadmium, copper, lead, nickel and zinc in
             ,umoles metal per gram dry weight sediment; (b) interstitial water
             toxic units (silver data from Berry et al., 1999; all other data
            • modified after Berry et al., 1996); and (c) SEM/AVS ratio.
             Species tested include: the oligochaete (Lumbriculus variegatus),
             polychaetes (Capitella capitata aadNeanthes arenaceodentata),
             amphipods (Ampelisca abditaand Hyalella aztecd), harpacticoid
             copepod (Amphiascus tenuiremis) and gastropod (Helisoma sp.).
             Data below the SEM detection limit are plotted at
             SEM/AVS=0.01.  Data below the detection limit of metals hi
             interstitial water are plotted at IWTU=0.01	
 1-69
Figure 3-18.  Percentage mortality of saltwater and freshwater benthic species
             in 10-day toxicity tests in spiked sediments (open symbols) and
             sediments from field (closed sediments) (silver data from Berry et
           .  al., 1999; all other data modified after Hansen et al.,  1996a).
             Mortality is plotted as a function of: (a) the sum of the
             concentrations of cadmium, copper, lead, nickel and zinc in '
             yumoles metal per gram dry weight sediment; (b) interstitial water
             toxic units; and (c) SEM/AVS ratio. Species tested include: the
             oligochaete (Lumbriculus variegatus), polychaetes ( Capitella
             capitata and Neanthes arenaceodentata), amphipods (Ampelisca
             abditaand Hyalella azteca), harpacticoid copepod (Amphiascus
             tenuiremis), and the snail (Helisoma sp.).  Data below the SEM
             detection limit are plotted at SEM/AVS = 0.01. Data below the
             detection limit of metals in interstitial water are plotted at IWTU
             = 0.01	-'.  /	•'.	
 1-70
Figure 3-19.  Percentage mortality of amphipods, oligochaetes and polychaetes
             exposed to sediments from four freshwater and three saltwater
             field locations as a function of the sum of the molar
             concentrations of SEM minus the molar concentration of A VS
             (SEM-AVS) (from Hansen et al., 1996a): the vertical dashed line
             at SEM-AVS = 0.0 indicates the boundary between sulfide-
             bound unavailable metal and potentially available metal	
  1-80
                                   Draft for SAB1-8

-------
                              FIGURES (Continued)

                                                                            Page
Figure 4-1.   SEM minus AVS values versus AVS concentrations in EMAP-
            Great Lakes sediments from Lake Michigan.  Data are from
            surficial grab samples only (this figure is taken from Leonard et
            al., 1996, see data in Appendix C).  The upper plot shows all
            values, the lower plot has the ordinate limited to SEM minus
            AVS values between -10  and +10	1-96

Figure 4-2.   SEM minus AVS values versus AVS concentrations in EMAP-
            Estuaries Virginian Province (U.S. EPA, 1996), REMAP-
            NY/NJ Harbor Estuary (Adams et al., 1996) and the NOAA NST
            Long Island Sound (Wolfe et al., 1994), Boston Harbor (Long et
            al., 1996), and Hudson-Raritan Estuaries (Long et al., 1995);
            (see data in Appendix D for sources). The upper plot shows all
            values, and the lower plot has the ordinate limited to SEM minus  -
            AVS values between -10  and +10	 1-98
                                 Draft for SAB 1-9

-------
                                      SECTION 1
        ^
                                   INTRODUCTION

 1.1    GENERAL INFORMATION

       Under the Clean Water Act (CWA) the U.S. Environmental Protection Agency (U.S.
EPA) is responsible for protecting the chemical, physical and biological integrity of the
nation's waters.  In keeping with this responsibility, the U.S. EPA published ambient water
quality criteria (WQC) in 1980 for 64 of the 65 toxic pollutants or pollutant categories
designated as toxic hi the CWA. Additional water quality documents that update criteria for
selected consent decree chemicals and new criteria have been published since 1980. These
WQC are numerical concentration limits that are the U.S. EPA's best estimate of
concentrations protective of human health and the presence and uses of aquatic life. While
these WQC play an important role in assuring a healthy aquatic environment, they alone are
not sufficient to ensure the protection of environmental or human health.

     .  Toxic pollutants in bottom sediments of the nation's lakes, rivers, wetlands, estuaries
and marine coastal waters create the potential for continued environmental degradation even
where water-column concentrations comply with established human health and aquatic life
WQC. In addition, contaminated sediments can be a significant pollutant source that may
cause water quality degradation to persist, even when other pollutant sources are stopped (U.S.
EPA 1997a, b, c). The scarcity of defensible sediment guidelines and the single chemical
nature of those available make it difficult to: accurately assess the extent of the ecological
risks of contaminated sediments; to establish pollution prevention  strategies; and to identify,
prioritize and implement appropriate clean up activities and source controls.

       As a result of the need for a procedure to assist regulatory agencies in making decisions
concerning contaminated sediment problems and their prevention, a U.S. EPA Office of
Science and Technology and Office of Research and Development research team was
established to review alternative approaches (Chapman, 1987).  All of the approaches
reviewed had both strengths and weaknesses and no single approach was found to be
applicable for sediment guidelines derivation in all situations (U.S. EPA,  1989). The
equilibrium partitioning (EqP) approach was selected for nonionic organic chemicals because it
                                   Draft for SAB 1-10

-------
presented the greatest promise for generating defensible national chemical specific sediment
guidelines applicable across a broad range of sediment types.  The term EqP sediment
guidelines (ESG) refers to numerical concentrations for individual chemicals that are
applicable across the range of sediments encountered in practice.  The three principal
observations that established the EqP method of deriving sediment guidelines for nonionic
organic chemicals were:
                            f
       1.     The concentration of nonionic organic chemicals in sediments, expressed on an
              organic carbon basis, and in interstitial water, correlate to observed biological
              effects on sediment dwelling organisms across a range of sediments.

       2.     Partitioning models can relate sediment concentrations for nonionic organic
              chemicals on an organic carbon basis to freely dissolved concentrations in .
                                                                      \
              interstitial water.

       3.     The distribution of sensitivities of benthic and water column organisms to
              chemicals are similar, thus, the currently established WQC final chronic values
              (FCV) can be used to define the acceptable effects concentration of a chemical
              freely dissolved in interstitial water.

       Due to their wide-spread release and persistent nature, metals such as cadmium,
copper, lead, nickel, silver and zinc are commonly elevated in aquatic sediments.  These
metals are a potential aquatic environmental concern in addition to nonionic organic chemicals.
Thus, there have been various proposals for deriving sediment guidelines for protecting
benthic communities from metal toxicity. Many such attempts have featured measurement of
total sediment metals followed by comparison to background metal concentrations, or in some
cases an effects-based endpoint (Sullivan et al.,  1985; Persaud et al., 1989; Long and Morgan,
1991; Ingersoll et al., 1996; MacDonald et al.,  1996). An important limitation to these types
of approaches is that causality can not be established in part because of the procedures used to
derive correlative values, and because values derived are based on total rather than
bioavailable metal concentrations; i.e., for any given total metal concentration, adverse
toxicological effects may or may not occur, depending upon physico-chemical characteristics
of the sediment of concern (Tessier and Campbell, 1987; Luoma et al., 1989; Di Toro et al.,
1990).
                     '                     .}
                                    Draft for SAB 1-11

-------
       Many researchers have used elaborate sequential extraction procedures to identify
sedimentary physico-chemical fractions with which metals are associated in an attempt to
understand the biological availability of metals in sediments (Tessier et al., 1979; Luoma and
Bryan, 1981). Key binding phases for metals in sediments included iron and manganese
oxides and organic carbon. Shortcomings with these approaches have limited their application
largely to aerobic sediments instead of anaerobic sediments where metals are found in the
greatest concentrations (See Section 2).

       In developing ESG for metals that are causally-based and applicable across sediments,
it is essential that bioavailability be understood.  Different studies have shown that while total
(dry weight) metal concentrations in anaerobic sediments are not predictive of bioavailability,
metal concentrations in interstitial water are correlated with observed biological effects (Swartz
et al.,1985; Kemp and Swartz,  1986). However, as opposed to the situation for nonionic
organic chemicals and organic carbon (Di Toro et al., 1991), sediment partitioning phases
controlling interstitial water concentrations of metals .were not readily apparent.  A key
partitioning phase controlling cationic metal activity and metal-induced toxicity in the
sediment-interstitial water system is acid volatile sulfide (AVS) (Di Toro et al., 1990). Acid
volatile sulfide binds, on a molar basis, a number of cationic metals of environmental concern
(cadmium, copper, nickel, lead, silver and zinc) forming insoluble sulfide complexes with
minimal biological availability. (Hereafter in this document, the use of the term "metals" will
apply only to the six metals:  cadmium, copper, lead, nickel, silver and zinc.)
                                 1                                                    '
        The data that support the EqP approach for deriving sediment guidelines for nonionic
organic chemicals are reviewed by Di Toro et al. (1991) and U.S. EPA (1997a).  Recently,
EPA evaluated the potential utility of the EqP approach for deriving sediment guidelines for
metals (U.S. EPA, 1994a), which was reviewed by EPA's Science Advisory Board (U.S.
EPA, 1995a).  The data that support the EqP approach for deriving sediment guidelines for
metals presented in this document are taken largely from a series of papers published in the
December, 1996 issue of Environmental Toxicology and Chemistry by Ankley et al (1996);
Berry et al. (1996); DeWitt et al. (1996); Di Toro et al, (1990; 1992; 1996a,b) Hansen et al.
(1996a,b); Leonard et al. (1996a); Liber et al. (1996); Manony et al!  (1996); Peterson et al.
(1996); and Sibley et al. (1996). In addition, publications by Di Toro et al. (1990, 1992),
Ankley et al. (1994) and the U.S. EPA (1995a) were of particular importance in the
preparation of this document.
                                    Draft for SAB 1-12

-------
       The same three principals observed in applying the EqP approach to nonionic organic
chemicals listed above, also apply with only minor adjustments to deriving ESG for mixtures
of the cationic metals cadmium, copper, lead, nickel, silver and zinc.
                             i  •                         •     '          •
       1.     The concentrations of these six metals in sediments, normalized to the
              concentration of acid volatile sulfide (AVS) and simultaneously extracted metals
              (SEM; the metals extracted with AVS) in sediments, and dissolved in interstitial
              waters correlate to observed biological effects on sediment dwelling organisms
              across a range of sediments.

       2.   >  Partitioning models can relate sediment concentrations for divalent cationic
              metals (and silver) on an AVS basis to the absence of freely dissolved
             'concentrations in interstitial water.

       3.     The distribution of sensitivities of benthic and water column organisms to
              organic chemicals and metals are similar (U.S.  EPA,  1998a), thus, the currently
              established WQC final chronic values (FCV) can be used to define the
              acceptable effects concentration of the metals freely dissolved in  interstitial
              water.
                                     i   .
       The EqP approach, therefore, assumes that: (1) the partitioning of the metal between
sediment AVS (or any other binding factors controlling bioavailability) and interstitial water is
at equilibrium; (2) organisms receive equivalent exposure from interstitial water-only exposure
or from exposure to any other equilibrated sediment phase: either from interstitial water via
respiration, sediment via ingestion,  sediment-integument exchange, or from a mixtures of
exposure routes; (3) for the cationic metals cadmium, copper, lead, nickel, silver and zinc, no-
effects concentrations hi sediments can be predicted using the difference between the total
molar concentration of SEM for these metals and the total molar concentration of AVS. This
difference is the amount of either excess metal or excess AVS. So long as the molar
concentration of AVS equals or exceeds the sum of the  molar concentrations of these metals,
the sediment is not expected to cause acute or chronic toxicity in  benthic organisms; and (4)
the WQC FCV concentration is an appropriate effects concentration for freely dissolved metal
in interstitial water and the toxicity  of metals  in interstitial water  is no more than additive.
                                    Draft for SAB 1-13

-------
       Two equally applicable ESG for metals, a solid phase and an interstitial water phase,
        >                          .    '     •                                    '. ' '
are proposed.  For the first time, the Agency is publishing ESG that account for bioavailability
in sediments and potential for effects of a mixture in the aquatic environment. The mixtures
approach for these six metals will provide an ecologically relevant benchmark by resolving the
                                   .1   ,.          '              » '.
longstanding toxicological problem of their inter-dependent geochemistry.  The solid phase
ESG is defined as the S[SEM]-[AVS]<;0 (total molar concentration of simultaneously extracted
metal - total molar concentration of acid volatile sulfide is less than or equal to zero). Note
that cadmium, copper, lead and nickel are divalent metals so that one mole of each metal can
bind with one mole of AYS. The molar concentrations of these metals are compared to AYS
on a one to one basis. Silver however exists predominantly as a mpnovalent metal so that
silver monosulfide (Ag2S) binds two moles of silver for each mole of AYS.  Therefore SEMAg
by convention will be defined as the molar concentration of silver divided by two, [Ag]/2,
which is compared to the molar AYS concentration.  The interstitial water phase ESG is
£[Mitd]/[FCVi)d] £l (the sum over all of the six metals of the concentration of each individual
metal dissolved in the interstitial water/the metal-specific Final Chronic Value based on
dissolved metal is less than or equal to one).  This latter value is termed interstitial water
guidelines toxic units (IWGTUs). The IWGTUs approach by definition requires that the IW
metals are additive.  The data presented in this document support additiviry.

       Importantly, both the solid phase ESG and interstitial water ESG are no-effect
guidelines; i.e., they predict sediments that are acceptable for the protection of benthic
organisms,  these ESG, when exceeded, do not predict sediments that are unacceptable  for the
protection of benthic organisms. The solid phase (SEM-AYS) guideline avoids the
methodological difficulties of interstitial water sampling that may lead to an overestimate'of
exposure and provides information on the potential for additional metal binding.  The use of
both the solid phase and interstitial water guidelines will improve estimates  of risks of
sediment-associated metals.  For example, the absence of significant concentrations of metal in
interstitial water in toxic sediments having SEM-AYS  > 0 and in non-toxic  sediments having
SEM-AYS £ 0 demonstrates that metals in these sediments are unavailable.  Because of the
known spatial and temporal cycling of the important metal-binding phases in sediments,
Section 5 of this document provides implementation  guidance on sediment collection, handling
and analysis that will improve estimates  of risk.
                                   Draft for SAB 1-14

-------
       The ESG developed using the latest available scientific data are suitable for providing
guidance to regulatory agencies because they are:
                                                               .**
       1.     numeric values,
       2.  .   chemical specific,      -
       3.     causal,
       4.     applicable to most sediments, and
       5.     protective of benthic organisms.

.'  ,    It should be emphasized that these guidelines are intended to protect benthic organisms
from the direct effects of these six metals associated with sediments. ESG are intended to
apply to sediments permanently inundated with water, intertidal sediment, and to sediments
inundated periodically for durations sufficient to permit development of benthic assemblages.
They do not apply to occasionally inundated soils containing terrestrial organisms.  These
guidelines do not address the question of possible contamination of upper trophic level
organisms or the synefgistic, additive or antagonistic effects of other substances.  The ESG
presented in this document represent the U.S. EPA's best recommendation at this time of the
concentration of a mixture of cadmium, copper, lead, nickel, silver and zinc in sediment mat
will not adversely affect most benthic organisms. ESG values may be adjusted to account for
future data or site specific considerations  (U.S. EPA, 1998c).

       This document presents the theoretical basis and the supporting data relevant to the
derivation of the ESG for the metals cadmium, copper, lead, nickel, silver and zinc.  An  .
understanding of the "Guidelines for Deriving Numerical National Water Quality Criteria for
the Protection of Aquatic Organisms and Their Uses" (Stephan et al., 1985), response to
public comment (U.S. EPA, 1985a);  "Ambient Water Quality Criteria for Cadmium" (U.S.
EPA, 1985b); "Ambient Water Quality Criteria for Copper" (U.S. EPA, 1985c); "Ambient
Water Quality Criteria-Saltwater Copper Addendum" (U.S. EPA, 1995c); "Ambient Water
Quality Criteria for  Lead" (U.S.  EPA, 1985d); "Ambient Water Quality Criteria for Nickel"
(U.S. EPA, 1986); "Ambient Water Quality Criteria for Silver" (U.S. EPA, 1980); "Ambient
Water Quality Criteria for  Zinc" (U.S. EPA,  1987) is necessary in order to understand the
following text, tables and calculations. Guidance for the acceptable use of ESG values for
metals mixtures is contained in "Users Guide for Multi-Program Implementation of Sediment
Guidelines" (U.S. EPA, 1998b).

                                   Draft for SAB 1-15

-------
 1.2    OVERVIEW OF DOCUMENT

        Section 1: "Introduction" provides a brief review of the EqP methodology as it applies
 to the individual metals cadmium copper, lead, nickel, silver and zinc and their mixtures.
 Section 2: " Partitioning of Metals in Sediments" reviews published experimental results that
 describe the partitioning and bioavailability of these metals in freshwater and marine
 sediments, the role of AVS, SEM and interstitial water concentrations of metals. Section 3:
 "Toxicity of Metals in Sediments" reviews the results of acute and chronic toxicity tests
 conducted with spiked and field sediments that demonstrate that the partitioning and
 bioavailability of metals in sediments can be used to accurately predict the toxicity of
 sediment-associated metals.  Section 4: "Derivation of Sediment Guidelines for Metals"
 describes the SEM-AVS and interstitial water guidelines toxic unit approaches for the
 derivation of the ESG for individual metals and mixtures of metals. Published WQC values
 for dissolved metal for five of these six metals (the silver FCV for freshwater is not available)
 are summarized for use in the Interstitial Water Guidelines Toxic Units (IWTU) Procedure.
 The ESG for metals  is then compared to chemical monitoring data on the environmental
 occurrence of metals and AVS in sediments from Lake Michigan, the Virginian Province from
.EPA's Environmental Monitoring and Assessment Program (EMAP) and NOAA's National.
 Status and Trends. Section 5:  "Implementation" describes procedures for collection, handling,
 analysis of sediments and interpretation of data from sediment samples required if the
 assessments of the risks of sediment-associated metals are to be accurate.  Section 6:
 "Guidelines Statement" concludes with EPA's guidelines statement for the metals cadmium,
 copper, nickel, lead, silver and zinc.  The references cited in this document are listed in
 Section?.
                                   Draft for SAB 1-16

-------
                                      SECTION!

                     PARTITIONING OF METALS IN SEDIMENTS

2.1    METAL TOXICITY IN WATER-ONLY AND IN INTERSTITIAL WATER OF
       SEDIMENT EXPOSURES

       The equilibrium partitioning methodology for establishing sediment guidelines (ESG)
requires that the chemical concentration be measured in the bioavailable phase and that the
chemical potential of the chemical be determined. This section-demonstrates that biological
effects correlate to metal activity.  Secondly it demonstrates that biological response is the
same for water-only exposures and for sediment exposures using the interstitial water
concentrations. Therefore, for both metals and nonionic chemicals this fundamental tenant of
                                                     A                        ,
the equilibrium partitioning model is satisfied.

       A direct approach, to establishing sediment guidelines for metals would be to apply the
water quality criteria final  chronic values to measured interstitial water concentrations.  The
validity of this approach depends on both the degree to which interstitial water concentration
represents free metal activity, and whether or not free metal activity can be accurately
measured in both systems.   For most metals, free metal activity'cannot be measured at water
quality criteria concentrations. Therefore, present water quality criteria are not based on free
metal activity. The problem with this is that many metals readily bind to dissolved (actually
colloidal) organic carbon (DOC), and DOC complexes do not appear to be bioavailable
(Bergman and Dorward-King, 1997). Hence guidelines based on interstitial water
concentrations of metals may be overly protective with the direct use of the concentration of
metals  in interstitial water.

       By implication this  difficulty extends to any complexing ligand that is present in
sufficient quantity. The decay of sediment organic matter can cause substantial changes in
interstitial water chemistry. In particular, bicarbonate increases due to sulfate reduction.  This
increases the importance of the metal-carbonate complexes and further complicates the
question of the bioavailable metal species (Stumm and Morgan,  1996).

       The sampling of sediment interstitial water for metals is not a routine procedure. The

                                   Draft for SAB 1-17

-------
least invasive technique employs a diffusion sampler which has cavities covered with a filter
membrane (Hesslein, 1976; Carignan, 1984; Carignan et al., 1985; Allen et al., 1993; Bufflap
and Allen, 1995).  The sampler is inserted into the sediment and the concentrations on either
side of the membrane equilibrate!  When the sampler is removed the cavities contain filtered
interstitial water samples. Since the sampler is removed after equilibration, the concentrations
of metals inside the sampler should be equal to the concentrations of freely dissolved metals in
the interstitial water.  The time required for equilibration, which usually exceeds one day,
depends on the size of the filter membrane and the geometry of the cavity.

       An alternate technique to separate the interstitial water is to obtain a sediment core,
slice it, filter or centrifuge the slice, and then filter the resultant interstitial water twice. For
anaerobic sediments this must be done in a nitrogen atmosphere to prevent the precipitation of
iron hydroxide which would scavenge the metals and yield artificially low dissolved
concentrations (Troup, 1974; Allen et al.. 1993).

       Although either of these techniques are suitable for research investigations, they require
more than the normally available sampling capabilities.  If solid phase chemical measurements
were available from which interstitial water metal activity could be deduced, it would obviate
the need for interstitial water sampling and analysis, circumvent the need to deal with
complexing ligands, and provide fundamental insight into metal binding phases in sediments
needed to predict bioavailabiliry.

2.1.1  Toxicity correlates to metal activity
                                                                    i
       A substantial number of water-only exposure experiments discussed below point to the
fact that biological effects can be correlated to the divalent metal activity {M2*}. The claim is
not that the only bioavailable form is M2* - for example MOH+ may also be bioavailable - but
that the DOC and certain other ligand-complexed fractions are not bioavailable.

       The acute toxicity of cadmium to  grass shrimp (Palaemonetes pugio) has been
determined at various concentrations of chloride (as salinity) and the complexing ligand NTA,
both of which form cadmium complexes  (Sunda et al., 1978).  The concentration response .
curves as a function of total cadmium (Figure 2-1, top panels) are quite'different at varying
concentrations of chloride, indexed by salinity, and NTA. However, if the organism response
                                    Draft for SAB 1-18

-------
                                                                     m
                                                                     a
                                                                     o
P«r

5S
D e

i§£  r


§ll "
O »-»
    «
'U.  §
 O  w
95o
XXu
QCQU.

"-wffi
UWW

c<
DC
oo
    o

    §



    1
    LL
                             (0
             o
             o
                  • o
                   in
                               \
                               o
                               o
                                  2

                                 «!
                                  a

                                  2
»•
j *
i •
•

*l^ >
v •
- \
» t\.
• ^k.
---^^, *
• ^~ "Nk
X
.\

fj
- .' •
f^a..^*!^
"? T«OO B
V»?ECM04
O
£<«•»••*
It f ^ 1 *
Mnt 4Kb • *—

o
I"* +"
M
S
"S
fc "
>
u
<
oi
0E
Q
<
U
n
8
^S „•
=5?2
5 2
S **
-»*»
l|
•&1
fa
.E §
I*
2.2
i £
P
S s
.1s5
                                                'O
                                                
*i *^ nt
        o
        o
                               •o
                               o

                               §
                              Q
                                         o  co
                                                                  tJ'
                                                                  O

                                                                •  *£


                                                                  f
                                                                  >

                                                                  o
                                                                  <


                                                                pi
                                                               O
                                                 to
                                                            o
          HVAIAynS lN30H3d
                                    Draft for SAB 1-19

-------
is evaluated with respect to Cd2+ activity in the solution then the data become a single
concentration-response relationship (bottom panels). Comparable results have been reported
for copper-EDTA complexes (Anderson and Morel, 1978) for which concentration-response
correlates to Cu2+ activity (Figure 2-2, left top and bottom).

       When the concentration of zinc is held constant and the concentration of the
complexing ligand NTA is varied, the effect on growth of the phytoplankter, Microcystis
aeruginosa, decreases as NTA added increases (Figure 2-2, right top and bottom; Allen et al.,
1980). The cell density decreases rather than increases in time and reaches control levels at
the highest NTA concentration (left top and bottom panels). The data can all be correlated to
free zinc activity as shown (right top and bottom panels).  Similar results for diatoms exposed
to copper and the complexing ligand Tris (Figure 2-3, top; Sunda and Guillard, 1976).
Variations in Tris concentrations and pH produce markedly different growth rates (left top and
bottom) which can all be correlated to the Cu2+ activity (right bottom). A similar set of results
have been obtained by Sunda and Lewis (1978) with DOC from river water as the complexing
ligand (Figure 2-3, right top and bottom).

       Metal bioavailability as measured by metal accumulation into tissues of organisms has
also been examined (Zamuda and Sunda, 1982). Uptake of copper by oysters is correlated not
to total copper concentration (Figure 2-4, top) but to copper activity (bottom).

:      The implication to be drawn from these experiments is that the partitioning model
required for establishing sediment guidelines should predict dissolved metal in interstitial
water. The following subsection examines the utility of this idea.

2.1.2  Toxicity correlates to interstitial water concentration

       This subsection presents some early data that first indicated the equivalence of
interstitial water concentrations and water-only exposures. Much more data of this sort are   .
presented in Section 3 of this document. Swartz et al.(1985) tested the acute toxicity of
cadmium to the marine amphipod Rhepoxynius abronius, in sediment and seawater. An
objective of the study was to determine the contributions of interstitial and particle-bound
cadmium to toxicity. A comparison of the 4-day LC50 of cadmium in interstitial water (1.42
mg/L) with the 4-day LC50 of cadmium in seawater without sediment (1.61 mg/L) resulted in
                                    Draft for SAB 1-20

-------
(0
111
o
I
I-
IU
o
K
UJ
0.
      ACUTE TOXICITY OF COPPER
         TO A DINOFLAGELLATE
     (FROM ANDERSON AND MOREL. 1972)
100
 90
 80
 70
 60
 SO
 40
 30
 20
 10
.   o
                  VWTHOUT-EDTA
           5    6789
        TOTAL COPPER (-LOG (Cuj))
                                          CHRONIC TOXICITY OF ZINC
                                        ON M1CROCYSTIS AERUGINOSA
                                             (FROM ALLEN, et. at.. 1530)
                                          10
                                                             o AAP v»rt:l M F«:£OTA vrta T.M
                                                                •»*.3»W7 MZn
                                                                •1.0x10-7MNTA
                                                                » 2.5 > 10-7 M NT*
                                                                * 40x10-7*1 NTA
                                                                « 3.0* 10-7 M NTA
                                                                 S.S«1fl'*MNTA
                                                                 t.3 » 10-5 M NTA
 ii
 o
  o
  UJ
  o
100
 90
 80
 70
 60
 50
 40
 30
 20
  10
   0
          I    10   11   12   13   14
           COPPER ACTIVITY ( p Cu )
                                        ta .
                                     Ul
                                     o
                                     u.
                                     O
                                     I  io4
                                           103
                                                              K1 ttpH7.Q
                                                           O - EDTA   12.4S
                                                           i -NTA    7.5*
                                                           X -OOS  '  6.31
                                                           O -CMOS  5.2S
                                                           • - Builder M 4.33
                                                           * -Control
                                          0  1.0  2.0  3.0  4.0  .5.0
                                             FREE ZINC (moles/liter x 107)
   2-2. Acute toxicity to a dinoflagelate (left) of total copper (top) and copper activity
   (bottom), with and without the complexing ligand EDTA (this portion of the figure is
   from Anderson and Morel, 1978). Toxicity of zinc toMicrocystisaeruginosa (right)
   showing growth as cells/ml versus time with different levels of the complexing ligands
   EDTA and NTA (top) and number of cells at five days as a function of free zinc
   concentration (bottom) (this portion of the figure is from Allen et al., 1980).
                               Draft for SAB 1-21

-------
1= £
£S •?
01
Oi
151
f-o §
IS 6
o
OP
ICITY OF

DIATOM


ANDGUILURD,
o


o
cr

u
    °
         555

         211
           I
       s!

               •H^N
               i  \ i

                  -<.
                           S
                             0

                           Is
                           -
                             g
                                       •



                                       V
ill

ill
                 *  "^--

                            U
                                       H1MOMO OldlOidS
                             Draft for SAB 1-22

-------
                   UPTAKE OF COPPER BY OYSTERS
       o  o
       U  o)
       0.  w
       o-~
       U
             150
             100
              50
                0
                0.01
        f
         zs
         go
         a *
         s 1
         P
         fc
         8
         U
         *
         A
    0.1          1.0
TOTAL COPPER CONC. (|IM)
10.0
200
180
160
140
120
100
80
60
40
2
TOTAL NTA (MM)
Winter
A 1.0
0 10.0
t Summer
1 ' - *02 •
\ \ A 1.0


1
. • J 	
: +- •
: ^
                            9        10        11
                          COPPER ACTIVITY (pCu)
Figure 2-4.  Body burdens of copper in oysters (Crassostrea virginica) versus total
copper (top) and copper activity (bottom) with different levels of the complexing Ugand
NTA (this figure is from Zamuda and Sunda, 1982).
                           Draft for SAB 1-23

-------
no significant difference between the two (Figure 2-5).

       Experiments were performed to determine the role of acid volatile sulfides in cadmium
spiked sediments using the amphipods Ampelisca abdita and Rhepoxynius hudsoni (Di Toro et
al., 1990). Three sediments were used, a Long Island Sound sediment with high AYS, a
Ninigret Pond sediment with low AVS concentration, and a 50/50 mixture of the two
sediments. Figure 2-6 presents a comparison of the observed mortality in the three sediments
to the interstitial water cadmium activity measured with a specific ion electrode.  Four-day
water-only and 10-day exposure sediment toxicity tests were perfonned.  The water-only
response data for Ampelisca and Rhepoxynius are included for comparison although they
represent a shorter duration exposure. These experiments also demonstrate the equivalence of
organism response to metal concentrations in interstitial water and in water-only exposures.

       An elegant experimental design was employed by Kemp and Swartz (1986) to examine
the relative acute toxicity of particle bound and dissolved interstitial cadmium. They
circulated water of the same cadmium concentration through different sediments.  This
resulted hi different bulk sediment concentrations, but the same interstitial water
concentrations. They found no statistically significant difference in organism response for the
different sediments.  Since the interstitial water concentrations were the same in each treatment
-the circulating water concentrations established the interstitial water concentrations - these
experiments confirmed the hypothesis of equal response to concentrations in water-only and
interstitial water.

       A series of 10-day toxicity tests using the amphipod Hyalella azteca were perfonned to
evaluate the bioavailability of copper hi sediments from two sites highly contaminated with this
metal: Steilacoom Lake, Washington and Keweenaw Watershed, Michigan (Ankley et al.,
1993). A water-only, 10-day copper toxicity test was also conducted with the same organism.
The mortality resulting from the water-only test was strikingly similar to that from the
Keweenaw sediment tests when related to interstitial water (Figure 2-7).  The LC50s show
strong agreement for the water-only (31  jug/L) and the Keweenaw sediment test (28 jug/L),
using the average of day 0 and day 10 interstitial water concentrations. Steilacoom Lake 10-
day interstitial water concentrations were less than the 7 ,ug/L detection limit and were
    f                              /             .          '
consistent with the observed lack of toxicity to H, azteca (Ankley et al, 1993).
                                   Draft for SAB 1-24

-------
       20
        15
   O)

   111
   5
   0)
   •-*
   2
   IU
   K
                       COMPARISON OF WATER AND
                          SEDIMENT EXPOSURE
                        (AFTER R.C. SWARTZetal., 1985)
                                                SEAWTER

                                                INTERSTITIAL
                                                WATER
                                          t»'0
                        DISSOLVED CADMIUM CONC. (mg/L)
Figure 2-5. Rhepoxynius abronis mean survival versus dissolved cadmium for 4-day
toxidty tests in seawater (symbols) and interstitial water at time 0 and 4 days (bars) (this
ficure is from Swartz et al  1985V
                           Draft for SAB 1-25

-------
 O
 or
 IJJ

 I
 -4
(0
a:
UJ
a:
O
             (%)Ainviyow
II
                  Draft for SAB 1-26

-------
fi
o
5
   100
     80
60
 o
 H
 0)
0.
40
     20
            • Water Only Exposure
            O Sediment Exposure
                                                      O   Q\
                _1	I   I   I	I
                                  10
                                                      100
1000
                                        Copper (ug/L)
       Figure 2-7. Toxicity of copper to Hyalella azteca versus copper concentrations in a
       water-only exposure (open symbols) and interstitial water copper concentrations in
       sediment exposures (closed symbols) using Keweenaw Waterway sediments (this figure is
       from Ankley et aL, 1993).
                                   Draft for SAB 1-27

-------
       The data presented in this subsection, and data to be presented in Section 3 of this
document, demonstrate that in water-only exposures metal activity and concentration can be
used to predict toxicity.  The results of the four experiments above demonstrate that mortality
data from water-only exposures can be used to predict sediment toxicity using interstitial water
concentrations. Therefore, the metal activity or concentration in interstitial water would be an
important component of a partitioning model needed to establish sediment guidelines. The
solid metal-binding phases of this partitioning model need to be identified.  The following
subsection presents data that identifies solid phase sulfides as the important metal-binding
phase.

2.2    SOLID PHASE SULFIDE AS THE IMPORTANT BINDING COMPONENT

       Modeling metal sorption to oxides in laboratory systems is well developed, and detailed
models are available for cation and anion sorption, see the articles in Stunun (1987) and
Dzombak and Morel (1990) for recent summaries. The models consider surface complexation
reactions as well as electrical interactions via models of die double layer.  Models for natural
soil and sediment particles are less well developed.  However, recent studies suggest that
similar models can be applied to soil systems (Allen et al., 1980; Barrow and Ellis, 1986a,b,c;
Sposito et al., 1988). Since the ability to predict partition coefficients is required if interstitial
water metal concentration is to be inferred from the total concentration, some practical model .
is required.  This subsection presents state of the science in the theoretical development of
metals partitioning in sediments.

2.2.1   Metal Sorption Phases

       The initial difficulty that one confronts in selecting an applicable sorption model  is that
the. available models are quite complex and many of the parameter estimates may be specific to
individual soils or sediments. However, the success of organic carbon based non-ionic
chemical sorption models suggests that some model of intermediate complexity that is based on
an identification of the sorption phases may be more generally applicable.

       A start in this direction has been presented (Jenne et al., 1986; Di Toro et al., 1987).
The basic idea was that instead of considering only one sorption phase as is assumed for  '
non-ionic hydrophobic chemical sorption, multiple sorption phases must be considered.  The
                                   Draft for SAB 1-28

-------
conventional view of metals speciation in aerobic soils and sediments is that metals are
associated with the exchangeable, carbonate and Fe and Mn oxide forms, as well as organic
matter, stable metal sulfides, and a residual phase.  In oxic soils and freshwater sediments,
sorption phases have been identified as paniculate organic carbon (POC) and the oxides of
iron and manganese (Jenne, 1968, 1977; Oakley et al., 1980; Luoma and Bryan, 1981).
These phases are important because they have a large sorptive capacity. Further they appear
as coatings on the particles and occlude the other mineral components. It was thought that
they provided the primary sites for sorption of metals. These ideas have been applied to metal
speciation in sediments. However, they ignore.the critical importance of metal sulfide
interactions which dominate speciation hi the anaerobic layers of the sediment.

2.2.2  Titration Experiments
                                                                             i

       The importance of sulfide in the control of metal concentrations in the interstitial water
of marine sediments is well documented (Boulegue et al., 1982; Emerson et al., 1983; Davies-
Colley et al., 198S; Morse et al., 1987).  Metal sulfides are very insoluble and the equilibrium
interstitial water metal concentrations in the presence of sulfides are small.   If the interstitial
water sulfide, S, concentration in sediments is large, then the addition of metal, M,  to the
sediment, metal sulfide, MS; would precipitate following the reaction:

         M2t  + S2' - MS(s)                        '                             (2-1)
This appeared to be happening during a spiked cadmium sediment toxicity test (Di Toro et al.,
1990) since a visible bright yellow cadmium sulfide precipitate formed as cadmium was added
to the sediment. However, interstitial water sulfide activity, {S2'}, measured with a sulfide
electrode indicated that there was little or no free sulfide hi the unspiked sediment. This was,
at the time, a most puzzling result.

       The lack of significant quantity of dissolved sulfide in the interstitial water and the
evident formation of solid phase cadmium sulfide suggested the following possibility. The
majority of the sulfide in sediments is in the form of solid phase iron sulfides. Perhaps the
source of the sulfide is this solid phase sulfide initially present.  As cadmium is added to the

                                    Draft for SAB 1-29

-------
sediment it causes the solid phase iron sulfide to dissolve releasing sulfide which is available
for the formation of cadmium sulfide.  The reaction is:             '
         Cd2* + FeS(s) - CdS(s)  + Fe24
(2-2)
Cadmium titrations with amorphous FeS and with sediments were performed to examine this
possibility.                                                             •

2.2.2.1 Amorphous FeS
                                                                       ;
       A direct test of the extent to which this reaction takes place was performed (Di Toro et
al., 1990). A quantity of freshly precipitated iron sulfide was titrated by adding dissolved
cadmium. The resulting aqueous cadmium activity, measured with the cadmium electrode
versus the ratio of cadmium added, [Cd]A, to the amount of FeS initially present, [FeS(s)]j, is
shown in Figure 2-8.  The plot of dissolved cadmium versus cadmium added illustrates the
increase in dissolved cadmium that occurs near [Cd]A / [FeS(s)]r = 1. A similar experiment
has been performed for amorphous MnS with comparable results. It is interesting to note that t
these displacement reactions'among metal sulfides have been observed by other investigators
(Phillips and Kraus, 1965).  The reaction  was also postulated by Pankow (1979) to explain an
experimental result involving copper and synthetic FeS.

       These experiments plainly demonstrate that solid phase amorphous iron and manganese
sulfide can be readily displaced by adding cadmium. -As a consequence, it is the source of
available sulfide which must be taken into account when evaluating the relationship between
solid phase and aqueous phase cadmium in sediments.

       A direct confirmation that the removal of cadmium was via the displacement of iron
sulfide is shown in Figure 2-9, The supernatant from a titration of FeS by Cd2+ was analyzed
for both cadmium and iron; The solid lines are the theoretical expectation based on the
stoichiometry of the reaction (Equation 2-9) (DiToro et al., 1990).       ,   .

2.2.2.2 Sediments

       A similar titration procedure has been used to evaluate the behavior of sediments taken
                                   Draft for SAB 1-30

-------
         "S



         s
         D
         Ul


         o
         w
         
-------
0)
u.
tb
0
mm
°E
ll
** "•»
4>
U ,
O
o
600
475
350
225
100-
-25!
                    RESULTS OF FeS + Cd** = CdS + Fe**

                            Concentration of Fe**
                               Analysis of Filtrate
                             Cd**/FeS Molar Ratio
s
•s
^ __
icentratioi
fppmj
o .
700
600
500
400
300-
200-
100-
0;
        O
                                               f
                           Concentration of Cd**
                              Analysis of Filtrate
0.2 0.4  0.6  0.8   1   1.2 1.4 1.6

       Cd**/FeS Molar Ratio
                                                       1.8
Figure 2-9. Concentrations of Fe2* and Cd2+ in supernatant from titration of FeS by Cd2*
(personal communication with DiToro 1992).
                           Draft for SAB 1-32

-------
from four quite different marine environments: sediments from Black Rock Harbor, the
Hudson River, and the sediments from Long Island Sound and Ninigret Pond used in the  .
toxicity tests. The binding capacity for cadmium is estimated by extrapolating a straight line
fit to the dissolved cadmium data. The equation is:
              x     '       .    .   '          '                                 •
         [SCd(aq)] =max{0,m([Cd]A -,[Cd]B)}                                   (2-3)

                                                               /
where [SCd(aq)] is the total dissolved cadmium, [Cd]A is the cadmium added, [Cd]B is the
bound cadmium, and m is the slope of the straight line. The sediments exhibit quite different
binding capacities for cadmium, listed in Table 2-1, ranging from approximately 1 jtmol/gm to
more than 100 pmol/g.  the question is whether this binding capacity is explained by the solid
phase sulfide present in the samples.

       Table 2-1. Cadmium binding capacity and AVS of sediments (from Di Toro et
                                      al., 1990).
Sediment
Black Rock Harbor
Hudson River
LI Sound(c>
Mixture(c>
Ninigret Pohd(c'd)
Initial AVS
Otmol/g)w
175(41)
12.6(2.80)
15.9 (3.30)
5.45 (-)
2.34 (0.73)
Final AVS Cd Binding Capacity
Otmol/g)^ (Mmol/g)
-
•
13.9 (6.43)
3.23(1.18)
0.28(0.12)
114(12)
8.58 (2.95)
4.57 (2.52)
-
1.12(0.42)
       (a)Average (Standard Deviation) AVS of repeated measurements of the stock.
                                         i
       ^Average (Standard Deviation) AVS after the sediment toxicity experiment.
       From original cadmium experiment (Di Toro et al., 1990).
       (d)50/50 mixture of LI Sound and Ninigret Pond.                           .
       
-------
labile fraction, acid volatile sulfide (AVS), is associated with the more soluble iron and
manganese monosulfides.  The more resistant sulfide mineral phase, iron pyrite, is not soluble
in the cold acid extraction used to measure AVS. Neither is the third compartment, organic
sulfide which is associated with the organic matter in sediments (Landers et al., 1983).
       «"
       The possibility that acid volatile sulfide is a direct measure of the solid phase sulfide
that reacts with cadmium is examined in Table  2-1, which lists the sediment binding capacity
for cadmium and the measured AVS for each sediment, and in Figure 2-10, which indicates
the initial AVS concentration. The sediment cadmium binding capacity appears to be
somewhat less than the initial AVS for the sediments tested. However a comparison between
the initial AVS of the sediments and that remaining after the cadmium titration is completed,
Table 2-1, suggests that some AVS is lost during titration experiment. In any case, the
covariation of sediment binding capacity and AVS is clear. This suggests that AVS is the
proper quantification of the solid phase sulfides that can be dissolved by cadmium.  The
chemical basis for this is examined below.

2.2.4  Solubility Relationships and Displacement Reactions

       Iron monosulfide, FeS(s), is in equilibrium with aqueous phase sulfide and iron
concentration via the reaction:
                     2* ^ e 2-'
FeS(s) - Fe2*  + S
                                                                                 (2-4)
If cadmium is added to the aqueous phase, the result is:
           12+
Cd2+ + FeS(s) ~
2+  + S
                                  2-
                                                                                 (2-5)
As the cadmium concentration increases, [Cd2+][S2'] will exceed the solubility product of
cadmium sulfide and CdS(s) will start to form. Since cadmium sulfide is more insoluble than
iron monosulfide, FeS(s) should start to dissolve in response to the lowered sulfide
concentration in the interstitial water.  The overall reaction is:
Cd2* + FeS(s) - CdS(s) + Fe2+
                                                                                  (2-6)
                                   Draft for SAB 1-34

-------
(0

z
UJ
a
UJ

CD
CD
i I 	
^t CM
0 CD
•
•
C
c
h/6iu)
                             QHAIOSSIQ
                      Draft for SAB 1-35

-------
The iron in FeS(s) is displaced by cadmium to form soluble iron and solid cadmium sulfide,
CdS(s). The consequence of this replacement reaction can be seen using an analysis of the
M(ID-Fe(H)-S(-H) system with both MS(s) and FeS(s) present in Appendix B. M(II)
represents any divalent metal that forms a sulfide that is more insoluble than FeS.  If the added
metal, [M]A, is less than the AYS present in the sediment then the ratio of metal activity to
total metal in the sediment-interstitial water system is less than the ratio of the MS to FeS
solubility  products:                     -
r2+
          {M2+}/[M]
                                eS
                                                                       (2-7)
This is a general result that is independent of the details of the interstitial water chemistry.  In
particular it is independent of the Fe2* activity. Of course the actual value of the ratio
(M2+}/[M]A depends on aqueous speciation, as indicated by Equation 2-6. However, the ratio
is still less than the ratio of the sulfide solubility products.

      : This is an important finding since the data presented in Section 2.1.1 indicates that
toxicity is related to metal activity, {M2*}. This inequality guarantees that the metal activity,
in contrast to the total dissolved metal concentration, is regulated by the iron sulfide - metal
sulfide system.

       The sulfide solubility products and the ratios are listed in Table 2-2.  The ratio of
cadmium activity to total cadmium is less than 10 ~10 **.  For nickel the ratio is less than W~559.
By inference this reduction in metal activity will occur for any other metal that forms a sulfide
that is significantly more insoluble than iron monosulfide. The ratios for the other metals in
Table 2-2, Zn, Cd, Pb, Cu and Ag, indicate that metal activity for these metals will be very
small in the presence of excess AVS.          .
                                    Draft forSAB 1-36

-------
                       Table 2-2. Metal sulfide solubility products.
Metal Sulfide
FeS
NiS
ZnS
CdS
PbS
CuS
Ag2S
log
K«P,2
-3.64
-9.23
-9.64
-14.10
-14.67
-22.19
-36.14
log
K*.
-22.39
-27.98
-28.39 -
-32.85
-33,42
-40.94
-54.71
Log
(Kws/KFeS)

-5,59
-6.00
-10.46
-11.03
-18.55
-32.32
•Solubility products, K^2> for the reaction NP+ + HS' - MS(s) + H+ for CdS (greenockite),
FeS (mackinawite), and NiS (millerite) from Emerson et al., 1983.  Solubility products for
CuS (covellite), PbS (galena), ZnS (wurtzite), and Ag2S (acanthite) and pK2 = 18.57 for the
reaction HS-  ~ H+  +„ S2" from Schoonen, and Barnes, 1988.  K^, for the reaction M2*  + S2' -
MS(s) is computed from log K^ *uid PK2-

2.2.5  Application to Mixtures of Metals

       A conjecture based on the sulfide solubility products for the metals listed in Table 2-2
is that the sum of the molar concentrations of metals should be compared to AVS. Since all
these metals have lower sulfide solubility parameters than FeS, they would all exist as metal
sulfides if their molar sum (assume [Ag]/2 because it is monovalent) is less than the AVS.  For
this case:
          [AVS]
< 1
                                                                                  (2-8)
no metal toxicity would be expected where [MT]i is the total cold acid extractable ith metal
molar concentration in the sediment (divide by 2 for silver). On the other hand, if their molar
sum is greater than the AVS concentration, then a portion of the metals with the largest sulfide
solubility parameters would exist as free metal and potentially cause toxicity. For this case the
following would be true:
                                   Draft for SAB 1-37

-------
          [AVS]
                                                                                  (2-9)
These two equations are precisely the formulas that one would employ to determine the extent
of metal toxicity in sediments assuming additive behavior and neglecting the effect of
partitioning to other sediment phases. Whether the normalized sum is less than or greater than
1.0 discriminates between non-toxic and potentially toxic sediments.  The additivity does not
come from the nature of the mechanism that causes toxicity. Rather, it results from the equal
ability of the metals to form metal sulfides with the same stoichiometric ratio of M and S.

       The appropriate quantity of metals to use in the metals/A VS ratio is referred to as
"simultaneously extracted metal" or SEM.  This is the metal which is extracted in the cold
acid used in the AVS procedure. This is the appropriate quantity to use because some metals
form sulfides which are not labile in the AVS extraction (e.g., nickel, copper).  If a more
rigorous extraction were used to increase the fraction of metal extracted which did not also
capture the additional sulfide extracted, then the sulfide associated with the additional metal
release would  not be quantified. This would result in an erroneously  high metal to AVS ratio
(DiToroetal., 1992).

       The above discussion is predicated on  the assumption that all the metal sulfides behave
similarly to cadmium sulfide.  Further it has been assumed that only acid soluble metals are
reactive enough to affect the free metal activity. That is, the proper metal concentration to be
used is the SEM.  Both of these hypotheses were tested directly with benthic organisms using
sediment toxicity tests.  Results of these sediment spiking experiments with cadmium, copper,
lead, nickel, silver, zinc and a mixture of these will be presented in Section 3 which follows.
                                    Draft for SAB 1-38

-------
                  \  '  •               SECTIONS

                       TOXICITY OF METALS IN SEDIMENTS

3.1    GENERAL INFORMATION             ,    •
                         ....                    '
       This Section summarizes data from acute and chronic toxicity tests which demonstrate
that the absence of sediment toxicity due to metals can be predicted by use of the interstitial
water concentrations^ metals or by comparison of the molar concentrations of AVS and
SEM.  This ability to predict .the toxicity of metals in sediments, through a fundamental
understanding of their bioavailability, is in sharp contrast with the absence of a causal
correlation between metals induced sediment, metal-induced toxicity,. and metals
concentrations based on the dry weight of sediment (Berry et al.,  1996; Di Toro et al., 1990;
1992; Luoma, 1989, Berry et al., 1999) and the use of sequential extractions (Ankley et al.,
1996).

3.2    PREDICTING THE TOXICITY OF METALS IN SEDIMENTS

   1    The SEM-AVS method for evaluating the toxicity of metals (Di Toro, 1990; Di Toro,
1992) has proven to be quite successful at predicting the lack of toxicity in spiked and field
contaminated sediments (Berry et al., 1996; Hansen et al., 1996).  However, it does not
                                                                      •>
appear to be able to predict very well the onset of toxicity in a sequence of sediments spiked
with increasing concentrations of metals, or for sets of field contaminated sediments In fact,
in a recent article (Long et al., 1998) it was claimed that the empirically derived methods
(Long et al., 1995) are better at making this prediction.  The purpose of the note is to
introduce a modification of the SEM - AVS procedure that significantly improves the
prediction of mortality.  Additionally, the inability of the empirically derived methods to make
this prediction is demonstrated.

       The Equilibrium Partitioning (EqP) model gives the prescription for the development of
sediment concentrations that predict the toxicity or lack of toxicity in sediments. The sediment
concentration C'^ that corresponds to a measured LC50 in a water only exposure of the test
organism is
                                  Draft for SAB 1-39"

-------
                           CSed = KPCw                                        (1)
where C*^ is the sediment concentration (ug/g dry wt.), KP (L/kg) is the partition coefficient
between pore water and sediment solids, and C*w is the LC50 concentration (ug/L) (Di Toro
et al., 1991). For application to metals that react with AVS to form insoluble metal sulfides
(Di Toro et al., 1990), this formula becomes

                           CM = AVS + KPCW                     '          £)
where AVS is the sediment concentration of acid volatile sulfides. The formula simply states
that since AVS can bind the metal as essentially insoluble sulfides, the concentration of metal
in a sediment that will cause toxicity is at least as great as the A VS. that is present. The
sediment metal concentration that should be used is the SEM concentration since any metal
that is bound so strongly that IN hydrochloric acid cannot dissolve it is not likely to be
bioavailable (Di Toro et al.,  1992). Of course, this argument is just speculation, which  is why
so much effort has been expended to demonstrate experimentally that this is actually the case
(Di Toro et al., 1991, Di Toro et al., 1992, Berry et al., 1996 and Hansen et al., 1996).
Therefore eq.(2) becomes                                                           .
                                                  PW     '                     (3)
The basis for the SEM/ AVS method is to observe that if the second term in eq.(3) is neglected
then the critical concentration is
                           SEM =        AVS
or the criteria for toxicity or lack of toxicity is
or, equivalently
SEM/AVS = 1
 *•
SEM -AVS = 0
(4)


(5)

(6)
       The failure of die either the ratio condition (eq.5) or the difference condition (eq.6) to
predict toxicity is due to the neglect of the partitioning term Kp C*w- Note that ignoring the
term does not affect the prediction of lack of toxicity since it makes the condition conservative
(i.e. smaller concentrations of SEM are at the boundary of toxicity and no toxicity).

       The key to improving the prediction of toxicity is to approximate the partitioning term
                                    Draft for SAB 1-40

-------
rather than ignoring it. In anaerobic sediments, the organic carbon fraction is an important
partitioning phase and partition coefficients for certain metals at certain pHs have been
measured (Mahony et al., 1996).  This suggests that the partition coefficient Kp in eq.(3) can
be expressed using the organic carbon based partition coefficient, KOC, together with the
fraction organic carbon in the sediment,
                           Kp = KQC foe       ,    '                              (7)

Using this expression in eq.(3) yields

                           SEM =       AVS + K
                                                                                (8)

Then moving the known terms to the left-hand side of the equation yields

                           (SEM- AVS)/ ^  = KocCv                         (9)
If we knew both KOC md C*w we could use eq.(9) to predict toxicity.  The method evaluated
below uses (SEM -  AVS)/ foe as the predictor of toxicity and evaluates the critical
concentrations (the right hand side of eq.9) based on observed SEM, AVS, foe, and toxicity  .
data.

       Data from toxicity tests using both laboratory-spiked and field-collected sediments were
compiled from the literature.  Three sources of laboratory spiked tests using marine sediments
were included, Casas and Crecelius (1994) and  Berry et al.  (1996, 1999).   Data reported
included total metals, simultaneously extracted metals (SEM), acid volatile sulfide (AVS), organic
carbon fraction (f^) and 10 day mortality.  In the study by Casas and Crecelius (1994) , the
toxicity of zinc, lead and copper were tested on the marine polychaete Capitella capitata.  In
Berry et al. (1996)  the toxicity of cadmium, copper, lead, nickel, zinc and a mixture of four
metals (cadmium, copper, nickel and zinc) were tested on the marine amphipod Ampelisca abdita.
In Berry et al. (1999) the toxicity of silver to A.  abdita was tested.  Three sources for metal
contaminated field sediments were included, Hansen et al. (1996), Kemble et al. (1994), and Call
et al. (1999).  Data reported included total metals, SEM, AVS, fraction organic  carbon and
mortality. The metals included in total SEM were cadmium, copper, nickel, lead and zinc.  In
Hansen et al. (1996), data were reported for five saltwater and four freshwater locations. For the
                                                         v
                                   Draft for SAB 1-41

-------
freshwater stations, organic carbon and total metals data were not provided. Organic carbon data
for  one  location,  Keweenaw  Watershed,  were  obtained  separately  [Berry,  personal
communication]. Ten day bioassays were conducted using A. abdita and a freshwater amphipod,
Hyalella azteca. In Kemble et al. (1994)  14-day bioassays were conducted on Chironomus
riparius, a freshwater midge. Data included from Call et al. (1999) were the control freshwater
sediment data with 10-day mortality to the midge, Chironomus tertians.

       Laboratory spiked and field-contaminated sediment data were grouped together,and
analyzed as one data set. Mortality data were compared against the SEM-AVS difference and this
difference normalized to the fraction organic carbon, (SEM-AVS)/^. For each comparison two
bounds were computed for the SEM - AVS comparison and the (SEM - AVS)/ f^: a lower bound
concentration equivalent to a 95 percent chance that the mortality observed would be less than 50
percent and an upper bound concentration equivalent to a 95 percent chance that the observed
mortality would be greater than 50 percent.  The lower bound limit was computed by evaluating
the fraction of correct classification starting from the lowest abscissa value. When the fraction
correct dropped to below 95%, the 95* percentile was interpolated.  The same procedure was
applied to obtain the upper bound. Upper and lower bounds were calculated for both SEM-AVS
difference and the organic carbon normalized difference.

       Using the same data set, individual total metal concentrations were divided by the effects
range median (ERM) values of Long et al. (1995) for that metal or metals and the mean quotient
value was determined.  This was compared to observed mortality. Upper and lower bounds were
calculated using the same method as described above.

       Mortality in the laboratory spiked and field-contaminated sediment tests were organism
independent when plotted against the SEM-AVS difference (Figure 3-1, top panel).  The lower
and upper boundaries determined for the SEM-AVS difference were  1.7  and 190 jumol/g
respectively and are shown on the top panel of Figure 3-1. A line indicating 50 percent mortality
is also  shown. These boundaries can be used as predictive limits. For SEM-AVS values lower
than 1.7 //mol/g, there is a 95 percent chance that the observed mortality will be less than 50
percent. Similarly, for SEM-AVS values greater than 190 //mol/g, there is a 95 percent chance
that the observed mortality will be greater than 50 percent.  The uncertainty falls in the range of
SEM-AVS values between the upper and lower bounds, which in this case are approximately two
orders  of magnitude wide.  When normalized to the organic carbon content of the sediment, the
                                   Draft for SAB 1-42

-------
        120
        100
         80
         <°
         20
o Field Amphlpod

v Reid Midge

• Lab Amphlpod

OufaPolychaete
                                    • • o
                                          o  •
                                       a       o
              <»V V
J   v     " o
 I  „   v
   v   •  n  <
    •   •   o
          ami    aoi     0.1
                          10     100     1000   10000
                              SEM-AVS(umoI/g)
         120
         too
          SO
     £.   60

     2?

          •
          20
          •20
                                       0   •
                            >--•-

            0.1
            10     100    1000    10000   100000  1000000
                             SEM-AVS/f oc (umol/goc)



Figure 3-1. Mortality versus [SEM]-[AVS] difference (top panel) and this difference

normalized to organic carbon (bottom panel) on a sum total SEM basis.
                            Draft for SAB 1-43

-------
 range of uncertainty is narrowed to approximately one order of magnitude (Figure 3-1, bottom
panel). The lower and upper boundaries are 86 and 2000 ^mol/g,,,. respectively.

       This same analysis is shown on a metal specific basis for cadmium, copper, nickel, lead
and zinc in Figures 3-2 through 3-6 together with the upper and lower boundaries determined
above.  It appears that for each metal, the boundaries could be adjusted slightly to encompass a
narrower range in uncertainty.  However, these metals all appear to be behaving similarly in this
analysis. In fact, this analysis was tested using the four metal mixture experiment (cadmium,
copper, lead  and nickel) and the area of uncertainty fell within the boundaries (Figure 3-7)
suggesting that these metals behave similarly and it is appropriate to analyze them together.

       A metal which appears to be more toxic than the other five metals is silver (Figure 3-8).
Mortality was observed at much lower SEM-AVS values. Upper and lower boundaries could not
be .computed  for silver due to lack of data at low SEM-AVS values that had an absence of
mortality.

       Long et al. (1998) have suggested that sediment quality guidelines (SQGs) based on dry
weight normalizations were equally if not more accurate, in predicting the non-toxicity or toxicity
of sediment associated metals than AVS-normalized SEM concentrations.  It was concluded that
using the mean ERM  quotient (metal concentration/ERM) was more effective at predicting
whether or not a sample was toxic compared to SEM/AVS ratios. The data set from Hansen et
al. (1996) was used.

       We re-examine this question using the complete data set described above, which includes
spiked as well as field contaminated sediments. Upper and lower bounds corresponding to 95%
correct predictions for the mean ERM quotient were determined as described above. The results
are shown hi  Figure 3-9. The lower and upper bounds were 0.24 and 270 respectively.  The
range of uncertainty is three orders of magnitude.   By contrast, the range of uncertainty for the
SEM-AVS difference was two orders of magnitude, indicating that the AVS normalized SEM
concentrations were more accurate at predicting  the absence or presence of metal-associated
toxicity.  However,  normalizing these concentrations to the sediment organic carbon content
lowered the range of uncertainty to close to one order of magnitude, therefore providing the most
reliable method of predicting toxicity or non-toxicity of sediment associated metals.
                                   Draft for SAB 1-44

-------
                                CADMIUM
    o
120

100

eo
60

40

20

0


•VI
	 ' 	 ' 	 •
o Field Amphipod
V Reid Midge
• Lab Amphipod
DLabPolychaete


•

, v
" ' o
•

'




4
0



s
0
f °
I
I
I
1. . .'.....i 	 4
. ,....,, 1 1 M.I
•
-
•
O , .
.







•


          0.001    0.01     0.1      1      10     100    1000    10000

                              SEM-AVS(umol/g)




g
'»
i
^




120
100

80
60

40

20

0
-20
C
. • 1 T 1

1

""



* •
, - ' ' o

.01 0.1 1 10



0
'

• • •
s
o
o
ll 	 .1
100 100C

.« . -
•
o
m






1 10000 100
                             SEM-AVS/foc (umol/goc)
                       y         .
Figure 3-2.  Mortality versus [SEM]-[AVS] difference (top panel) and this difference
normalized to organic carbon (bottom panel) for cadmium.
                           Draft for SAB 1-45

-------
£
            120
            100
            BO
            60
            40
            20
           0 Field Amphlpod
           ^ Field Mldg»
           •Lab Amphlpod
           OUbPolychaete
                                     COPPER
                                      *  • mo OH1
                                          0   O
                                           O
                                             O
           ISO
           100
           so
           60
           40
                   V


                   V
                                  °0
                           0.1      1      10
                                                 100     1000   10000
                                SEM-AVS(iimotfg)
                            1 e  a a »d
                                 o
                                  o
                              o
                              o

                              o
                                                   • *
                                                     o
                                                      100000  1000000
                             SEM-AVS/foc (umol/goc)
Figure 3-3. Mortality versus [SEM]-[AVS] difference (top panel) and this difference
normalized to organic carbon (bottom panel) for copper.
                            Draft for SAB 1-46

-------
                                  NICKEL
     o
120

100

80
80

40


20

0
jtn

° Reid Amphlpod
V Reid Midge
• Lab Amphipod
OubPolychaete

.______..-.___
'
' mm '

V
0°
*
J . ,.J „.!
' ' """1 — • t«"l"l — »

• t •

o
.
_ - _ - «Q. —
•
1
o
o
o
o

1 	 .1 . 	 .J
-' ''""1 — ' 	 «
•
• *
.
o
"

U — — — — — .






t . ......I . , 	
             0.1
           OJM1    0.01     0.1      1       10      100     1000    10000

                               SEM-AVS (umol/g)


I
i



100
BO
60
40
20
0
-30



'
V
V
0°
'

• *
0

° 0
o
o


.....
o


• .


10     100     1000    10000   100000  1000000

   SEM-AVS/foc (umol/goc)
Figure 3-4. Mortality versus [SEM]-[AVSJ difference (top panel) and this difference
normalized to organic carbon (bottom panel) for nickel.
                            Draft for SAB \A1

-------
                                      LEAD
      i
100
80
60
40

20
0
0.0
' ORaldAmphlpod
• VReld Midge
• Lab Amphipod
OtabPolychaete


V
• . v
. o
m
. . .....j . ......j 	 ..j . f 	 j 1
MI 0,001 0.01 ai i
CJ
•
*
9

V . o

IB ««««
-
•





                                                               tooo
                                SEM-AVS (umol/g)
           120
           too
           80
      £   60
           40
           20
          -20 ' ' '****
                                              Q   •
                                                 •
            0.01     0.1
                                  10
                                         100    1000    10000   100000
                              SEM-AVS/foc (umol/goc)


Figure 3-5  Mortality versus [SEM]-[AVS] difference (top panel) and this difference
normalized to organic carbon (bottom panel) for lead.
                            Draft for SAB 1-48

-------
                                    ZINC
     €
     o
12U

100

80
60

40

M


0
.9ft

o Field Amphlpod
V Field Midge
• Lab Amphlpod
OLabPolychaete .


.______„ — — _..
V* v
v V
0
' « O O
' '
, . .....J . . ..»J . . ,[«.J


• a •

•
^*
.
.. _ _ _ «,^, _
o
'
V v
•
V°3
*
1. . .....j . 	 	

•
O
.
-
'
"





-

,
           0.001    0.01      0.1
10     100     1000    10000
                                SEM-AVS(umol/g)
120
100
BO
& 60
*
1 40
S '
20


0
.9n


•i



•V • ^

O

"


a
/
^
„ ' 	 3
o
•
v
*
'Gp? o

li 	

0 • • -

,

	






             0.1      1      10     100     1000    10000   100000  1000000

                               SEM-AVS/foc (umol/goc)

Figure 3-6. Mortality versus [SEM]-[AVS] difference (top pane!) and this difference
normalized to organic carbon (bottom panel) for zinc.
                            Draft for SAB 1-49

-------
120
100
80
Mortality (%)
S 5 8
0
•20
MIXTURE - LAB AMPHIPOD TF=RT
• \
• '
-
<
• t
,. . .....,, P , 1 Ml*!—
* • * ••***

1 ' '
,
,
1 * .
i
i
.
•
•
•
_
.
            oiooi    0.01     0.1       i
                                           10      100     1000
                                 SEM-AVS(umol/g)
           120
           100
           «0
     ^
      €

      I
           20
           •20
             0.1
                            10      100     1060    10000   100000  1000000

                              SEM-AVS/foc (umol/goc)
Figure 3-7. Mortality versus [SEM]-[AVS] difference (top panel) and this difference
normalized to organic carbon (bottom panel) for an amphipod bioassay test spiked with
four metal mixture.
a
                            Draft for SAB 1-50

-------
                         SILVER -LAB AMPHIPOD TEST
12U
100
80
SS. 60
1 "
20
0
-«W1
• * * <


'
1
*
•
: ' '".
•



'

-
-




           OJHM1   0.001    0.01    0.1
10     100    1000
                               SEM-AVS (umol/g)
          120
          100
           80
           40
           20
           -20
             0.01     0.1     1      10      100     1000   10000   100000

                             SEM-AVS/foc (umol/goc)


Figure 3-8. Mortality versus [SEM]-[AVS] difference (top panel) and this difference
normalized to organic carbon (bottom panel) for silver.
                           Draft for SAB 1-51

-------
      e
             0.001    OU»     0.1      1       10     100     1000    10000

                                  Average Metal/ERM
Figure 3-9. Mortality versus the average Total Metal/ERM quotient for field data (o) and
laboratory spiked data (V).
                              Draft for SAB 1-52

-------
       The use of the organic carbon normalized SEM-AVS as a predictor of toxicity reduces the
 uncertainty of the prediction from three orders of magnitude using average ERM ratios, and two
 orders of magnitude using SEM-AVS, to one order of magnitude using (SEM-AVS)/ f^. There
 appears to be no basis for the claim that average ERM ratios are preferable.

 3.3    PREDICTING METAL TOXICITY:  SHORT-TERM STUDIES

 3.3.1  Spiked sediments: Individual experiments

       A key to understanding  the bidavailability of sediment-associated  contaminants  was
 provided by Adams et al. (1985) who observed that the effects  of kepone, a nonionic organic
 pesticide, were  similar  across  sediments  when toxicity was related to  interstitial water
 concentrations.   Swartz et al.  (1985) and Kemp and  Swartz (1986) first observed that metal
 concentrations in interstitial waters of different sediments are correlated with observed biological
 effects. However, as opposed to the situation for nonionic organic chemicals and organic carbon
 (Di Toro et al., 1991), the sediment partitioning phases  that controlled  interstitial water
 concentrations of metals and metal-induced sediment toxicity were not known.

       Di Toro et al. (1990) first investigated the significance of sulfide partitioning in controlling
 metal bioavailability and metal-induced toxicity in marine sediments spiked with cadmium.  In
 these experiments, the operational definition of Cornwell'and Morse (1987) was used to identify
 that fraction of amorphous sulfide, or AVS, available to interact with cadmium in the sediments.
 Specifically, the  AVS was defined as the sulfide liberated from wet  sediment by treatment with'
 IN HC1 acid. Di Toro et al. (1990) found that when the molar concentration of AVS in the test
 sediments was larger than that of the molar concentration of total cadmium (i.e., when the
cadmium: AVS ratio was less than 1, or the difference between cadmium and AVS difference was
 less than 0), interstitial water concentrations of the metal were small and no toxicity was observed
 in 10-d tests with the amphipods Rhepoxynius httdsoni or Ampelisca abdita.  Studies by Carlson
et al. (1991) with cadmium-spiked freshwater sediments yielded similar results; when there was
more  AVS than total cadmium, significant toxicity was not observed in 10-d tests with an
 oligochaetes (Lumbriculus variegatus) or snail (Helisoma sp).  Based upon these initial studies,
another study with nickel-spiked sediments using A. abdita and field sediments contaminated with
 cadmium and nickel using  the freshwater amphipod  Hyalella azteca, Di Toro et al.  (1992)
provided further support to the importance of AVS in controlling metal bioavailability in

                                  Draft for SAB 1-53

-------
sediments. These studies suggested that it may be feasible to derive ESG for metals by direct
comparison of molar AVS concentrations to the molar sum of the concentrations of cationic metals
(specifically, cadmium, copper, nickel, lead and zinc) extracted with the AVS; i.e., £SEM.  They
observed that the expression of metals concentrations based on the £SEM is required because a
significant amount of nickel sulfide is not completely soluble in the AVS extraction. Hence, AVS
must be used as the measure of reactive sulfide and £SEM as the measure of total reactive metal.

       Casas and Crecelius (1994) further explored the relationship of SEM and AVS, interstitial
water concentrations, and toxicity by conducting 10-d toxicity tests with the marine polychaete
Capitella capitata exposed to sediments spiked with zinc, lead and copper. As was true in earlier
studies, elevated  interstitial  water metal concentrations were observed only when  SEM
concentrations exceeded those of AVS. Sediments were not toxic when SEM concentrations were
less than AVS and when the concentration in interstitial water were less than the water-only LC50.
Green et al. (1993) reported results of another spiking experiment supporting the general EqP
approach to deriving sediment guidelines for metals. In their study, metal-sulfide partitioning was
not directly quantified but it was found that toxicity of cadmium-spiked marine sediments to the
meiobenthic copepod Amphiascus tenuiremis was predictable based upon interstitial water (but not
sediment dry wt) cadmium concentrations. Further spiking experiments by Pesch et al. (1995)
demonstrated that 10-d survival of the marine polychaete Neanthes  arenceodentata was,
comparable to controls  in cadmium- or nickel-spiked sediments with more AVS than SEM.

 ;      Berry et al.(1996) described experiments in which A.  abdita was exposed for 10 days to
two or three sediments  spiked either singly,  or hi combination, with cadmium, copper, nickel,
lead and zinc.  As in previous studies, significant toxicity to the amphipod did not occur  when
AVS concentrations exceeded those of SEM. They compared observed mortality to interstitial
water metal concentrations expressed as toxic units (IWTU):          >
         IWTU=[MJ/LC50
(3-1)
where
[Mj  is the dissolved metal concentration in the interstitial water, and  the  LC50 is  the
concentration of the metal causing 50% mortality of the test species in a water-only .test.  If
                                   Draft for SAB 1-54

-------
 interstitial water exposure in a sediment test is indeed equivalent to that in a water-only test, then
 1.0 IWTU should result in 50% mortality of the test animals.  Berry et al. (1996) reported that
 significant (>24%) mortality of the saltwater amphipod occurred hi only 3.0% of sediments with
 less than 0.5JWTU, while samples with greater than 0.5 IWTU were toxic 94 A% of the time.
 Berry  et al.  (1996) also  made an important observation relative  to interstitial water metal
 chemistry in then* mixed-metals test.  Chemical equilibrium calculations suggest that the relative
 affinity of metals for AVS should be silver> copper > lead > cadmium >zinc> nickel (Emerson
 et al.,  1983; Di Tore et al.,  1992). Hence, the appearance of the metals in  interstitial water as
 AVS  is "exhausted" should occur in an inverse order (e.g., zinc  would  replace nickel in a
 monosulfide complex and nickel would be liberated to the interstitial water, etc). Berry et al!
 (1996) observed this trend hi sediments spiked with cadmium, copper , nickel and zinc (Figure
 3-10).  Furthermore, an increase in the concentration of a metal in a sediment with a low sulfide
 solubility product constant (K^) theoretically would displace a previously unavailable and non-
 toxic metal with a higher K^ making that metal available to bind to other  sediment phases or
 enter interstitial water to become toxic.  Berry et al. (1999) also exposed the  saltwater amphipod
Ampelisca abdita to sediments spiked with silver. When AVS was detected hi the sediments they
 were not toxic and interstitial water contained no detectable silver.   For sediments that contain
 no detectable AVS,  any SEM silver that is detected is dissolved interstitial silver, because silver
 sulfide and silver chloride precipitate are not extracted using the standard AVS procedure.

 3.3.2   Spiked sediments: Chromium experiments

       Toxicity tests were conducted with the amphipod (Ampelisca abdita) exposed to sediment
spiked with chromium VI as potassium dichromate (Berry and Boothman, 1999). Tests examined
the lexicological implications of  the reduction of toxic  chromium VI to  insoluble nontoxic
chromium  ffi (Kaczynski and Kleber,  1994) in  anaerobic  sediments.   This reaction  is
environmentally significant  because  (1)  chromium in  is insoluble-and nontoxic  hi typical
 freshwater and marine  sediments that are anaerobic, contain measurable concentrations of acid
volatile sulfide and have interstitial water pHs of from about 6.5 to 11.5; (2) chromium III once
formed, does not oxidize to chromium VI; (3) soluble and toxic chromium VI can only occur in
sediments with no detectable AVS; and (4) only chromium VI will occur dissolved hi interstitial
water (See section 3.3.2.1 of this document).
                                   Draft for SAB 1-55

-------
             JW Metal vs SEM/AVS: SW Mixed Metals - US
 10000
£  1000-1
XX

S   100
<1>

*     10
      1
1
.2
I    0.1
    0.01
                  LOG
                 JCsp_

              Nl .27.98


              Zn -28.39
               Cu  .40.94
           Nl— »

           Cu— »
           Zn— *
           Cd— 4
       0.01
                     0.1
1          10

 SEM/AVS
                                                        100
1000
  10000
o
    1000-1
o
o
 O
       100


        10


          1
 g    0.1
     0.01
                IW Metal vs SEM/AVS: SW Mixed Metals - NIN
           Nl.
        0.01
                      0.1
  1           10

   SEM/AVS
                                                          100
  1000
 Figure 3-10. Concentrations of individual metals in interstitial water of sediments from
 Long Island Sound (top) and Ninigret Pond (bottom) in the mixed metals experiment as a
 function of SEM/AVS ratio (this figure is from Berry et al., 1996). Concentrations below
 the IW detection limits, indicated by arrows, are plotted at one half the detection limit.

 is the sulfide solubility product constant.            -   ,

                            Draft for £45.1-56

-------
        Ten day toxicity tests with A. abdita were conducted by Berry and Boothman (1999,
" personal communication) using methods described by Berry et al. (1996). Sandy sediments from
 saltwater Ninigret Pond, RI with 1.7 ^umole AVS/g and about 0.15% TOC that were spiked to
 achieve nominal chromium concentrations of 11 to 520 yug/g. The pH of the spiking solution was
 adjusted with sodium hydroxide to 7.6 prior to addition to the sediment.  Sediment interstitial
 water pH ranged between 7.3 (Control) and 8.2 (highest concentration). A modification of the
 Cranston and Murray (1978) method was used to determine concentrations of total chromium and
 chromium VI in interstitial water.  Chromium VI in SEM was determined using the Wang et al.
 (1997) modification of the method of Cranston and Murray (1978).

       In the sediments that were not toxic to  A. abdita, all of the chromium in the SEM was
 present as chromium HI; concentration ranged from 0.24 to 2.08 ^moles/g.  Sediments with 3.73
 and 6.54 /umoles total chromium/g were lethal (Table 3-1).  Chromium VI was found in SEM of
 the two toxic sediments with concentrations of 1.01 and 4.08 //moles chromium Vl/g. Average
 measured total chromium concentrations were  not appreciably  different  from nominal
 concentrations. AVS concentrations hi nontoxic sediments ranged from 0.64 to 1.42 Ambles/ g
 sediment, but AVS was not detected hi toxic sediments.  The concentrations of AVS exceeded
 those of SEM hi two of four nontoxic sediments and chromium VI was absent from the interstitial
 waters of nontoxic sediments.  IWTUs of chromium VI explained observed toxicity: those
 sediments hi which chromium was not detectable hi the interstitial water were not toxic, those
 with greater than one toxic unit were toxic.

       These results indicate that chromium in sediments is not toxic if AVS is present and that
IWTUs can  be used  to identify nontoxic sediments.   Conversely,  sediments  containing
concentrations of SEM chromium when AVS is not present or when IWTUs are of lexicological
significance might be toxic. This approach does not apply to unique sediments having interstitial
water pHs of less than about 6.5 or greater than about 11.5 that may occur in water bodies with
low water column pHs or near acid mine drainage.

3.3.2.1      Chromium Chemistry hi Sediments
                                  i
       The purpose of this section is to present a review of the aqueous chemistry of
chromium as it influences its toxicity in sediments. The speciation chemistry of chromium is
                                .   Draft for SAB 1-57

-------
 •I

 i
 s
 a
2
I


1

!

-t

1
«
«M
O
2
rn
42

3
             II
                «"
              o
              U

              73

              S
              .
            50 J>
          jao
          V3
              U


              d.
       •a
                       o   oo   5(.  "!
                                                 o\
                       88  8   8  8  *   oo
                       O   o  d   ef.  «  o   ">p
                       VI   VI
                                    v   vi
                                                 ''*
                                             p   oo
                                             cs   so
                                        _
                                        C-  ^-
                           §   §  i   3
                           ^*   O  CS   ^


                           —•   ~  o*   o
                                            o  o
                          ^   vt  oo   oo   m  ^>

                          o   o  ^   oj  . m'  5
                                        O   O  O
                                        —   ^r  ts
                      f  2   5
                          o   o  • —   ci
                                  Draft for SAB 1-58

-------
  presented in Figure 3-11, and E^- pH diagram (Richard and Bourg 1991). The principle species
 that exist at the indicated E* (redox potential) and pH is listed.  Chromium exists in two oxidation
 states in natural systems.  Oxidized Cr(VI) species are present as oxyanions CrO42' and HCrO4'.
 The reduced Cr(III) species are Cr3+ and the hydroxide complexes CrOH2*, Cr(OH)2+, Cr(OH)3°
. and Cr(QH)4'. Although chromium sulfide minerals are known to exit, e.g. Cr2S3(s) (Wells 1962)
 and Cr3S4 (Brezinaite) (Vaughan and Craig 1978) they hydrolyze to chromium hydroxide upon
 exposure to water (Durrant and Durrani 1970).

        Cr(m) concentrations are regulated by an insoluble hydroxide Cr(OH)3(s) in the pH range
 from 6 to 11 (the shaded region in Figure 3-11).  The actual concentrations can be seen in Figure
 3-12 from (Rai, Sass et al. 1987). Solubility limits the dissolved concentration to approximately
 10'7 M in the pH range of approximately 6-7  to 10.  The water quality criterion for Cr(III) is
 approximately lO* M (74 pg Cr/L) (EPA 1984).  Therefore in the pH range of approximately 6-7
 to. 10 we expect no toxicity since the  interstitial water toxic unit concentration is approximately
 10'7 M/10* M = 0.1. By contrast to Cr(m), the oxidized Cr(VI) species are soluble and are also
 toxic. The EPA chronic water quality criteria for Cr(VI) are 11 and 50 i*g Cr/L in freshwater and
 saltwater, respectively.

        The reduction of Cr(VI) to Cr(ID) can occur only in reducing environments (Figure 3-11).
 The reduction can occur by bacterial action (Ohtake and Hardoyo 1992; Blake, Choate et a!. 1993;
 Schmieman, Petersen et al. 1997; Wang and Shen 1997), by ferrous iron (Eary and Rai 1989;
 Fendorf and Li 1996; Buerge and Hug 1997), by organic matter (Elovitz and Fish 1995; Wittbrodt
 and Palmer 1995; Deng and Stone 1996), photochemically (Hug, Laubscher et al. 1997), in soils
 (Eary and Rai  1991), in aquifer material (Anderson, Kent et al. .1994), by hydrogen sulfide
 (Smillie, Hunter et al. 1981; Pettine, Barra et al. 1998), pyrite fines (Zouboulis, Kydros et al.
 1995) and amorphous iron sulfide (Patterson, Fendorf et al. 1997). We have also examined the
 reduction of chromate using amorphous iron sulfide hi neutral solutions in support of the effort
 to develop an EqP based sediment criteria for chromium by the EPA.

        For the data presented in Figure3-13, the stoichiometry of the reaction appears to be

        CrCM^ + FeS + 3H2O - Cr(OH)3(s) + FeOOH(s) + S° + 2OH'             (1)

 i.e. one mole of chromate reduced per mole of FeS initially present. We have also seen titrations

                                   Draft for SAB 1-59

-------
                                   6    8  10   12   14
                                   PH
Figure 3-11. Stability fields for chromium species (Richard and Bourg, 1991).
                            Draft for SAB 1-60

-------
o
-1

-2

-3
 (•
-4

-5

-6

-7

•8,

-9
 O   6-8 d
 V 18-22 d
- 0 63-67 d
 D   134 d
                                                    WQC
                       o  o
                               Cr(OH)°
                 I    t
         3   45   6   7   89   10

                                 PH
                                   11   12   13  14  15
                       Draft for'SAB 1-61

-------
             2000
                    Initial AVS  - 475 umol/L
                          500        1000       15QO

                             Initial CiO42~ (jimol/L)
2000
Figure 3-13.  Titration of AVS with chromate. The final chromate concentration in
solution is plotted vs. the initial chromate concentration. The tine corresponds to a 1:1
stoichiometry for chromate reduction by FeS.
                             Draft for SAB 1-62

-------
 that correspond to complete oxidation of the sulflde to sulfate when sediments are used. The
important fact, from the point of view, of sediment criteria, is that if any FeS  is present in
sediments, then all the chromate would have  been reduced to chromium hydroxide by the
reduction reaction (eq.l).

       The reverse reaction: the oxidation of Cr(IH) to Cr(VI), is  of direct concern because
Cr(VI) is soluble and toxic. The oxidation rate of Cr(III) to Cr(VI) with oxygen as the only
oxidant is quite slow with .half-lives of months reported (see (Pettine and Millero 1990) for a
review). Oxidation can occur rapidly (in minutes) with hydrogen peroxide as the oxidant (I*ettine
and Millero 1990) leading to the suggestion that the photochemical production of H202 in surface
waters can be producing Cr(VI). For sediments, however, the likely oxidant is manganese
dioxide, which has been shown to oxidize Cr(HI) to Cr(VI) rapidly  (Schroeder and Lee 1975;
Eary and Rai 1987; Johnson and Xyla 1991; Fendorf and Zasoski 1992; Silvester, Charlet et al.
1995).

       At acid pHs the stoichiometry of the reaction appears  to be (Takacs 1988)
             3 MnO2(s) + 2 Cr(OH)2* - 3 Mn2+ + 2HCKV
                                                     (2)
which goes to completion very quickly (minutes to hours). We have modeled the reported data
for pH = 4.5 assuming that the oxidation rate is 1/2 order with respect to Cr3*, which suggests
a surface mass transfer limitation, and saturation kinetics for MnO2.  The kinetic equation for the
concentration, Cr, of Cr(ni) is
dCr
 dt
                                     MnO1
                                  MnO2+,KUHO)
                                                    (3)
where Cr and MnO2 are both functions of time, k is the reaction rate constant and KMn02 is the half
saturation constant. The concentration of MnO2 is determined by mass balance based on the
stoichiometric eq.(2).  The data for experiments at varying MnO2(s) concentrations are shown in
Figure 3-14.  Both the Cr(VI) and Mn(H) produced by the oxidation are shown.  The Mn(H)
concentrations are multiplied by 2/3 from eq.(2) so. that their concentrations should be the same
as the Cr(VI). Other data are available for pH = 3-5 which indicates  that the oxidation rate
increases with pH (Silvester, Charlet et al. 1995).                     .
                                   Draft for SAB 1-63

-------
      •  O 4.8 uM
      T  V 9.1 uM
O  29 uM
T  290 uM
         10    20    30    40    50
              Time (min)
10   20.  30   40   50
      Time (min)
Figure 3-14. Oxidation of Cr(III) by MnO2. Initial concentrations of MnO2 indicated in
the legend.  Open symbols are Cr(VI) concentrations, closed symbols are 2/3 Mn(II)
concentrations. The lines are model computations.
                            Draft for SAB 1-64

-------
       The situation at neutral pHs is less clear.  Since Cr(III) forms in insoluble Cr(OH)3(s), the
issue is whether chromium hydroxide can be oxidized by Mn02(s). Takacs (1988) reports a slow
rate of oxidation from a single experiment. However Johnson and Xyla (Johnson and Xyla 1991)
report that the rate is Independent of pH.  It is clear that further experimentation is required to
settle this important issue.

       The sorption of Cr(III) and Cr(VI) are important reactions that limit the bioavailability of
chrome to organisms in both the water column and sediment. In addition, complexed forms of
chrome can diffuse from the pore water of the sediment to the overlying water. The sorption of
Cr(VI) to hydrous iron oxide has been studied extensively and the surface complexation model
has been shown to apply (Dzombak and Morel 1990; Mesuere and Fish 1992; Mesuere and Fish
1992).

       The modeling of the  sorption of Cr(UI) is complicated by the precipitation kinetics of
Cr(OH)3.  Studies of Cr(HJ) sorption to silica have been carried out at low pHs (Fendorf, Lamble
et al. 1994;  Fendorf and Sparks  1994). We have analyzed a set of sorption data generated at
neutral pH (Masscheleyn, Pardue et al. 1992) with a partitioning model. Varying amounts of DOC
and suspended matter were added to solutions of Cr(IU) at initial concentrations of Cr(III)T = 1.0,
0.5,0.2 and 0.1 mg/L. Dissolved Cr(ffl) and Cr(VI) concentrations were measured. No Cr(VI)
was detected, indicating that no significant oxidation occurred.  However, the concentration of
dissolved Cr(TH) varied systematically  with increasing DOC and suspended solids.  The results
of a linear partitioning model which considers the species: Cr3"1", Cr=DOC,  and Cr=SS, and
Cr(OH)3(s) is shown in Figure 3-15.

       The linear partitioning between Cr3* and Cr(OH)3(s) is modeling the initial stages of
precipitation of Cr(OH)3(s).  Figure 3-15A (DOC = 9.4 mg/L) and Figure 3-15B (DOC = 24
mg/L) present the data and the model results. The total dissolved chrome is plotted versus the
suspended solids (SS) concentration.  Each line represents a different initial concentration of
Cr(JU.)T, which is plotted, on the left edge of the plot as a star for visual reference. Figure 3-15C
presents the concentrations of species computed from the model for one case: Cr(III)T = 1.0 mg/L
and DOC = 24 mg/L.

       A summary of the chromium cycle in natural waters is presented in Figure 3-16 (Richard
and Bourg 1991). The cycling from Cr(VI) to Cr(JJI) in sediments is illustrated, as is the

                                   Draft for SAB 1-65

-------
       DOC = 9.4mg/l
  i.oo*
DOC = 24 mg/L
                               100*
X
  001
       0  200 400 600 1001000
                               D.10i
                               0.01
Speciafion
0 ZOO 400 600 100 1000

Suipendetf Solldl (fflf/L)
                                                               0  200 400 COO SOO 1000
    Figure 3-15. Model of Cr(III) sorption to suspended solids and DOC. Varying initial
    concentrations of CrCm^ (see the stars plotted on the y-axis). Plots of total and dissolved
    chromium vs total suspended solids. Alternating filled and hatched symbols represent the
    data for each CrOIIV (A) DOC - 9.4 mg/L. (B) DOC = 24 mg/L. C) An example of the
    computed speciation for the DOC = 24 mg/L and CrCDO^ - 1.0 mg/L case.
                                 Draft for SAB 1-66

-------
                         input

                       + FeODor-^
                  /*"   org. matter   ^*
             CKVI)
>
•



>
\
\
weak
adsorption
•I
settling

-
adsorption
or
precipitation
sdiling
+ organic
matter

/
CrOID-org
• t
              diffusion
                  \
  1
I
                                    CrOH)
                      + dissolved
                      organics
    dffiston
CrOro-org
 s>
Figure 3-16.  Chromium cycling in the aquatic environment (Richard and Bburg, 1991).
                           Draft for SAB 1-61

-------
 oxidation of Cr(III) to Cr(VI) by manganese dioxide MnO2. It is this latter reaction that is the
 only source of concern.  Preliminary experiments that we have performed indicate that the
 reaction is extremely slow, and is limited by the solubility of Cr(OH)3(s).  Therefore we do not
 expect that this is a source of concern.
           f
       The formation of Cr(OH)3(s) lowers mat dissolved concentration at SS=0 below Cr(III)T.
 Note that if the normal solubility product formulation were applied,  [Cr3*] = 10'7 M and since
 [Cr=DOC]  = KODQC ICr?][DOC],  the dissolved Cr(III) = [Cr3*] + [Cr=DOC] would also be
 constant.  The fact that dissolved chrome is varying is a consequence of the system not having
 reached equilibrium with respect to Cr(OH)3(s). As shown in FigureS- 15C, most of the dissolved
 chrome is complexed to DOC and, therefore, is presumably, not bioavailable (Campbell  1995).
 The results of this modeling exercise are sorption constants of Cr(HI) to DOC and SS.

 3.3.3  Spiked Sediments: All experimental results summarized                    -

       Berry et al. (1996) summarized the results of all of the above studies using 10-day toxicity
tests with saltwater sediments spiked with cadmium, copper, lead, nickel, silver or zinc and metal
mixtures using amphipods (DiToro etal., 1990; Berry etal., 1996; 1999), polychaetes exposed
to sediments spiked with cadmium, copper, lead, nickel or zinc (Casas and Crecelius, 1994; Pesch
et al., 1995), copepods exposed to sediments spiked with cadmium (Green et al..  1993),  and
freshwater tests using oligochaetes and snails exposed to sediments spiked with cadmium (Carlson
et a., 1991).  These data describe tests with seven freshwater  and saltwater species and sediments
from seven locations, with AVS concentrations ranging from 1.9 to 65.7 pimol/g dry wt and TOC
ranging from 0.15 to 10.6% (Green et al., 1993 measured interstitial cadmium but not AVS).
Similar results from 10-day tests with amphipods hi two marine sediments are summarized in
Berry etal., 1999.

       Overall, the results of these experiments demonstrate that it is not possible to predict the
toxicity of sediments spiked with metals using the total metal concentration on a dry weight basis
(Figure 3-17a). Sediments having s 24%  mortality are considered non-toxic as defined by Berry
et al. (1996), and as indicated by the horizontal line in each panel of Figures 3-17 and 3-18.
Much of this variability is caused by the fact that the relationship between mortality and total
metal concentrations hi tests was sediment specific as it was in the cadmium results shown in
Figure 2-10.  The dry weight metal concentrations required to cause acute mortality in these
                                   Draft for SAB 1-68

-------
        c
        o
100 .


 80


 60


 40
                                       O  O*»dBODn> O'D 09   (BPQQCBJ 00O ODO *™
                                            00     °,0 „
                 0.01        0.1         1         10        , 100       1000
                               Total Metal or SEN! [umol/g)

         o

100

80
60


40


20
ft
1 1 I I Mill 1 1 M Mill
B

o



. o

0 0 ,
_' 	 V"t>0*TJ 	
,,8 ° ofi°° *
,58 - °o
-------
100
-r 80
JT 60
*ca
r
20
Q
0.(
100
1 -
I -
20
f
0
' 100
.* "
1 '6C
i «
' Z(
Figure 3-18. Pera
spiked sediments (
al., 1999; all other
sumoftheconcen
sediment- fit) inte
riTrrmr- i ninni r iiirrrn — rn 	 inin — i r'liini
A •
" o o0cB>o CDOO-
v o oo Tp o
O Spiked Sediment O o 0 o * o
• O Field Sediment O . rf» . O c
- " s S?
0 . o •
0
O o o
O o o /\
g 000 <>
' ./i^s^^f^00,,,,^

1 0.1 1 10 100 1000
Total Metal or SEM (umol/gj
i i i nun -iii inn i i IIMIII i i IIIIIH ~l 1 IDiii
B
o 00*00 OOOD* o &OCKB ooxo.a
O O o ou o o,
P0°00 °0 -
0 °
. „ 0
oT°o
• o !° °0 o o H
0 Oft ° 0
l^^^^^j ° ~

01 0.1 1 10 100 1000
Interstitial Water Toxic Units
i 111!
C -jj^j.
«»o Q 8Oj
A «° ° ° ^ 2
m ? 0 Vn O "
•
4«° *
ff 0 ° 0
" ^ ft^JL °" 536 " "o" •*"-"**" "o "J*
^^ ^ ••
SO -25 0 25 50 75 100
SEM-AVS
sntage mortality of saltwater and freshwater benthic species
open symbols) and sediments from the field (closed symbol
data modified after Hansen et al., 1996a). Mortality is plo
[rations of cadmium, copper, lead, nickel and zinc in jumol
rstitial water toxic units: and (c) SEM/AVS ratio. Soecies
,
!
125
> in 10-day
s) (silver (
ttedasafu
ss metal pe
tested incl
oligochaete Lumbriculus variegatvs, polychaetes (Capitella capitota and Neanthes arenaceodentata), the
barpacticoid {Amphiascus tenuiremis)', amphipods (Ampelisca abditasad Hyalella azteca) and the snail
(Helisoma sp.).  Data below the SEM detection limit are plotted at SEM/AVS = 0.01. Data below the
detection limit of metals in interstitial water are plotted at IWTU - 0.01.
                                 Draft for SAB 1-70

-------
 experiments are very high relative to those often suspected to be of lexicological significance in
field sediments. This has sometimes been interpreted as a limitation of the use of SEM and AVS
to predict metal-induced toxicity.  However, the range in AVS hi these sediments spiked with
metals is similar to sediments commonly occurring in the field. The important point is that even
a  sediment  with only  a moderate  concentration of AVS has  a considerable capacity  for
sequestering metals as a metal sulfide, a form which is not bioavailable (Di Toro et al., 1990).

       In contrast, the combined data from all available freshwater and saltwater spiked-sediment
experiments supports the use of IWTU to predict mortality of benthic species in spiked sediment
toxicity tests (Figure 3-17b).  Mortality in these experiments was sediment independent when
plotted against IWTU.  Sediments with IWTUs of < 0.5 were generally not toxic.  Of the 96
sediments with IWTU  < 0.5, 96.9%  were not toxic, while 76.4%  of the 89 sediments with
IWTU 2 0.5 were toxic (Table 3-2). This close relationship between IWTU and sediment toxicity
in sediments spiked with metals was also observed hi studies with field sediments contaminated
with metals (See Section 3.2.3 below), sediments spiked with nonionic organic chemicals (Adams
et al., 1985; Di Toro et al., 1991;  1985; Swartz et al., 1990) and  field sediments contaminated
with nonionic organic chemicals (Hoke et al., 1994; Swartz et al., 1994).
                                  Draft for SAB 1-71

-------
Table 3-2.  Toxicity of sediments from freshwater (FW) and  saltwater (SW) field locations,
spiked-sediment tests and combined field and spiked-sediment tests as a function of the difference
between the molar concentrations of SEM and AVS (SEM-AVS), interstitial water toxic units
(IWTUs) and both SEM-AVS and IWTUs (modified from Hansen et al., 1996a).
Study
Type Parameter .
Lab-Spike SEM-AVS
(FW & SW)
IWTU

SEM-AVS, IWTU

Field SEM-AVS
(FW&SW) ' . •
IWTU

SEM-AVS, IWTU

All SEM-AVS
^
IWTU
%
SEM-AVS, IWTU

Percent of Sediments
Value
sO2
>(?
<0.5
*0.5
* O2, < 0.5
> O3, * 0.5
sO2
X)3
<0.5
* 0.5
s O2, < 0.5
X)3, 2 0.5
sO2
> O3, a 0.5
n
101
95
96
89
83
78
57
79
79
53
49
45
158
174
175
142
132
123
Non-toxic1
98.0
26.3
96.9
23.6
97.6
14.1
98.2
59.5
98.7
45.3
100.0
33.3
98.1
42.0
97.7
31.7
98.5
21.1
Toxic1
2.0
73,7
3.1
76.4
2.4
85.9
1.8
40.5
1.3
54.7
0.0
66.7
1,9
58.0
2.3
68.3
1.5
78.9
1 Non-toxic sediments < 24 percent mortality. Toxic sediments > 24 percent mortality.
2 SEM-AVS sO is the same as an SEM/AVS ratio of * 1.0.
3 SEM-AVS > 0 is the same as an SEM/AVS ratio of > 1.0.
                                Draft for SAB 1-72

-------
          The interstitial water metal concentrations in all spiked-sediment studies were usually
below the limit of analytical detection in sediments with SEM/AVS ratios below 1.0 (Berry et al.,
1996). Above an SEM/AVS ratio of 1.0, the interstitial metals concentrations increased up to five
orders of magnitude with increasing SEM/AVS ratio/This several order of magnitude increase
in interstitial water metals concentration with only a factor of two or three increase in sediment
concentration is why mortality is most often complete hi these sediments, and why the chemistry
of anaerobic sediments controls the toxicity of metals to organisms living in aerobic micro-
habitats.  It also explains why the toxicity of different metals in sediments to different species is
so similar.  Interstitial water metals were often below or near detection limits when SEM/AVS
ratios were only  slightly above 1.0 indicating the presence of other metals binding phases in
sediments.
                                 ^
          The combined data  from all available saltwater and freshwater spiked sediment
experiments also supports the use of SEM/AVS ratios to predict sediment toxicity to benthic
species in spiked-sediment toxicity tests. All tests yield similar results when mortality is plotted
against SEM/AVS ratio (Figure 3-17c). Mortality in these experiments was sediment Independent
when plotted on an SEM/AVS basis. With the combined data, 98.0% of the 101 metals-spiked
sediments with SEM/AVS ratios s 1.0 were not toxic, while 73.7% of the 95 sediments with
SEM/AVS ratios >  1.0 were toxic (Table 3-2).

          These overall data show that when both SEM/AVS ratio and IWTU are used,
predictions of which sediments would be toxic were unproved.  Of the 83 sediments with
SEM/AVS  ratios <,  1.6  and IWTU < 0.5, 97.6%  were not toxic,  while 85.9% of the 78
sediments with SEM/AVS ratios > 1.0 and  IWTU *  0.5 were toxic (Table 3-2). ( Note: Table
3-2 uses  SEM-AVS instead of SEM/AVS  ratios.  An SEM-AVS of sO is the same as an
SEM/AVS ratio of * 1.0. An SEM-AVS of > 0 is the' same as an SEM/AVS ratio of > 1.0.)  '

          These results show that SEM/AVS and IWTU are accurate predictors of the absence
of mortality in sediment toxicity tests, however, predictions of which sediments might be toxic
are less accurate.  The fact that a significant number of sediments  (26.3%) tested had SEM/AVS
ratios of  > 1.0,  but were not toxic indicates that other binding phases/ such as organic carbon
(Mahony et al.,  1996),  may also  control bioavailability  hi anaerobic sediments.  While the
SEM/AVS model of bioavailability accurately predicts which sediments will not be toxic, a model
which utilizes SEM/AVS ratios or (SEM-AVS) (Hansen et al.,  1996aJ and incorporates other

             .                     Draft for SAB 1-73        ,   '    .

-------
binding phases might more accurately predict which sediments will be toxic (Di Tord et al., 1987;
Mahonyetal., 1996).
     -•                                                  '
          Organism behavior may also explain why sediments with SEM/AVS ratios of >  1.0
were not toxic. Many of the sediments which had the highest 'SEM/AVS ratios in excess of 1.0
that produced little or no mortality were from experiments using the polychaete, Neanthes
arenaceodentata (see Pesch et al., 1995, Figure 8). This appeared to be related, in part, to the
ability of this polychaete to avoid burrowing into the test sediments, thereby limiting its exposure
to the elevated concentrations of metals in the interstitial water and sediments.  This  same
phenomenon may also explain the low mortality of snails, Heliosoma sp., in freshwater sediments
with high SEM/AVS ratios.  These snails are epibenthic and also have the ability to avoid
contaminated sediments (G. Phipps, personal comm.).  Increased mortality was always observed
in sediments with SEM/AVS ratios > 5.9 hi tests with the other five species.

          Similarly, a significant number of sediments with i 0.5 IWTUs were not toxic. This
is likely the result of IW ligands which reduce the bioavailability and toxicity of dissolved metals,
sediment avoidance by polychaetes, or snails, or methodological problems in contamination-free
sampling of IW. Ankley et al. (1991) suggested-that a toxicity correction for the hardness of the
IW is needed to compare toxicity in IW to that in water-only tests. Absence of a correction for
hardness might affect the accuracy of predicting metal-induced sediment toxicity using IWTUs hi
freshwater.   Further, a significant improvement in  the accuracy  of  metal-induced toxicity
predictions using IWTUs might be achieved if DOC binding hi the IW is taken into account.
Green et al. (1993) and Ankley et al. (1991) hypothesized that increased DOC hi the IW reduced
the bioavailability of cadmium in sediment exposures, relative to the water-only exposures. Green
et al. (1993) found that the LC50 value for cadmium in  an  IW exposure without sediment was
more than twice that hi a water-only exposure, and that the LC50 value for cadmium hi  IW
associated with sediments was more than three times that hi a water-only exposure.
3.3.4
Field sediments
          In addition to short-term laboratory experiments with spiked sediments, there have been
several published studies of laboratory toxicity tests with metal-contaminated sediments from the
field.  Ankley et al. (1991) exposed L. variegatus and the amphipod Hyalella azteca to 17
sediment  samples  along  a  gradient  of  cadmium  and  nickel  contamination  from  a
                                   Draft for SAB U74

-------
 freshwater/estuarine site in Foundry Cove, NY.  In 10-d toxicity.tests, H. azteca mortality was
 absent in all sediments where SEM (cadmium plus nickel) was less than AVS.  Mortality was
 greater than controls only in sediments with more SEM than AVS.  Lumbriculus variegatus was
 far less sensitive to the sediments than H. azteca, which correlates with the differential sensitivity
 of the two species in water-only tests with cadmium and nickel.

          In 10-day toxicity tests with the saltwater amphipod A. abditam these same sediments
 from Foundry Cove, Di Toro et al. (1992) observed that metals concentrations ranging from 0.1
 to 28 /umoles SEM/g sediment were not toxic in some sediments, whereas metals concentrations
 ranging from 0.2 to 1000 pinoles SEM/g were lethal hi other sediments.  These results indicate
 that the bioavailable fraction of metals hi sediments varies from sediment to sediment. In contrast,
   i     •                            •                       -
 the authors also observed a clearly discernable mortality-concentration relationship when mortality
 was related to the SEM/AVS molar ratio (i.e., there was no significant mortality where SEM/AVS
 ratios were < 1.0, mortality increased in sediments having SEM/AVS ratios 1.0-3.0, and there
 was 100 % mortality in sediments with ratios > 10).  The sum of the interstitial water toxic units
 (IWTU) for cadmium and nickel ranged from 0.08 to 43.5.  Sediments with <. 0.5 IWTUs were
 always non-toxic, those with >2.2 IWTUs were always toxic, and two of seven sediments with
 intermediate IWTUs (0.5 to 2.2) were toxic. Molar concentrations  of cadmium and nickel in the
 interstitial water were similar. However, cadmium contributed over 95 percent to the sum of the
 toxic units because cadmium is 67 tunes more toxic to A. abdita than nickel. The latter illustrates
the utility of interstitial water concentrations of individual metals in assigning the probable cause
 of mortality hi benthic species (Hansen et al., 1996a).

          In tests with the same sediments also from Foundry Cove, Pesch et al. (1995) observed
that six of the 17 sediments tested had SEM/AVS ratios  < 1.0, interstitial water toxic units < 0.5,
and none of the six were toxic to the polychaete  Neanthes arenaceodentata.  Nor were 11
sediments that contained SEM/AVS ratios > 1.0 toxic. These results were not surprising because
only one sediment had > 0.5 IWTUs, and because N. arenaceodentata is not sensitive to cadmium
and nickel and can avoid sediments containing toxic concentrations of these metals.
          ,                                                     <            •
          Ankley et al. (1993) examined the significance of AVS as a binding phase for copper
 in freshwater sediments from two copper-impacted sites.  Based upon interstitial water copper
concentrations in the test sediments, the 10-d LC50 for H. azteca  was 31 Aig/L; this compared
favorably with  a measured LC50 of 28 jUg/L hi a 10-d water-only test.  Sediments having

                                  Draft for SAB 1-75

-------
SEM/AVS ratios < 1.0 were not toxic. They also observed no toxicity in several sediments with
markedly more SEM than AVS suggesting that copper was not biologically available in these ,
sediments. Absence of copper in interstitial water from these sediments corroborated this lack of
bioavailability.  This observation suggested the presence of binding phases hi addition to AYS for
copper in the test sediments. Recent studies suggest that an important source of the extra binding
capacity hi these sediments was organic carbon (Mahony et al., 1996; U.S. EPA, 1994a).

          Hansen et al. (1996a) investigated the biological availability of sediment-associated
divalent metals  to A.  abdita and H. azteca in sediments from five saltwater locations and one
freshwater location hi the United States, Canada and China using 10-day lethality tests. Sediment
toxicity was not related to dry weight metals concentrations. In the 49 sediments evaluated where
metals were the likely cause of toxicity (i.e., those with less SEM than AVS and those with less
than 0.5 IWTU), no toxicity was observed. One third of the 45 sediment samples with more SEM
than AVS and more than 0.5 IWTU were toxic.

          Hansen et al. (1996a) made an observation that is important to the interpretation of the
toxicity of sediments from field locations, particularly those  from industrial harbors.   They
observed that if these sediments are toxic and SEM/AVS ratios are < 1.0, non-metals associated
toxicity should always be suspected even if metals concentrations are very high on a dry weight
basis. Further, they stated that the use of such data to reach the conclusion that this EqP approach
is not valid is incorrect.  This is because when SEM/AVS ratios were less than 1.0, there was an
almost complete absence of toxicity in spiked sediments, and Meld sediments where metals were
the only known  source of contamination and  IWTUs for metals  were  <0.5.   Metals
concentrations, when expressed on a sum of the IWTU basis, can therefore provide insight that,
hi part,  may explain apparent anomalies between SEM/AVS ratios and the  observed toxicity of
these sediments and; sediments from other field sites. The joint use of both SEM/AVS ratios and
interstitial water concentrations are powerful tools for explaining the presence of toxicity when
SEM/AVS ratios are < 1.0, and the absence of toxicity when SEM/AVS ratios are > 1.0.  Over
all saltwater and freshwater field sediments tested in the laboratory, 100% were not toxic when
SEM-AVS was  sO.O and IWTUs were  <0.5 and 66.7% were toxic when SEM-AVS was  >0.0
and IWTUs were *0.5 (Table 3-2).
                                   Draft for SAB 1-76

-------
3.3.5     Field Sites and Spiked Sediments Combined

          Figures 3-18, a,b, and c and Table 3-2 summarize available data from saltwater and
freshwater sediments spiked with individual metals or metal mixtures, saltwater field sites and
freshwater field sites on the utility of metals concentrations in sediments normalized by dry
weight, interstitial water toxic units (IWTUs) or SEM/AVS ratios to explain the bioavailability
and acute toxicity of metals in sediments. Data are from Hansen et al.(1996a) and Berry et al.,
1999.  This analysis  contains all available data from 10-day lethality tests where mortality,
IWTUs, and SEM/AVS ratios are known from experiments with sediments most certainly toxic
only because of metals. The relationship between benthic organism mortality and total dry weight
metals concentrations in  spiked and field sediments is  not useful to causally relate metal
concentrations  to organism response (Figure  3-18a).   The  overlap  among bulk metals
concentrations which  cause no toxicity and those which are  100 percent lethal is almost four
orders of magnitude.

          Data  in Figure 3-18b show  that over all tests, the toxiciry of sediments whose
concentrations are normalized on an IWTU basis are typically consistent with the interstitial water
toxic unit concept; that is if IWTUs are s 1.0, then sediments should be lethal to <; 50 percent of
the organisms exposed, and significant mortality probably should be absent at <  0.5 IWTU
(Figure 3-18c).  Of the spiked and field sediments evaluated which had IWTUs < 0.5, 97.7
percent of 175 sediments were non-toxic (Table 3-2).  For the 142 sediments having IWTUs ;>
0.5,  68.3 percent were toxic (Table 3-2).  However, and as  stated above, given the effect on
toxiciry or bioavailability of the presence of other binding phases (e.g., DOC) in interstitial water,
water quality (hardness or  salinity)  and organism behavior, it is not surprising that many
sediments having IWTUs * 0.5 are not toxic.

          Data in Figure 3-18c show that over all tests, organism response in sediments whose
concentrations are normalized on an SEM/AVS basis is consistent with metal-sulfide binding on
a mole to mole basis  as first described by Di Toro et al. (1990), and later recommended for
assessing the bioavailability of metals in sediments by Ankley et al. (1994).  Saltwater and
freshwater sediments spiked with metals and from field locations with SEM/AVS ratios  £ 1.0
were uniformly (98.1 percent of 158 sediments) non-toxic (Figure 3-18b; Table 3-2). The majority
(58.0 percent) of 174 sediments having SEM/AVS ratios > 1.0 were toxic. On the  other hand,
given the effect on toxicity or bioavailability of the presence of other sediment phases that also

                                   Draft for SAB 1-77

-------
 affect bioavailability (Di Toro et al., 1987; Mahony et al., 1996), it is not surprising that many
 sediments having SEM/AVS ratios > 1.0 are not toxic.

          Over all tests, the data in Figure 3-18a, b, and c indicate that the use of both IWTUs
.and SEM/AVS ratio's together did not improve the accuracy of predictions of sediments that were
non-toxic (98.5 percent of 132 sediments; Table 3-2).  However, it is  noteworthy that 78.9
percent of the 123 sediments with both SEM/AVS  > 1.0 and IWTUs ;> 0.5 were toxic (Table 3-
2). Therefore, the approach of using SEM/AVS ratios, IWTUs, and especially both indicators
to identify sediments of concern is very useful.

          The results of all available data demonstrate that using SEM, A VS and interstitial water
metals concentrations to predict the toxicity of cadmium, copper, lead, nickel, silver and zinc in
sediments is quite certain.  This is very useful, because the vast majority of sediments found in
the environment hi the U.S. have SEM/AVS ratios <. 1.0. This suggests that there should be little
concern about metals in sediments (Wolfe et al.,  1994; Hansen et al., 1996a; Leonard et al.,
1996a; Section 4 of this document) on a national basis, even though localized areas of biologically
significant metal contamination do exist. However, a very important consideration is that most
of these data are from field sites where sediment samples were collected in the summer. This is
the tune  of the year when the seasonal cycles of AVS produce the maximum metal-binding
potentials (Boothman and Helmstetter, 1992; Leonard et al., 1993). Hence, sampling at seasons
and conditions when AVS is at minimal values is a must in establishing the true level of overall
concern about metals hi sediments and in evaluations of specific sediments. Predicting which of
the sediments with SEM/AVS > 1.0 will be toxic is presently less certain.   Importantly, the
correct classification rate seen hi these experiments (accuracy of predicting which sediments were
toxic was 58.0% using the SEM/AVS ratio alone, 68.3% using IWTUs and 78.9% using both
indicators) is high.  An SEM/AVS ratio > 1.0, particularly at multiple adjacent sites, should
trigger additional tiered assessments which might include characterization of the spatial (both
vertical and horizontal) and temporal distribution of chemical concentration (AVS and SEM) and
toxicity, measurements of interstitial water metal and toxicity identification evaluations (TIE's).
In this context, the SEM,  AVS, IWTU approach should be viewed as only one of the  many
sediment evaluation methodologies.                         ~

          Because AVS can bind divalent metals hi proportion to then- molar concentrations,
Hansen et al. (1996a) proposed the use of the difference between the molar concentrations of SEM
and AVS (SEM-AVS)  rather than SEM/AVS ratios used previously.  The molar difference
provides  important insight into the extent of additional available binding  capacity and  the
                                   Draft for SAB 1-78

-------
magnitude by which AVS binding has been exceeded (Figure 3-19). Further, absence of organism
response when AVS binding is exceeded can indicate the potential magnitude of importance of
other binding phases in controlling bioavailability.  Figure 3-19 shows that for most non-toxic
saltwater and freshwater field sediments, one to 100 //moles of additional metal would be required
to exceed the sulfide binding capacity (i.e., SEM-AVS = -1 to -100 ^moles/g). In contrast, most
toxic field sediments contained 1.0 to 1000 jumoles of metal beyond the binding capacity of sulfide
alone. Data on non-toxic field sediments whose sulfide binding capacity is exceeded (SEM-AVS
is > 0.0 ^moles/g) indicates that other sediment phases, in addition to AVS, have great
significance in controlling metal bioavailability., In comparison to SEM/AVS ratios, the use of
SEM-AVS differences is particularly informative where AVS concentrations are low, such as
those from Steilacoom Lake and the Keweenaw Watershed, where the SEM-AVS difference is
numerically low and SEM/AVS ratios are high (Ankley et  al., 1993).

          EPA believes that results from tests using sediments spiked with metals and sediments
from the field in locations where toxicity is metals-associated demonstrate the value in explaining
the biological availability of metals  concentrations normalized by SEM/AVS ratio and IWTUs
instead of dry weight metals  concentrations. Importantly, data from spiked  sediment tests
strongly indicate that metals are not the cause of most of the toxicity observed in field sediments
when both SEM/AVS ratios are.* 1.0 and IWTU are < 0.5 (Table 3-2).  Expressing
                                  Draft for SAB 1-79

-------
CO
uu

•

80-
.
60-
40-

20-


n
w
-11
	 , 	 . 	 . o m-p 	 Q.
i oo
! . . o ••
. - o *
D D • "
• • ' ' 'o . .
'
D
' ° •
• • • • •
A
in
"•••-'•
A .......«.„„. ,„,„ „.......,., 	 	 	 .„ 	 	 	 L..............L........... 	 ,..............^...^A,.,^....... ,»^.«...
A Of • • - - A* •
A fl^H- . ' ^^ 1 ^H 4^^^i ^v
^^ 7L A OA A^ •
^T1«A^ Jf- A Vbo 4A *
4^ AA * 0 i AO^A A1A ++ *
-i_A-AT* fji_ ^A .jurQt/t- ? _*. _A_ /VWL JL J_-A__-». "
00 -10 '-1 <0 1 10 100 100








•




0
                                SEM-AVS 0/mol/g dry wt)
         Figure 3-19. Percentage mortality of amphipods, oligochaetes and potychaetes exposed to
         sediments from three saltwater and four freshwater field locations as a function of the sum
         of the molar concentrations of SEM minus the molar concentration of AVS (SEM-AVS)
         (from Hansen et al., 1996a): The vertical dashed tine at SEM-AVS = 0.0 indicates the
         boundary between sulfide-bound unavailable metal and potentially available metal.
                                     Draft for SAB 1-80

-------
 concentrations of metals in sediments on an SEM-AVS basis provides important insight into
available additional binding capacity of sediments and the extent to which sulfide binding has been
exceeded.  It,  along with measurement  of interstitial water  concentrations of metals, can
potentially identify the specific metal causing toxicity. This can theoretically be accomplished by
subtracting the metals-specific molar concentrations in order of their sulfide solubility product
constants (K^). Predictions of sediments not likely to be toxic, based on use of SEM-AVS and
rWTUs for all data from freshwater or saltwater field sediment and spiked sediment tests are
extremely accurate (98.5 percent) using both parameters (Table 3-2).  While the predictions of
sediments likely to be toxic are less accurate, the use of SEM-AVS  is extremely useful  in
identifying sediments of potential concern (Table 3-2).  Hansen (1995) summarized data from
amphipod tests using freshwater and saltwater laboratory metals-spiked sediments and field
sediments where metals were a known problem by comparing the percentage of sediments that
were toxic to the SEM-AVS concentration (tests with polychaetes and gastropods were excluded
because these organisms avoid exposure). Seventy percent of the sediments in these amphipod
studies with an SEM-AVS concentration of aO.76 //moles of excess SEM/g were toxic. The
corresponding values for 80,  90 and 100% of the sediments being toxic were 2.7, 16 and 115
jumoles of excess SEM/g, respectively.

           Of course, SEM, AVS and IWTUs can only predict toxiciry or the lack of toxicity due
 to metals in sediments.  They cannot be used  alone to predict the toxicity  of sediments
 contaminated with toxic concentrations of other contaminants. However, SEM/AVS ratios have
 been used in sediment assessments to rule out metals as probable causative agents of toxicity
 (Wolfe et al., 1994). Also, the use of SEM and AVS to predict  the biological availability and
 toxicity of cadmium, copper, lead, nickel, silver  and zinc  is applicable only to anaerobic
 sediments that contain AVS.  In  aerobic sediments binding  factors other then AVS control
 bioavailability (Di Toro et al., 1987; Tessier et al., 1993). Measurement of interstitial water
 metal may be useful for evaluations of these and other metals in aerobic and anaerobic sediments
 (Ankley et al., 1994).   Even with these caveats, EPA believes that the use of SEM, AVS and
 interstitial measurements in combination are superior to all other currently available sediment
 evaluation procedures to causally assess the implications of  these six metals associated with
 .sediments (see discussion in  Section 5 "Implementation" for further guidance).

 3.4      PREDICTING METAL TOXICITY: LONG-TERM  STUDIES

           Taken as a whole, the short-term laboratory experiments with metal-spiked and field-
 collected sediments present a strong argument for the ability to predict an absence of metal

                                    Draft for SAB 1-81

-------
toxicity  based upon  sediment  SEM:AVS  relationships  and/or  interstitial  water  metal
concentrations.  However, for this approach to serve as a valid basis for ESG derivation,
comparable predictive success must be demonstrated in long-term laboratory and field experiments
where chronic effects could be manifested (Luoma and Carter, 1993; Meyer et al.t 1994). This
demonstration was the goal of experiments described by Hare et al. (1994), DeWitt et al. (1996),
Hansen et al. (1996b), Liber et al. (1996) and Sibley et al. (1996). An important experimental
modification to these long-term studies, as opposed to the short-term tests described in Section
3.3, was the collection of horizon-specific chemistry data.  This is required because AVS
concentrations often increase, and SEM/AVS ratios decrease, with an increase in sediment depth
(Howard and Evans, 1993; Leonard et al., 1996a); hence, chemistry performed on homogenized
samples might not reflect the true exposure of benthic organisms dwelling hi surficial sediments
(Luoma and Carter, 1993; Hare et al., 1994; Peterson et al., 1996).
                                                       (
3.4.1     Life-cycle toxicity tests

          DeWitt et al. (1996) conducted an  entire life-cycle toxicity test with the marine
amphipod Leptocheirus plumulosus exposed for 28 d to cadmium-spiked estuarine sediments
(Table 3-3).  The test began with newborn amphipods and measured effects on survival, growth
and reproduction relative to interstitial water and SEM/AVS normalization. Seven treatments of
Cd were tested: 0 (control), 0.34, 0.74, 1.31,  1.55, 2.23 and 4.82 molar SEM^/AYS ratios
(measured concentrations). Gradients hi AVS concentration as a function of sediment depth were
greatest in the control treatment, decreased as the SEM^/AYS ratio increased, and became more
pronounced over tune.  Depth gradients in SEM^/AYS were primarily due to the spatial and
temporal changes in AYS concentration, because SEM^ concentrations changed very little with
time or depth. Thus, in most treatments SEM^/AYS ratios were higher at the top of sediment
cores than at the bottom. This is expected because the oxidation rate of iron sulfide in laboratory
experiments is very rapid (100% in 60 to 90 minutes) while for cadmium sulfide it is quite slow
(10% hi 300 hours) (Mahony et al.,  1993). Interstitial cadmium concentrations increased hi a
dramatic step-function fashion in treatments having SEM/AYS ratios ;>2.23; and were below the
96-h LC50 for this amphipod in lesser treatments. There were no significant effects on survival,
 growth or reproduction in sediments                                       •          .
                                    Draft for SAB 1-82

-------
   -M
   1
I
  H

  •t

  "I
          s?
I


I.

\o
•rj

of
oo
               ON <1
               _.' (S
               cs «n
               \O~
          3
               N
                                    *  -a"
                                    ON  «
                                    r.  «
                                  «*  S2  3.2
                        00
                            cs
                            9
                                      9 2  
-------
 containing more AVS than cadmium (SEM/AVS ratios 0.34,0.74,1.31 and 1.55), in spite of the
 fact that these samples contained from 183 to 1370 jug cadmium/g sediment. All amphipods died
 hi sediments having SEM/AVS ratios 2.23. These results are consistent with predictions of metal
 bioavailability from acute tests with metal-spiked sediments (i.e., that sediments with SEM^/AVS
 ratios si are not toxic, interstitial water metal concentrations are related to organism response,
 and sediments with SEMc/AVS ratios > 1 may be toxic).
    •                                      '                     '
          Sibley et al. (1996) reported similar results from a 56-d life-cycle test conducted with
 the freshwater midge Chironomus tentans exposed to zinc-spiked sediments (Table 3-3). The test
 was initiated with newly hatched larvae and lasted through one complete generation during which
 survival, growth, emergence and reproduction were monitored. In sediments where the molar
 difference between SEM and AVS was <0, at dry wt zinc concentrations as high as 270 mg/kg,
 concentrations of zinc in the sediment interstitial water were low and no adverse effects were
 observed for any of the biological endpoints measured.  Conversely, when SEM-AVS exceeded
 0, AVS and interstitial water concentrations of zinc increased with increasing treatments (being
 highest in surficial sediments), and reductions in survival, growth,  emergence and reproduction
 were observed.  Over the course of the study, the absolute concentration of zinc in the interstitial
 water in these treatments decreased due to the increase in sediment AVS and loss of zinc from
 twice daily renewals of the overlying water.
3.4.2
Colonization tests
          Hansen et al. (1996b) conducted a 118-d benthic colonization experiment in which
sediments were spiked to achieve nominal cadmium/AVS molar ratios of 0.0 (control), 0.1, 0.8
and 3.0 then held in the  laboratory in a  constant flow of unfiltered seawater (Table 3-3).
Oxidation .of AVS hi the surficial 2.4 cm  of the control treatment within two to four weeks
resulted in sulfide profiles similar to those occurring in sediments in nearby Narragansett Bay, RI
(Boothman and Helmstetter, 1992).  In the nominal 0.1 cadmium/A VS treatment, measured
SEMo, was always .less than AVS, interstitial cadmium concentrations (<3-10 jig/L) were less
than those likely to cause biological effects, and no significant biological effects were detected.
In the nominal 0.8 cadmium/AVS treatment, measured SEM^ commonly exceeded AVS in the
surficial 2.4 cm of sediment, and  interstitial cadmium concentrations (24-157 Mg/L)  were
sufficient to be of toxicological significance to highly sensitive species.  In this treatment, shifts
in the presence or absence over all taxa, and fewer macrobenthic polychaetes (Mediomastus
ambiseta,  Streblospio benedicti and Podarke obscura) and unidentified meiofaunal nematodes,
were observed.  In the nominal  3.0 cadmium/AVS treatment, concentrations of SEMoi  were
                                   Draft for SAB 1-84

-------
 always greater than AVS throughout the sediment column. Interstitial cadmium ranged from
 28,000 to 174,000 pg/L. In addition to the effects observed in the nominal 0 j cadmium/AVS
 treatment, sediments in the 3.0 cadmium/AVS treatment were colonized by fewer macrobenlhic
 species, polychaete species and harpacticoids, had lower densities of diatoms, lacked bivalve
 molluscs, and'exhibited other impacts.  Over all treatments, the observed biological responses
 were consistent with predicted possible adverse effects resulting from elevated SEM^/A VS ratios
 in surficial sediments and interstitial water cadmium concentrations.

          Boothman et al. (1996) conducted a field colonization experiment in which sediments
 from Narragansett Bay, RI were spiked with an equi-molar mixture of cadmium, copper, nickel
 and zinc at nominal SEM:AVS ratios of 0.1, 0.8 and 3.0, placed in boxes, and replaced in
 Narragansett Bay (Table 3-3):  AVS concentrations decreased with time in surface (0-3 cm)
 sediments in all treatments where  SEM < AVS, but did  not change in subsurface (6-10 cm)
 sediments or in the entire sediment column in the SEM > AVS treatment.  SEM decreased with
 time only where SEM exceeded AVS. The concentration of metals hi interstitial water was below
 detection limits when  SEM was  less than AVS.  When SEM exceeded AVS, significant
 concentrations of metals were present hi interstitial water hi order of then- sulfide solubility
 product constants.  Interstitial water concentrations in these sediments decreased  with tune
 exceeding the WQC hi interstitial water for 60 days for all metals, 85 days for cadmium and zinc,
 and for the entire experiment (120 days) for zinc.  Benthic fauna! assemblages in the spiked
 sediment treatments were not different from the control treatment. Lack of biological response
 was consistent with the vertical profiles of SEM and AVS. AVS was greater than SEM in all
 surface sediments, including the top 2 cm of the 3.0 SEM: AVS treatment, due to the oxidation
of AVS and loss of SEM. The authors speculated that  interstitial metal was  likely absent in the
surficial sediments  in  spite  of data  demonstrating  the presence  of significant  measured
concentrations of interstitial metal.  This is because the  interstitial water in the nominal 3.0
SEM/AVS treatment was sampled  from sediment depths  where SEM was  in excess.  It is  in
surficial sediments where settlement by saltwater benthic organisms first occurs.  Also, there was
a storm event which allowed a thin layer of clean sediment to be deposited on top of the spiked
sediment (Boothman, USEPA, personal communication). These data demonstrate the importance
of sampling of sediments and interstitial water in sediment horizons where benthic organisms are
active.

          Hare et al. (1994) conducted an approximately  1-yr field colonization experiment in
which  uncontaminated freshwater  sediments were spiked with cadmium and  replaced in the
oligotrophic lake from  which they originally  had been collected  (Table 3-3).   Cadmium

                                  Draft for SAB 1-85

-------
concentrations in interstitial waters were very low at cadmium:AVS molar ratios <1.0, but
increased markedly at ratios. > 1.0:  They reported reductions in the abundance of only the
chironomid Chironomus satinarius in the nominal 10.0 SEM/AVS treatment.  Cadmium was
accumulated by organisms from sediments with surficial SEM concentrations that exceeded those
of AVS.  These sediments also contained elevated concentrations of cadmium hi interstitial water.

           Liber et al. (1996) performed a field colonization experiment using sediments having
4.46 ftmole sulfide from a freshwater mesotrophic pond (Table 3-3). Sediments were spiked with
0.8, 1.5, 3.0, 6.0 and  12.0 ^mole zinc, replaced in the field, and chemically and biologically
sampled over 16 mo. There was a pronounced increase in AVS concentrations with increasing
zinc concentration;  AVS was lowest in the surficial  0-2 cm of sediment with minor seasonal
variations.  With the exception of the highest spiking concentration (ca., 700 nig/kg, dry wt),
AVS concentrations remained larger than those of SEM.  Interstitial water zinc concentrations
were rarely detected in any treatment, and Were never at concentrations that might pose a hazard
to benthic  macroinvertebrates.  The only observed difference in benthic community structure
across the  treatments was a slight decrease in the abundance of Naididae oligochaetes at the
highest spiking concentration. This absence of any noteworthy biological response was consistent
with the absence of interstitial water concentrations of biological concern.  This was attributed to
the  increase in concentrations  of iron and manganese sulfides, produced  during periods of
diagenesis, which were replaced by the more stable zinc sulfide which is less readily oxidized
during whiter months. In this experiment, 'and theoretically hi nature, excesses of sediment metal
might be overcome over tune due to the diagenesis of organic material.  In periods of minimal
diagenesis, the oxidation rates of metal sulfides, if sufficiently great, could release biologically
significant  concentrations of the metal into interstitial waters.  This phenomenon should occur
metal-by-metal in order of then- sulfide solubility product constants.
                                    Draft for SAB 1-86

-------
                                     SECTION 4

                         DERIVATION OF ESG FOR METALS
                                              •*i»
4.1       GENERAL INFORMATION

          Section 4 of this document presents the technical basis for establishing ESG for
copper, cadmium, nickel, lead, silver and zinc. The basis of the overall approach is the use of
EqP theory linked to the concept of maintaining metal activity for the sediment interstitial
water system below effects levels. Extensive toxicological data from short-term and long-term
laboratory and field experiments, with both marine and freshwater sediments and a variety of
                              t
species, indicates that  it is possible to reliably predict an absence of metal toxicity based upon
EqP theory. ESG for  all six metals collectively can be derived using two procedures: (a) by
comparing the sum of their molar concentrations, measured as SEM, to the molar
concentration of AVS  in sediments (AYS Guideline); or (b) by comparing the measured
interstitial water concentrations of the metals to WQC final chronic values (FCVs) (Interstitial'
Water Guidelines). These approaches are described in more detail below.  A lack of
exceedence of ESG based upon any one of the two procedures indicates that metal toxicity
should not occur. Exceedence of either the AVS or Interstitial Water Guidelines is indicative
of a potential problem that would entail further evaluation.
                                                     i,
          At present, EPA believes that the technical basis for implementing these two
approaches is supportable. The Organic Carbon and Minimum Partitioning Approaches as
proposed to the SAB and in Ankley et al. (1996) require additional research prior to their
implementation. Research issues for these latter two approaches include the development of
robust partitioning datasets for the six metals, as well as investigation of factors such as the
importance of other binding phases.  The four approaches have been presented to and
reviewed by the Science Advisory Board of EPA (U.S EPA, 1994a; 1995a).

          Additional  research required to fully implement other approaches for deriving ESG
for these metals and to derive ESG for other metals includes the development of uncertainty
estimates associated with any approach; part of this would include their application to a variety
of field settings and sediment types.  Research also is needed to establish the technical basis
for ESG for metals other than the six described herein, such as mercury, arsenic and
chromium. Finally, the ESG approaches are intended to protect benthic organisms from, direct
toxicity associated with exposure to metal-contaminated sediments.  They are not designed to

             ,                      Draft for SAB 1-87

-------
 protect aquatic systems from metal release associated, for example, with sediment suspension,
 or the transport of metals into the food web either from sediment ingestion or the ingestiori of
 contaminated benthos. This latter issue, in particular, should be the focus of future research
 given existing uncertainty in the prediction of bioaccumulation of metals by benthos (Ankley,
 1996).

        .  The following nomenclature is used in subsequent discussion of ESG derivation for
 metals. The ESG for the metals are expressed in molar units because of the molar
 stoichiometry of metal binding to  AVS.  Thus,  solid phase constituents (AVS, SEM) are in
 moles/g dry wt. The interstitial water metal concentrations are expressed in ^moles/L, either
 as dissolved concentrations [MJ or activities {M2*} (Stumm and Morgan, 1981). The
 subscripted notation, Mj, is used to distinguish dissolved aqueous phase molar concentrations
 from solid phase molar concentrations with no subscript.  For the combined concentration,
 [SEMT], the units are moles of metal per volume of solid plus liquid phase (i.e., bulk).  Note
 also that when [SEMAg] is summed and/or compared to AVS, V*  the molar Ag concentration is
 applied.

          One final point should  be made with respect to nomenclature. Use of the terms
 non-toxic  and having no effect mean, only with respect to the six metals considered in this
 paper.  The toxicity of field collected sediments can be caused by other chemicals. Therefore,
 avoiding exceedences of ESG for metals does not mean that the sediments are non-toxic. It
 only ensures that the six metals  being considered should not have an undesirable biological
 effect.  Moreover, as discussed  in detail below, exceedence of the guidelines for the  six metals
does not necessarily indicate that metals will cause toxicity.  For these reasons, we strongly
recommend the use of toxicity tests, TIEs, chemical monitoring hi vertical, horizontal and
temporal scales, and other assessment methodologies as integral parts of any assessment
concerned with the effects of sediment-associated contaminants (Ankley et'al., 1994).
4.2
SINGLE METAL SEDIMENT GUIDELINES
          Except hi rare instances, single metal guidelines are not usually applicable to field
situations since there is almost always more than one metal to be considered. As will become
subsequently clear, it would be technically indefensible to derive guidelines for one metal at a
time because of the competitive nature of AVS binding. Nevertheless,  it is illustrative to
present the logic for single metals as a prelude to the derivation of the multiple metal
                                   Draft for SAB 1-88

-------
 guidelines.

 4.2.1  AVS Guidelines

        It has been demonstrated that if the SEM of a sediment is less than or equal to the
 AVS then no toxic effects are seen. This is consistent with the results of a chemical
 equilibrium model for the sediment - interstitial water system (Di Toro et al., 1992). The
 resulting metal activity {M2*} can be related to the total SEM of the sediment and water, and
 to the solubility products of the metal sulfide (KMS) and iron sulfide (K^) . In particular, it is
 true.that at [SEM] s [AVS] then:
           [SEM^j£                                                           <4"2>

 Because the ratio of metal sulfide to iron sulfide solubility products (KMS/KFeS) is very small
 (< 10'5) even for the most soluble of the sulfides, the metal activity of the sediment is at least
 five orders of magnitude smaller than the SEM (see Di Toro et al. (1992) for data sources and
 references).  This indicates that no biological effects would be expected.  Therefore, the .
 condition [SEM] < [AVS] is a "no effect" ESG.
                                                        /

       The reason we use the term  "no effect" is that for the condition [SEM]  < [AVS] no
 biological impacts are expected. However, for [SEM] > [AVS], which might seemingly be
 considered a ESG violation, there are many documented instances where no biological impacts
 occur (e.g., because organic  carbon partitioning controls metal bioavailability hi the interstitial
.water, or the species of concern avoid or are insensitive  to metals).     '"

 4.2.2 Interstitial Water Guidelines

      • The condition [SEM]  <;  [AVS] indicates that the metal activity of the sediment -
 interstitial water system is low and, therefore, below lexicologically-significant

                                   Draft for SAB 1-89

-------
concentrations. Another way of ensuring this is to place a condition on the interstitial water
activity directly.  Suppose that we knew the metal activity, denoted by {FCV}, that
corresponded to the [FCV].  Then the ESG corresponding to this effect level is:
{M2
                                                                                   (4-3)
It is quite difficult, however,  to measure and/or calculate metal activity in a solution phase, at
the low concentrations required, since it depends on the identities, concentrations and
thermodynamic affinities of other chemically reactive species that are present.  Also, the WQC
are not expressed on an activity basis. An approximation to this condition is:
         [Md]£[FCVd]
                                                                       (4-4)
where [FCVJ is,the FCV applied to total dissolved metal concentrations. That is, we require
that the total dissolved metal concentration in the interstitial water [M J be less than the FCV
applied as a dissolved guideline.  Although this requirement ignores the effect of chemical
speciation on both sides of the equation - compare Equations (4-3) and (4-4) - it is .the
approximation that is currently being suggested by EPA for the WQC for metals (Prothro,
1993). That is, the WQC should be applied to the total dissolved - rather than the total acid
recoverable - metal concentration (Table 4-1; U.S. EPA,1995b). Hence, if this second
condition is satisfied it is consistent with the  level of protection afforded by the WQC.

       In situations where the SEM exceeds the AVS ([SEM] > [AVS]), but the interstitial
water total dissolved metal is less than the final chronic value ([MJ  <  [FCVJ), this sediment
would not violate the guidelines.  These cases occur when significant binding to other phases
occurs. It should be noted that using the FCV for metals hi freshwater samples requires that
the hardness of the interstitial water be measured since the WQC vary with hardness.

4.3    MULTIPLE METALS GUIDELINES

       As described in the previous subsection, and from a practical standpoint it is
insufficient and inappropriate to consider each metal separately because of the interactive
                                   Draft for SAB 1-90

-------
nature of metal-sulfide binding.  This is of particular concern for the AVS guidelines.


Table 4-1.  Water quality criteria (WQC) criteria continuous concentrations (CCC) based on
the dissolved concentration of metal"1'. These WQC CCC values are for use in the Interstitial
Water Guidelines approach for deriving sediment guidelines based on the dissolved metal
concentrations in interstitial water.            ,             -
Metal
Cadmium
Copper6
Lead
Nickel
Silver
Zinc
" OJ-.S.EPA. 1995b).
b Rounded all criteria
c Vnr ^iramnto rfi*» frt
Saltwater CCC, ng/L
9.3
3.1
8.1
8.2
NAf
81
to two significant figures.
>ofiu/atf*r fTY"1 at a ViarHn«»cc nf ^
Freshwater CCC, //g/Lc
CF" re<0-7S32lIn]-3-49°1 -
Q 950iyo.8M5ito(iiaKineis)]-i.«63)i
0 79 1 [e( 1 .273Ita(hantae«)H.705) j
Q 997fe(O.S460lto(luntae»)]-H.1645n
' NAf
0.986[e<0-8473[ta(tol*1*>s)J"t"OJ614)l
n inn O«H onn mo r*rc\j\ a™ n **
  0.94, and 1.6 //g cadmium/L; 6.2, 12, and 20/^g copper/L; 1.0, 2.5, and 6.1jug lead/L; 88,
  160, and 280/^g nickel/L; and 58, 108, and 187 ^g zinc/L.
  CF= Conversion factor to calculate the dissolved CCC for cadmium from the total CCC for
  cadmium: CF=1.101672-[(ln hardness)(0.041838)]
  The saltwater CCC for .copper is from the "Ambient Water Quality Criteria- Saltwater
  Copper Addendum" (U.S. EPA, 1995c).
  The silver criteria are currently under revision to reflect water quality factors that influence
  the criteria such as hardness, DOC, chloride and pH among other factors. Since silver has
  the smallest solubility product (see Table 2-2) and the greatest affinity for AVS, it would be
  the last metal to be released from the AVS or the first metal to bind to the AVS. Therefore,
  it is unlikely that silver would occur in the interstitial water. However in sediments
  contaminated with silver the user should be aware of the limitations in the above criteria for
  silver.  AVS Guidelines can be applied although the Interstitial Water Guidelines can not.
  If the AVS Guideline is exceeded (£SEM > AVS) and the sediment is contaminated with
  silver, further testing and evaluations would be warranted to access toxicity.
                                   Draft for SAB\\~9l

-------
4.3.1
AVS Guidelines
   The results of calculations using chemical equilibrium models indicate that metals act in an
                !
competitive manner when binding to AVS. That is, the six metals: silver, copper, lead,
cadmium, zinc and nickel will bind to AVS and be convened to their respective sulfides in this
sequence (i.e., in the order of increasing solubility).  Therefore, they must be considered
together.  There cannot be a guideline for just nickel, for example, since all the other metals
may be present as metal sulfides and, therefore, to some extent, as AVS. If these other metals
are not measured as SEM, then the £SEM will be misleadingly small, and it may appear that
[£SEM] < [AVS] when in fact this would not be true if all the metals are considered
together.  It should be noted that EPA currently restricts this discussion to the six metals listed
above; however, in situations where other sulfide forming metals (e.g.,  mercury) are present
at high concentrations, they also must be considered.

  The equilibrium model prediction of metal activity is similar to the single metal example
when a mixture of the metals is present. If the molar sum of SEM for the six metals is less  '
than or equal to the AVS, that is:
         £s [SEM.MAVS]
                                                                       (4-5)
then:
                                                                                  (4-6)
where [SEAMED] is the total SEM (^moles/L(bulk)) for the 1th metal. Thus the activity of each
metal, {Mj, is unaffected by the presence of the other sulfides.  This can be understood as
follows.  Suppose that the chemical system starts initially as iron and metal sulfide solids and
that the system proceeds to equilibrium by each solid dissolving to some extent.  The iron
sulfide dissolves until the solubility product of iron sulfide is satisfied.  This sets the sulfide
activity.  Then each metal sulfide dissolves until reaching its solubility. Since so little of each
dissolve relative to the iron sulfide, the interstitial water chemistry is not appreciably changed.
Hence, the sulfide activity remains the same and the metal activity adjusts to meet each
solubility requirement.  Therefore, each metal sulfide behaves independently of one another.
                                    Draft for SAB 1-92

-------
The fact that they are only slightly soluble relative to iron sulfide is the cause of this behavior.
Thus, the AVS Guidelines are easily extended to the case of multiple metals.
          ' '         •   '                       v,       .                         i
4.3.2     Interstitial Water Guidelines                    ,

   The application of the Interstitial Water Guideline to multiple metals is complicated, not by
the chemical interactions of the metals hi the sediment - interstitial water system (as in the case
with the AVS Guideline), but rather because of their possible toxic interactions. Even if the
individual concentrations do not exceed the FCV of each metal (FCVD). me metals could exert
additive effects that might result in toxicity (Besieger et al., 1986; Spear and Fiandt, 1986;
Enserink et al., 1991; Kraak et al.,  1994).  Therefore, to address this potential additivity, the
interstitial water metal concentrations are converted to toxic units (TUs) and these are
summed.  Since FCVs are used as the no effects concentrations these TUs are referred to as
interstitial water guidelines toxic units (IWGTUs). For freshwater sediments, the FCVs are
hardness dependent for all of the divalent metals under consideration, and thus, need to be
adjusted to the hardness of the interstitial water of the sediment being considered.  Because
there are no FCVs for silver in freshwater or saltwater, this approach is not applicable to
sediments containing significant concentrations of silver (i.e. , SSEM > AVS).   Since silver
has the smallest solubility product (see Table 2-2) and the greatest affinity for AVS, it would
be the last metal to be released from the AVS or the first metal to bind with AVS.  Therefore
it is unlikely that silver would occur in the interstitial water. For the i* metal with a total
dissolved  concentration [MKd], the IWGTU is:
A lack of exceedence of the ESG requires that the sum of the IWGTUs be less than or equal to
one:
Hence, the multiple metals guideline is quite similar to the single metal case (Equation 4-4),
except that it is expressed as summed IWGTUs.

                                   Draft for SAB 1-93

-------
   To summarize, the proposed ESG are as follows.  The sediment passes the ESG for the six
 metals if either of these conditions is satisfied:
(a)       AVS Guideline:   .


          £i [SEM.MAVS]

where
             V             .  '

£ i[SEMJ = [SEMcJ + [SEMoJ +
           f
(b)       Interstitial Water Guideline
                                                                                 (4-5)
                                           + [SEMNi] + [SEMJ + [l/2SEMAg]
                                                                                 (4-8)
where
                                               +  P
-------
characterizations of the SEM, AVS, and interstitial water concentrations would be appropriate
(Ankieyetal., 1994).

4.4       ESG FOR METALS VS. ENVIRONMENTAL MONITORING DATABASES
                     ,  .                                                            /
   The purpose of this  Section is to compare ESG based on SEM-AVS or IWGTUs to
chemical monitoring data from freshwater and saltwater sediments in the United States. This
comparison of AVS-SEM and interstitial water concentrations can indicate the extent of metals
contamination in the United States. When toxicity or benthic organism community health data
are available in conjunction with these concentrations it is possible to speculate as to potential
causes of the observed effects.

4.4.1     Data Analysis
                                                               i    ,
   Three sources were  identified which contain both AVS and SEM databases; one also had
                              /                                        •
data on concentrations  of metals in interstitial water.  Toxicity tests were also conducted on all.
sediments from these sources. The databases are from the Environmental Monitoring and
Assessment Program (EMAP) (Leonard et al., 1996a),  National Oceanic and Atmospheric
Administration, National Status and Trends Program (NOAA NS&T) (Wolfe et al., 1994;
Long et al., 1995; 1996) and from the Regional Environmental Monitoring and Assessment
Program (REMAP) (Adams et al., 1996).

Freshwater sediments:

   The AVS and SEM concentrations in the 1994 EMAP database from the Great Lakes were
analyzed by Leonard et al. (1996a). Forty-six sediment grab samples and nine core samples
were collected in the summer from forty-two locations in Lake Michigan. SEM, AVS, TOC,
interstitial water metals (when sufficient volumes were present), and 10-day sediment toxicity
to the midge,  Chironomits tertians, and the amphipod, Hyallela a&eca, were measured in
sediments collected by  the grab (Appendix C).

   The AVS concentrations vs. SEM-AVS differences from Appendix C are plotted in Figure
4-1. Grab sediment samples containing AVS concentrations below the detection limit of 0.05
^mol/g AVS are plotted at that concentration.  Forty-two of the 46 (91 percent) samples had

                                  Draft for SAB 1-95

-------
          10
      :§»
         -10
         -20
         -30 h
                                       1            10
                           Acid Volatile Sulfide (\tmol S/g)
100
                           Acid Volatile Sulfide (umol S/g)
Figure 4-1. SEM minus AVS values versus AVS concentrations in EMAP-Great Lakes
sediments from Lake Michigan. Data are from surficial grab samples only (this figure is
taken from Leonard et al., 1996, see data in Appendix C). The upper plot shows all
values, the lower plot has the ordinate limited to SEM minus AVS values between -10 and
+10.
                           Draft for SAB 1-96
                                                                                  t

-------
 SEM-AVS differences greater than 0.  Thirty-six of these had less than 1.0 ^mol of £SEM
 metal per gram sediment; and none had over 5.8 //moles/g of excess metal. In theory,
 sediments with SEM concentrations in excess of that for AVS have the potential to be toxic
 due to metals.  However, the majority of exceedence occur in places where the AVS is very.
 small and the amount of SEM is also very small. For these Lake Michigan sediments, a closer
 look at both interstitial water metal and toxicity test results is needed.  Measurement of the
 concentrations of metals in interstitial water can be used to determine if the excess metals are
 bound to other sediment phases, therefore, prohibiting toxicity due to interstitial metal.
 Interstitial water guidelines toxic units (IWGTU) can be calculated for each metal as the
 interstitial water concentration divided by the final chronic value for that metal.  Interstitial
 water volumes were sufficient to measure metals concentrations in 20 of the samples. The sum
.of the IWGTU for cadmium, copper, lead, nickel and zinc in these sediments was less than 0.4
 (Leonard et al., 199,6a).  In 10-d toxicity tests using Chironomus tentans and Hyalella azteca,
                                        '. •    •            l                  ^
 no toxicity was observed 81% of the 21 sediments not exceeding the ESG. They conclude that
 for the toxic sediments that did not exceed the metals ESG, the observed toxicity is not likely
 due to metals. Further, these sediments are unlikely to be contaminated by metals (Leonard et
 al., 1996a). These data demonstrate the value of using both SEM-AVS and IWGTUs to
 evaluate the risks of metals in sediments.

 Saltwater sediments:                                                    -

 : Saltwater data from a total of 398 sediment samples from five monitoring programs
 representing the eastern coast of the United States from  Chesapeake Bay to Massachusetts are
 included in Figure 4-2. The  EMAP Virginia Province database (U:S. EPA, 1996) consists, in
part, of 127 sediment samples collected from August to  mid-September 1993 from randomly
 selected locations in tidal rivers and small and large estuaries from the Chesapeake Bay to
 Massachusetts (Strobel et al., 1995).  The NOAA data is from Long Island Sound, Boston
Harbor and the Hudson River Estuary. Sediments were collected from 63 locations in the
 coastal bays and harbors of the Long Island Sound in August, 1991 (Wolfe et al., 1994).
 Sediment samples from 30 locations hi Boston Harbor were collected in June and July 1993
 (Long et al., 1996).  Sediment samples from 38 locations in the Hudson River Estuary were
 collected from March to May 1991 (Long et al., 1995).  Sediment samples were collected in
 the REMAP program from 140 locations  from the New  York/ New Jersey Harbor Estuary
 System (Adams et al., 1996). All of the above sediment grab samples were from
                                                      ,w
approximately the top 2 cm of undisturbed sediment.
                                  Draft for SAB 1-97

-------
         50

 CA

 <

 U
 w
   .
 O
 w
I
        •50
•100
       •150
       10.0
        5.0
                                                              J I I I If
           O EMAPVP
           0 REMAP NY-NJ HARBOR ESTUARY
           0 NOAALI,BO,HR
                                                        0
                                                        e
                                                      100
                                                                 1000
                                AVS (umol/g)
                                                                1000
                                AVS (umol/g)
Figure 4-2.  SEM minus AVS values versus AVS concentrations in EMAP-Estuaries
Virginian Province (U.S. EPA, 1996), REMAP-NY/NJ Harbor Estuary (Adams et al,
1996) and the NOAA NST Long Island Sound (Wolfe et al., 1994), Boston Harbor (Long
et al., 1996), and Hudson-Raritan Estuaries (Long et al., 1995); (see data in Appendix D).
The upper plot shows all values, and the lower plot has the ordinate limited to SEM minus
AVS values between -.10 and +10.
                          Draft for SAB 1-98

-------
   For saltwater sediments, molar concentration of AVS typically exceeds that for SEM (SEM-
 AVS <0) for most of the samples across the entire range of AVS concentrations (Figure 4-2).
 A total of,68 of the 398 saltwater sediments (17 percent) had an excess of metal, and only 4 of
 the 68 (6 percent) had over 2 /umol/g excess SEM. As AVS levels increase above this
 concentration fewer and fewer sediments have SEM-AVS differences that are positive; none
 occurred when AVS was >8.1 /umol/g.  Unlike the sediments from the freshwater EMAP
 survey in Lake Michigan, interstitial water was not measured in these saltwater  sediments.
 Only five of the 68 sediments (7 percent) having excess of up to 0.9 ^mol/g SEM were toxic
 in 10-d sediment toxicity tests with the amphipod Ampelisca abdita, whereas 79 of 330 (24
 percent) sediments having an excess of AVS were toxic. The data support the interpretation
 that (1) toxicity was NOT metals-related in the 79 sediments where AVS  was in excess over
, SEM; (2) metals might have caused the toxicity in the five toxic sediments having an excess of
 metal, but even in the absence of measurements of interstitial water metals concentrations, we
 speculate .that metals toxicity is unlikely because there was only £0.9 ^mol/g excess SEM (the
 molar concentration SEM most often exceeds that of AVS, in sediments having AVS
 concentrations <. 1 lonol/g); and (3) the absence of toxicity in sediments having an excess of
 SEM of up to 4.4 /zmol/g indicates mat significant metal-binding potential over that of AVS
                                                           t.
 existed in some sediments. Organic carbon concentrations of from 0.05% to 15.2% (average
 1.9 percent) provides for some of this additional metal-binding.
                                                                 v.
   The data above appear to suggest that in the United States direct toxicity caused by metals
 in sediments is extremely rare. .While this might be true, these data by themselves are
 inconclusive and it would be inappropriate to use the data from the above studies to reach this
 conclusion. All of the above studies were conducted in the summer when the seasonal
 biogeochemical cycling of sulfur should produce the highest concentrations of iron
 monosulfide which should make direct metal-associated toxicity less likely than in the
 winter/spring months. Accurate assessment of the extent of the direct ecological risks of
 metals in sediments requires mat sediment monitoring occur in the months of minimum AVS
 concentration; typically  November to early May. These yet to be conducted studies must
 monitor at a minimum SEM, AVS, and interstitial water metal and toxicity.  The data
 presented here are not intended to be used to draw conclusions about toxicity due to
 resuspension or bioaccumulation.
                                   Draft for SAB 1-99

-------
                                      SECTION 5

                                  IMPLEMENTATION

 5.1    CONSIDERATIONS IN PREDICTING METAL TOXICITY

        Results of the short- and long-term laboratory and field experiments conducted to date
 using sediments spiked with individual metals and mixtures of metals represent convincing
 support of the conclusion that an absence (but not necessarily a presence) of metal toxicity can
 be reliably predicted based upon metal:sulfide relationships and/or interstitial water metal
 concentrations.  In contrast, much confusion exists in the use of this convincing evidence to
 interpret the significance of metals concentrations in sediments from the field when toxicity
 and benthic community structure measurements are available.  In addition, the use of these
 observations as a basis for predicting metal bioavailability, or deriving ESG, raises a number
 of conceptual and practical issues related to sampling, analytical measurements and effects of
 additional binding phases.  Many  of these were addressed by Ankley et al. (1994), though the
 most salient to the proposed derivation of ESG are described below.

 5.2    SAMPLING AND STORAGE

        Accurate prediction of exposure of benthic organisms to metals is critically dependent
 upon sampling appropriate  sediment horizons at appropriate tunes.  This is because of the
 relatively high rates of AVS oxidation due to natural processes in sediments and the
 requirement that oxidation must be avoided during sampling of sediments and interstitial
 water.  In fact it is this seemingly labile nature that has led some to question the practical
 utility of using AVS as a basis for EqP-derived ESG for metals (Luoma and Carter, 1993;
 Meyer et al., 1994). For example, there have been many observations of spatial (depth)
 variations in AVS concentrations, most of which indicate that surficial AVS concentrations are
' less than those in deeper sediments (Besser et al., 1996;  Boothman and Helmstetter, 1992;
 Brumbaugh et al., 1994;  Hansen et al., 1996b; Hare et al., 1994; Howard and Evans, 1993;
 Leonard et al., 1996a; Liber et al.. 1996). This likely is due to oxidation of AVS at the
 sediment surface, a process that is enhanced by bioturbation (Peterson et al., 1996).  In
 addition to varying with depth,  AVS can vary  seasonally. For example, in systems where
 overlying water contains  appreciable oxygen during cold weather months, AVS tends to
                                   Draft for SAB 1-100

-------
 decrease, presumably due to a constant rate of oxidation of the'AVS linked to a decrease in its
 generation by sulfate-reducing bacteria (Herlihy and Mills, 1985; Howard and Evans, 1993;
 Leonard et al.,  1993).  Because of potential temporal and spatial variability of AVS, it appears
 that the way to  avoid possible under-estimation of metal bioavailability is to sample the
 biologically "active" zone of sediments at times when AVS might be expected to be present at
 small concentrations.  We recommend that at a minimum AVS and SEM measurements be
 made using surficial (0-2.0 cm)  sediments during the period from November to early May in
 aerobic aquatic ecosystems.  Minimum AVS concentrations may not always occur during cool-
 weather seasons; for example, systems that become anaerobic during the winter can maintain
 relatively large sediment AVS concentrations (Liber et al., 1996).  Therefore, seasonal
 measurements of AVS, SEM and interstitial metal concentrations may need to be determined.
 Importantly, the biologically active zones of some benthic communities may be within only the
 first few millimeters hi surface depth while in other communities the biologically active zones
 may be up to a  meter.  So in order to ensure sufficient characterization, multiple sediment
 horizons may require sampling of interstitial water, SEM and AVS to determine the potential
 for exposure to metals.

       The somewhat subjective aspects of these sampling recommendations have  been of
 concern. Recent research suggests that the transient nature of AVS may be over-stated relative
 to predicting the fate of all metal-sulfide complexes in aquatic sediments. Observations from
 the Duluth EPA laboratory made in the early 1990s indicated that AVS concentrations in
 sediments contaminated by metals such as cadmium and zinc tended to be elevated over
 concentrations typically expected in freshwater systems  (G.T. Ankley, unpublished data). The
probable underlying basis for these observations did not become apparent, however,  until a
recent series of spiking and metal-sulfide stability experiments. The field colonization study of
 Liber et al. (1996) demonstrated a strong positive correlation between the amount of zinc
added to test sediments and the resultant concentration of AVS hi the samples.  In  fact, the
initial design of then- study attempted to produce test sediments with as much as five-times
more  SEM^ (nomimal) than AVS; however, the highest measured SEM^ /AVS ratio
achieved was only slightly larger than 1.  Moreover, the expected surficial depletion and
seasonal variations in AVS were unexpectedly low in the zinc-spiked sediments. These
observations suggested that zinc sulfide, which comprised the bulk of AVS in the spiked
 sediments, was  more stable than the iron sulfide that presumably was the source of most of the
AVS hi the control sediments. The apparent stability of other metal sulfides versus iron
                       i

                                  Draft for SAB 1-101

-------
 sulfide also has been noted in laboratory spiking experiments with freshwater and saltwater
 sediments (Boothman et al,, 1996; De Witt et al.,  1996; Hansen et al., 1996b; Leonard et al.,
 1995; Peterson et al., 1996; Sibley et al.; 1996).

        In support of these observations, recent metal-sulfide oxidation experiments conducted
 by Di Toro et al. (1996b) have confirmed that cadmium and zinc form more stable sulfide
 solid phases than iron. If this is also true for sulfide complexes of copper, nickel and lead, the
 issue of seasonal/spatial variations in A VS becomes of less concern because most of the studies
 evaluating variations in AVS have focused on iron sulfide (i.e., uncontaminated sediments).
 Thus, further research concerning the differential stability of metal-sulfides, both from a
 temporal and spatial perspective, is definitely warranted.

 5.2.1  Sediments                               .

       At a minimum, sampling of the surficial 2.0 cm of sediment in between November and
 early May is recommended. A sample depth of 2.0 cm is more appropriate for remediation
 and monitoring. In some instances such as for dredging or where depths greater than 2 cm are
 important than sample depths should be planned based on particular study needs. Sediments
 can be sampled using dredges; grabs, or coring, but mixing of aerobic and anaerobic
 sediments must be avoided because the trace metal speciation in the sediments will be altered
 (See Bufflap and Allen, 1995 for detailed recommendations to limit sampling artifacts).
 Coring is generally less disruptive, facilitates sampling of sediment horizons and limits
potential metal contamination and oxidation if sealed PVC core liners are used.

       Sediments not immediately analyzed for AVS and SEM must be placed in sealed air-
tight glass jars and refrigerated or frozen.  Generally, 50 ml or more of sediment should be
added to nearly fill the jar.  If sediments are stored this way there will be little oxidation of
AVS even after several weeks. Sampling of the stored sediment from the middle of the jar
will further limit potential effects of oxidation on AVS. Sediments experiencing oxidation of
AVS during storage will become less black or grey if oxidized. Because the rate of metal-
 sulfide oxidation is markedly less than that of  iron sulfide, release of metal during storage is
unlikely.                                   ,              ,
                                   Draft for SAB 1-102

-------
5.2.2  Interstitial Water

       Several procedures are available to sample interstitial water in situ or ex situ. Carignan
et al. (1985) compared metals concentrations in interstitial water obtained by ex situ
centrifugation at 11,000 rpm followed by double filtration (0.45 jam and 0.2 or 0.03 ,wm) and
in situ diffusion samplers with a 0.002 fjm interstitial size.  For the metals of concern in this
guidelines document, concentrations of nickel and cadmium were equivalent using both
methods and concentrations of copper and zinc were higher and more variable using
centrifugation. They recommended the use of in situ dialysis for the study of trace
constituents in sediments because of its inherent simplicity and the avoidance of artifacts that
can occur with the handling of sediments in the laboratory.

       More recently Bufflap and Alien (1995) reviewed four procedures  for the collection of
interstitial water for trace metals analysis. These included ex situ squeezing and centrifugation
and in situ dialysis and suction filtration.  This paper should be read by those selecting a
interstitial water sampling method.  They observed that each method has its  own advantages
and disadvantages, and that each user must make their own choice given the inherent errors of
each method.  Importantly, interstitial water must be extracted by centrifugation or squeezing
in an inert atmosphere until acidified because oxidation will alter metal speciation. Artifacts
may be caused by temperature changes in ex situ methods that may be overcome by
maintaining temperatures to those in situ. Contamination of interstitial water by fine particles
is important in all methods as differentiation of paniculate and dissolved metal is a function of
interstitial size.  The use of 0.45 ,um filtration, while an often accepted definition of
dissolved, may result in laboratory to laboratory discrepancies. The use of suction filtration
devices is limited to coarser sediments, and they do not offer depth resolution. The use of
diffusion samplers is hampered by the time required for equilibrium (7-14 days) and the need
for diver placement and retrieval in deep waters. Acidification of interstitial water obtained
by diffusion or from suction filtration must occur immediately to limit oxidation.  Bufflap and
Allen (1995) conclude that in situ techniques have less potential for producing sampling
artifacts  than ex situ procedures. They concluded that of the in situ procedures, suction
filtration has the best potential for producing artifact free interstitial water samples directly
from the environment.  Of the ex situ procedures they concluded that centrifugation under a
nitrogen atmosphere followed immediately by filtration and acidification was the simplest
technique likely to result in an unbiased estimate of metal concentrations in interstitial water.

                                    Draft for SAB 1-103

-------
 At present, EPA recommends filtration of the surface water through 0.4 to 0.45 /u.
 polycarbonate filters to better define that fraction of aqueous metal associated with toxicity
 (Prothro, 1993). Thurman equates the organic carbon retained on a 0.45 micrometer glass-
 fiber filter to suspended organic carbon so that this filtration procedure under nitrogen
 atmosphere followed immediately by acidification is acceptable for interstitial waters.
 However in studies comparing collection and processing methods for trace metals, sorption to
 filter membranes or filtering apparatus has been identified when losses occur (Ozretich and
 Schults, 1998).  Ozretich and Schults, 1998 have recently presented a method combining
 longer centrifugation times with a unique single-step IW withdrawal procedure which has the
 potential for minimizing metal losses by eliminating the need for filtration.

       In contrast to the above recommendations, EPA recommends the use of dialysis
 samplers to obtain samples of interstitial water for comparison of measured concentrations of
 dissolved metals with WQC.  This is primarily because diffusion samplers obtain interstitial
 water with the proper in situ geochemistry thus limiting artifacts of ex situ sampling.  Further,
 EPA has found that in shallow waters where contamination of sediments is most likely,
 placement of diffusion samplers is easily accomplished and extended equilibration tunes are
 not a problem.  Secondly, EPA  recommends the use of centrifugation under nitrogen and
 double 0.45//m filtration using polycarbonate filters for obtaining interstitial water from
 sediments in deeper aquatic systems. Probably most importantly, the extremely large database
 comparing interstitial metals concentrations with organism responses from spiked and field
 sediment experiments hi the laboratory has demonstrated that, where the interstitial water toxic
 unit  concept predicted that metals concentrations in interstitial water should not be toxic,
toxicity was not observed when  either dialysis samplers or centrifugation were used (Berry et
al., 1996; Hansen et al., 1996a). Therefore, it is likely that when either methodology is used
to obtain interstitial water for comparison with WQC, if metals concentrations are below 1.0.
 IWGTU sediments should be acceptable for protection of benthic organisms.

5.3    ANALYTICAL MEASUREMENTS
                ».

       An important aspect to deriving "global" ESG values is that the methods necessary to
 implement the approach must be reasonably  standardized or have been demonstrated to
produce results that are comparable to those of standard methodologies.  Prom the standpoint
 of the proposed metal ESG, a significant amount of research has gone into defining
                                   Draft for SAB 1-104

-------
methodologies to obtain interstitial water and sediments (see Section 5-2 above), to extract
SEM and AVS from sediments, and to quantify AVS, SEM and the metals in interstitial water.
                   %                              .       '
5.3.1  Acid Volatile Sulfide

       The SEM/AVS extraction method recommended by EPA is that of Allen et al. (1993).
In terms of AVS quantification, a number of techniques have been successfully utilized
including gravimetric (Di Toro et al.', 1990; Leonard et al., 1993), colorimetric (Cornwell and
Morse, 1987), gas chromatography - photoionization detection (Casas and Crecelius, 1994;
Slotton and Reuter, 1995) and specific ion electrodes '(Boothman and Helmstetter,1992;
Brouwer and Murphy, 1994; Brumbaugh et al., 1994; Leonard et al., 1996b).  Allen et al.,
1993 report a limit of detection for 50% accuracy of 0.01 ^mol/g for a 10-g sediment sample
using the colorimetric method. Based on several studies Boothman reports a detection limit of
1 jumol AVS which translates to 0.1 /imol/g dry weight for a 10 g sediment sample using the
ion specific electrode method (personal communication).

5.3.2  Simultaneously Extracted Metal                                               -

       Simultaneously extracted metals are operationally-defined as metals extracted from
sediment into solution by the acid volatile sulfide extraction procedure. The "dissolved"
metals in this solution are also  operationally defined as the metal species which pass through
filter material used to remove the residual sediment, and thus are defined by the interstitial size
of the filtration material used. Common convention defines "dissolved" as metal species
<0.45-/xm in size. SEM concentrations measured hi sediments are not significantly different,
however, using Whatman 1 filter paper alone (< 1 l-/tm nominal interstitial size) or in
combination with a 0.45-jtm filter (W. Boothman, unpublished data).  SEM solutions
generated by the AVS procedure can be analyzed for metals, commonly including cadmium,
copper, lead, nickel, silver and zinc by routine atomic spectrochemical techniques appropriate
for environmental waters (e.g. inductively coupled plasma atomic emission or graphite furnace
atomic absorption spectrophotometry) (U.S. EPA, 1994b).  Because of the need to determine
metals at relatively low concentrations, additional consideration must be given to preclude
contamination during collection, transport and analysis (U.S. EPA,  1995d,e,f,g).
                                  Draft for SAB 1-105

-------
 5.3.3  Interstitial Water Metal

       Interstitial water can be analyzed for the metals cadmium, copper, lead, nickel, silver
 and zinc by routine atomic spectrochemical techniques appropriate for environmental waters
 (e.g. inductively coupled plasma atomic emission or graphite furnace atomic absorption
 spectrophotometry) (U.S.  EPA, 1994b). Because of the need to determine metals at
 concentrations at or below the threshold of biological effects (i.e., WQC concentrations);
 additional consideration must be given to preclude contamination during collection, transport
 and analysis (U.S. EPA, 1995d,e,f,g).  (See guidance on clean chemistry techniques in U.S.
 EPA, 1994c.) Generally, detection limits should be at *0.1IWGTU, because the toxic unit
 contributions of each of the metals must be summed.
                                                                         • -t'f
 5.4,   ADDITIONAL BINDING PHASES

       Although AVS is an important binding phase for metals, there clearly are other
physico-chemical factors that influence metal partitioning in sediments. In aerobic systems, .or
those with low productivity (i.e., where the absence of organic carbon limits sulfate
reduction), AVS plays little or no role in determining interstitial water concentrations of
metals.  For example, Leonard et al. (1996a) found that a relatively large percentage of
surficial sediments from open areas in Lake Michigan did not contain detectable AVS.  In fact
the great majority (42 of 46) of samples analyzed by Leonard et al. (1996a) contained less
AVS than SEM, yet interstitial water metal concentrations of cadmium, copper, nickel, lead
and zinc were consistently small or non-detectable.  Even in sediments where concentrations of
AVS are significant, other partitioning phases may provide additional binding capacity for
SEM (e.g., Ankley et al.,  1993; Calamono et al., 1990; Slotton and Reuter, 1995).  In aerobic
sediments both organic carbon and iron and manganese .oxides control interstitial water
concentrations of metals (Calamono et al,.  1990; Jenne, 1968; Luoma and Bryan, 1981;
Tessier et al., 1979). In anaerobic sediments, organic carbon appears to be an important
additional binding phase controlling metal partitioning, in particular for cadmium, copper and
lead (U.S. EPA, 1994a).

      Even in substrates with very little metal binding capacity (e.g., chromatographic sand),
surface adsorption associated with cation exchange capacity will control interstitial water  metal
concentrations to some degree (Hassan et al., 1996). Although an ideal ESG model for metals
                                  Draft for SAB 1-106

-------
 would incorporate all possible metal binding phases, current knowledge concerning
 partitioning/capacity of phases other than AVS is insufficient for practical application of a
 multiple phase model for deriving ESG in this sediment guidelines document.
5.5    PREDICTION OF THE RISKS OF METALS IN SEDIMENTS BASED ON EqP

       It is important to repeat that conclusions about sediment toxicity based on SEM-AVS
concentrations pertain only to cadmium, copper, lead, nickel, silver and zinc.  (1) When the
molar concentration of AVS exceeds that of SEM (negative SEM-AVS) sediment toxicity due
to these metals is unlikely and any observed toxicity is most likely from some other cause.
This is important because toxicity observed hi sediments having an excess of AVS is often
incorrectly assumed to disprove the EqP metals theory. The correct conclusion is that some.
factor other than metals caused the effect.  This can be further substantiated if the toxic unit
concept is applied to metal concentrations measured in interstitial water; the absence of
significant concentrations of metals coupled with the negative SEM-AVS are powerful
evidence that metals are an unlikely cause of the effect.  (2) Sediments can only be toxic from
the metals cadmium, copper, lead, nickel, silver and zinc when the molar concentrations  of
SEM exceed those of AVS (SEM-AVS differences are positive).  Measurements of interstitial
water concentrations of metals are invaluable hi demonstrating that the sediments are toxic
because of metals, and these measurements will provide insights into the specific metal(s)
causing the observed toxicity. (3) It is not uncommon for toxicity to be absent in sediments
having concentrations of SEM that exceed those of AVS (SEM-AVS is positive).  This is
because other metal binding phases in sediments often reduce the concentrations of
bioavailable metal. (4) When sediments are toxic, and SEM-AVS is greater than 0.0, the
toxicity may or may not be metals-related. Often sediments having SEM-AVS of up to 10
Mmoles SEM/g are not toxic because the excess metals are associated with other binding
phases.  Measurements of interstitial water concentrations of metals are invaluable in
demonstrating an absence or presence of bioavailable metal.
                                  Draft for SAB 1-107

-------
                                     SECTION 6

                             GUIDELINES STATEMENT

       The procedures described in this document indicate that, except possibly where a
locally important species is very sensitive, benthic organisms should be acceptably protected in
freshwater and saltwater sediments if any one or both of the following two conditions is
satisfied: (a) If the sum of the molar concentrations of SEM cadmium, copper, lead, nickel,
silver and zinc is less than or equal to the molar concentration of AYS or (b) the sum of the
dissolved interstitial water concentration of cadmium, copper, lead, nickel, silver and zinc
divided by their respective WQC is  less than or equal to 1.0.

(a)    AVS Guidelines:
          Ei [SEM.MAVS]
                                                                                (4-5)
where
£ ifSEMJ = [SEMcJ  +
(b)    Interstitial Water Guidelines
                                          + [SEMNi] +
         \-v  l  i,dj
         ^(FCVJ*
                                                                               (4-8)
where
       tM.d]
       7
If any one of these two conditions are violated, this does not mean that the sediment violates the
ESG and is unacceptable.  For example, if SEM exceeds AVS, or if the AVS in a sediment is non-
detectable, then condition (a) will be violated. However, if there is sufficient sorption to particles,
                                  Draft for SAB 1-108

-------
or organic carbon or other binding phases so that condition (b) is satisfied, then the sediment
meets the guideline and benthic organisms are acceptably protected from metals-induced sediment
toxicity.

       If both of these conditions are violated, or if the AYS Guideline is violated and the
sediment is contaminated with silver then there'is reason to believe that the sediment may be
unacceptably contaminated by these metals. Further testing and evaluations would, therefore, be
useful in order to assess actual toxicity and its causal relationship to the five metals.  These may
include acute and chronic tests with species that are sensitive to the metals suspected to be causing
the toxicity.  Also, in situ community assessments, sediment TIEs and seasonal characterizations
of the SEM, AYS and interstitial water concentrations would be appropriate (Ankley et al.,  1994).
       The ESG approaches are intended to protect benthic organisms from direct toxicity
associated with exposure to metal-contaminated sediments. They are not designed to protect
aquatic systems from metal release associated, for example, with sediment suspension, or the
transport of metals into the food web either from sediment ingestion or the ingestion of
contaminated benthos.  This latter issue, in particular, should be the focus of future research given
existing uncertainty in  the prediction of bioaccumulation of metals by benthos (Ankley, 1996).

       It is repeated here that these guidelines apply only to the six metals discussed in this
document, copper cadmium, lead, nickel, zinc and silver.  Procedures  for sampling and analytical
methods for interstitial water and  sediments are discussed in Section 5, Implementation.
                                   Draftjor SAB 1-109

-------
                                      SECTION?
                                                                         i

                                     REFERENCES

Adams, W.J., R.A. Kimerle and R.G. Mosher. 1985. Aquatic safety assessment of chemicals
       sorbed to sediments.  In Aquatic Toxicology and Hazard Assessment: Seventh Symposium.
       R.D. Card well, R. Purdy and R.C. Banner, Eds. STP 854. American Society for Testing
       and Materials, Philadelphia, PA, USA, pp. 429-453.

Adams, D.A., J.S. O'Connor, and S.B. Weisberg.  1996.  Draft Final Report Sediment Quality of
       the NY/NJ Harbor System, An Investigation under the Regional Environmental Monitoring
       and Assessment Program (R-EMAP). Monitoring and Assessment Branch, Division of
       Environmental Science and Assessment. United States Environmental Protection Agency,
       Region 2. Edison, New Jersey.                            ,      .

Allen, H.E., R.H. Hall and T.D. Brisbin.  1980.  Metal speciation effects on aquatic toxicity.
       Environ. Sci. Technol. 14:441-443.

Allen, H.E., G. Fu and B. Deng.  1993, Analysis of acid-volatile sulfide (AVS) and
       simultaneously extracted metals (SEM) for the estimation of potential toxicity in aquatic
       sediments.  Environ. Toxicol. Chem. 12:1-13.                   .

Anderson, D.M. and F.M.M. Morel. 1978.  Copper sensitivity of Gonyaulax tamarensis.
       Limnol. Oceanogr. 23:283-295.

Anderson, L.D., D.B. Kent, et al. 1994.  Batch experiments characterizing the reduction of
       Cr(VI) using suboxic material from a mildly reducing sand and gravel aquifer.  Environ.
       Sci. Technol. 28:178-185.

Ankley, G.T., G.L. Phipps,  E.N. Leonard, D.A. Benoit, V.R. Mattson, P.A. Kosian, A.M.
       Cotter, J.R. Dierkes, D.J.  Hansen and J.D. Mahony. 1991.  Acid volatile sulfide as a
       factor mediating cadmium and nickel bioavailability in contaminated sediments. Environ.
       Toxicol. Chem. 10:1299-1307.
                                 Draft for SAB 1-llQ

-------
 Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and J.L. Bennett.  1993. Predicting the
       acute toxicity of copper in freshwater sediments: Evaluation of the role of acid-volatile
       sulfide.  Environ. Toxicol. Chem.  12:315-320.

 Ankley, G.T., D.M. Di Toro, D.J. Hansen, J.C. Mahony, R.C. Swartz, R.A. Hoke, N.A.
       Thomas, A.W. Garrison, H.E. Allen and C.S. Zarba. 1994.  Assessing the potential
       bioavailability of metals in sediments: a proposed approach.  Environ. Management
       18:331-337.'
                     >              .                                •             -
 Ankley, G.T.  1996. Evaluation of metal/acid volatile sulfide relationships in the prediction of
       metal bioaccumulation by benthic macroinvertebrates. Environ. Toxicol.,Chem. 15:2138-
       2146.

 Ankley, G. T.,  D.M. Di Toro, D.J. Hansen and WJ. Berry. 1996. Technical basis and proposal
       for deriving sediment quality criteria for metals. Environ. Toxicol. Chem. 15:2056-2066.

 Barrow, N.J. and A.S. Ellis.  1986a.  Testing a mechanistic model,  m. The effects of pH on
       fluoride retention by a soil. J. Soil Sci. 37:287-295.

 Barrow, N J. and A.S. Ellis. 1986b. Testing a mechanistic model.  IV.  Describing the effects
       of pH on zinc retention by a soil. J. Soil Sci. 37:295-302.

Barrow, NJ. and A.S. Ellis.  1986c.  Testing a mechanistic model.  V. The points of zero salt
       effect for phosphate retention and for acid/alkali titration of a soil.  J. Soil Sci.  37:303-
       310.

Bergman,  H.L.  and E.J. Dorward-King. 1997. Reassessment of Metals Criteria for Aquatic Life
       Protection. Society of Environmental Toxicology and Chemistry (SETAC) Press:
       Pensacola, FL.

Berner, R.A. 1967.-  Thermodynamic stability of sedimentary iron sulfides.  Am. J. Sci. 265:773-
  v    785.
                                  Draft for SAB 1-111

-------
 Berry, W.J., DJ. Hanson, LD. Mahony, D.L.NRobson, D.M. Di Toro, B.P. Shipley, B. Rogers,
       J.^l. Corbin and W.S. Boothman. 1996.  Predicting the toxicity of metals-spiked
       laboratory sediments using acid-volatile sulfide and interstitial water normalization.
       Environ. Toxicol. Chem. 15:2067-2079.        ,
                      \                 .  •

 Berry, W. et al. 1999.  Predicting the toxicity of sediments spiked with silver. Environ. Toxicol.
       Chem.  18:40-48.

 Berry, W.J. and W.S. Boothman. 1999. Results of toxicity tests with Ampelisca abdita exposed
       to chromium-spiked sediments. Memorandum to the record. February 17, 1999.

 Besser, J.M., C.G. Ingersoll and J.P. Giesy.  1996. Effects of spatial and temporal variability of
       acid volatile sulfide on the bioavailability of copper and zinc in freshwater sediments.
       Environ. Toxicol. Chem. 15:286-293.
    i             -              ,
 Biesinger, K.E., G.M. Christensen and J.T. Fiandt. 1986.  Effects of metal salt mixtures on
       Daphnia magna reproduction. Ecotoxicol. Environ. Safety. 11:9-14.

 Blake, R.C., D.M. Choate, et al.  1993.  Chemical transformation of toxic metals by a
 ..     pseudomonas strain from a toxic waste site. Environ. Toxicol.  Chem.  12:1365-1376.

Boothman, W.S. and A. Helmstetter. 1992. Vertical and seasonal variability of acid volatile
       sulfides in marine sediments.  Research Report to the U.S. EPA Environmental Monitoring
       and Assessment Program, Narragansett, RI, USA.

Boothman, W., D. Hansen, W. Berry, D. Robson, A. Helmstetter and J. Corbin. 1996. Abstract:
       Variation of acid volatile sulfides and metals and biological response in field-exposed
       sediments. Society of Environmental Toxicology and Chemistry Annual Meeting.
       Washington, D.C.

Boulegue, J., C.J. Lord m and T.M. Church., 1982. Sulfur speciation and associated trace metals
       (Fe, Cu) in the interstitial waters of Great Marsh, Delaware. Geochim. Cosmochim. Acta.
       46:453-464.                                      ,
                                  Draft for SAB 1-112

-------
 Brouwer, H. and T.P. Murphy.  1994.  Diffusion method for the determination of acid-volatile
       sulfides (AVS) in sediment. Environ. Toxicol. Chem.  13:1273-1275.
                                                                  /
 Brumbaugh, W.G., C.G. Ingersoll, N.E. Kemble, T.W. May  and J.L. Zajcek.  1994.  Chemical
       characterization of sediments and interstitial water from the upper Clark Fork River and
       Milltown Reservoir, Montana.  Environ. Toxicol. Chem. 13:1971-1983.

 Buerge, LJ. and S.J. Hug.  1997. Kinetics and pH dependence of chromium (VI) reduction by .
       iron (II). Environ. Sci. Technol. 31:1426-1432.

 Bufflap, W.E. and H.E. Allen. 1995. Sediment interstitial water collection methods: A review.
       Water Res. 29: 65-177. Calder, E.R. Long and C.G. Ingersoll. 1996. Development and
     <•                i                    •
       evaluation of sediment quality guidelines for Florida coastal waters.  Ecotoxicology, In
       Press.

 Calamono, W., W. Ahlf and U. Forstner.  1990.  Exchange of heavy metals between sediment
       components and water.  In Metal Speciation in the Environment.  J.A.C. Broekart, S.
       Gucer and F. Adams, Eds. Springer-Verlag, Berlin, Germany, pp. 503-522.

.Call, D.J. et al.  1999. Silver toxicity to Chironomus tentans in two freshwater sediments.
       Environ. Toxicol. Chem.  18:30-39.

 Campbell, P. G. C.  1995. Interactions between trace metals and aquatic organisms: a critique of the
       free-ion activity model.  Metal Speciation and Bioavailability in Aquatic Systems. John
       Wiley & Sons.

 Carignan, R. 1984. Interstitial water sampling by dialysis: methodological notes. Limnol.
       Oceanogr. 20:667-670.

 Carignan, R., F. Rapin and A. Tessier. 1985. Sediment interstitial water sampling for metal
       analysis: a comparison of techniques. Geochim. Cosmochim. Acta. 49:2493-2497.

 Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A.  Kosian and A.M. Cotter. 1991. The role of
       acid volatile sulfide in determining cadmium bioavailability in sediments. Environ.

                                  Draft for SAB 1-113

-------
       Toxicol. Chem. 10:1309-1319.
Casas, A.M. and E.A. Crecelius. 1994. Relationship between acid volatile sulfide and the
       toxicity of zinc, lead and copper in ma line sediments.  Environ. Toxicol. Chem. 13:529-
       536.

Chapman, G.A. 1987. Establishing sediment criteria for chemicals - regulatory perspective.  In
       Fate and Effects of Sediment-Bound Chemicals in Aquatic Systems. K.L. Dickson, A.W.
       Maki and W.A. Brungs, Eds.  Pergamon Press, New York, pp
 355-376.
Cornwall, J.C. and J. W. Morse. 1987.  The characterization of iron sulfide minerals in anoxic
       marine sediments. Mar. Chem. 22:193-206.

Cranston, R.E. and J.W. Murray. 1978. The determination of chromium species in natural
       waters. Anal. Chim. Acta. 99:275-282.

Davies-Colley, R.J., P.O. Nelson and K.J. Williamson.  1985.  Sulfido control of cadmium and
       copper concentrations in anaerobic estuarine sediments.  Mar. Chem. 16:173-186.
Deng, B.L. and A.T. Stone.  1996.  Surface-catalyzed chromium (VI)
      comparisons of different organic reductants and different oxide surfaces
      Technol.  302484-2494.
reduction: reactivity
         Environ. Sci.
De Witt, T.H., R.C. Swartz, DJ. Hansen, W. J. Berry and D. McGovern. 1996. Interstitial
      metal and acid volatile sulfide predict the bioavailability of cadmium during a full life-
      cycle sediment toxicity test using the estuarine amphipod, Leptocheirus plumulosus.
      Environ. Toxicol. Chem. 15:2095-2101.

Di Toro, D.M., F. Harrison, E. Jenne, S. Karickoff, and W. Lick. 1987. Synopsis of Discussion
      session 2: Environmental fate and compartmentalization. In Fate and Effects of Sediment-
      Bound Chemicals in Aquatic Systems. K. L. Dickson, A.W. Maki and W.A. Brungs,
      Eds., Pergamon, New York, NY, USA, pp. 136-147..

Di Toro, D.M., J.J. Mahony, DJ. Hansen, K.J. Scott, M.B. Hicks, S.M. Mayr and M.S
                                 DraftforSAB 1-114

-------
       Redmond. 1990.  Toxicity of cadmium in sediments: the role of acid volatile sulfide.
       Environ. Toxicol. Chem. 9:1487-1502.

DiToro, D.M. et al.  1991.  Technical basis for the equilibrium partitioning method for
       establishing sediment quality criteria. 12.            •     .
              f                                                     . •
DiToro, D.M. et al.  1991.  Technical basis for the equilibrium partitioning method for
       establishing sediment quality criteria. Environ. Toxicol. Chem. 11(12): 1541-1583.

Di Toro, D.M., C. Zarba, D.J. Hansen, R.C. Swartz, C.E. Cowan, H.E. Allen, N.A. Thomas,
       P.R. Paquin and W.J. Berry.  1991.  Technical basis for establishing sediment quality
       criteria for nonionic organic chemicals using equilibrium partitioning. Environ. Toxicol.
       Chem.  10:1541-1583.

Di Toro, D.M., 1992. Personal communication relating to Di Toro et al., 1990 on titration of
       FeSbyCd2*.

Di Toro, D.M., J.D.  Mahony, DJ. Hansen, KJ. Scott, A.R. Carlson and G.T. Ankley.  1992.
       Acid volatile sulfide predicts the acute toxicity of cadmium and nickel in sediments.
       Environ. Sci. Techno!. 26:96-101.

Di Toro, D.M., J.D.  Mahony, and A.M. Gonzalez. 1996a Particle oxidation model of synthetic
       FeS and sediment acid volatile sulfide.  Environ. Toxicol. Chem. 15:2156-2167.

Di Toro, D.M., J.D.  Mahony, DJ. Hansen and W.J. Berry. 1996b.  A model of the oxidation of
       iron and cadmium sulfide in sediments. Environ. Toxicol. Chem. 15:2168-2186.

Durrani, P.J. and Bi Durrani.  1970.  Introduction to advanced inorganic chemistry. London,
       Longman.

Dzombak, D.A. and F.M.M. Morel.  1990. Surface Complexation Modeling, Hydrous Ferric
       Oxide.  New York, NY, John Wiley & Sons.

Eary, L.E. and D. Rai.  1987. kinetics of chromium (III) oxidation to chromium (VI) by reaction

                                 Draft for SAB 1-115

-------
       with manganese dioxide. Environ. Sci. Tech. 21:1187-1193.
Eary, L.E. and D. Rai.  1989.  Kinetics of chromate reduction by ferrous ions derived from
       hematite and biotite at 25°C. Am. J. Sci. 289:180-213.

Eary, L.E. and D. Rai.  1991.  Chromate reduction by subsurface soils under acidic conditions.
       Soil Sci. Soc. Am. J. 55:676-683.

Elovitz, M.S. and W. Fish.  1995.  Redox interactions of Cr (VI) and substituted phenols:
       products and mechanism.  Environ. Sci. Technol. 29:1933-1943.

Emerson, S., L. Jacobs and B. Tebo. 1983. The behavior of trace metals in marine anoxic
       waters: solubilities at the oxygen-hydrogen sulfide interface.  In Trace Metals in Sea
       Water. C. S. Wong.

Enserink, E.L., J.L.  Maas-Diepeveen andC.S. van Leeuwen.  1991. Combined toxicity of
       metals: an ecotoxicological evaluation. Water Res. 25:679-687.

Fendorf, S.E. and R.J. Zasoski.  1992.  Chromium (HI) oxidation by delta-
       MnO2.1.characterization. Environ. Sci. Technol.  26:79-85.

Fendorf, S.E., G.M. Lamble, et al. 1994. Mechanisms of chromium ((HI) sorption on silica, l.cr
                                                                structure spectroscopy.
           surface .structure derived by extended x-ray absorption fine
       Environ. Sci. Technol. 28:284-289.
Fendorf, S.E. and D.L. Sparks. 1994. Mechanisms of chromium (HI) sorption on silica.2. effect
       of reaction conditions. Environ. Sci. Technol.  28:290-297.

Fendorf, S.E. and G.C. Li.  1996. Kinetics of chromate reduction by ferrous iron.  Environ. Sci.
       Technol.  30:1614-1617.

Goldhauber, M.B. and I.R. Kaplan. 1974. The sulfur cycle.  In The Sea, Vol. 5, Marine
       Chemistry. E.D, Goldberg, Ed. J. Wiley & Sons, NY.,~USA, pp. 569-655.
                                  Draft for SAB 1-116

-------
                              \.
                                                                             rf
 Green, A.S., G.T. Chandler and E.R. Blood.  1993.  Aqueous-, interstitial water-, and sediment-
        phase cadmium: toxicity relationships for a meiobenthic copepod.  Environ. Toxicol.
        Chem. 12:1497-1506.

 Hansen, D.J. 1995. Memorandum to Catherine Fox. Subject: Assessment tools that can be used
        for the National Sediment Inventory. February 28, 1994.

 Hansen, D.J., W.J. Berry, J.D. Mahony, W,S. Boothman, D.L. Robson, G. T. Ankley, D. Ma,
        Q Yan and C.E. Pesch. 1996a.  Predicting toxicity of metal-contaminated field sediments
        using interstitial water concentrations of metals and acid-volatile sulfide normalization.
        Environ. Toxicol. Chem. 15:2080-2094.

' Hansen, D.J., J.D. Mahony, W.J. Berry, S.J. Benyi, J.M! Corbin, S.D. Pratt and M.B. Able.
        1996b.  Chronic effect of cadmium in sediments on colonization by benthic marine
        organisms: an evaluation of the role of interstitial cadmium and acid-volatile sulfide in
        biological  availability. Environ. Toxicol. Chem. 15:2126-2137.

 Hare, L.,  R. Carignan and M.A. Huerta-Diaz. 1994. An experimental study of metal toxicity
      :  and accumulation by bentnic invertebrates: implications for the acid volatile sulfide (AVS)
        model.  Limnol. Oceanogr. 39:1653-1668.

 Hassan, S.M., A.W. Garrison, H.E. Allen, G.T. Ankley and D.M. Di Toro.  1996.  Using the
        distribution coefficients of five trace metals in sandy sediments to select the lowest capacity
    ,    natural adsorbent. Environ. Toxicol. Chem.  15:2198-2208.

 Herlihy, A.T. and A.L. Mills. 1985. Sulfate reduction in freshwater sediments receiving acid
        mine drainage.  Appl. Environ. Microbiol. 49:179-186.

 Hesslein,  R.H. 1976. An in situ sampler for close interval interstitial water studies. Limnol.
        Oceanogr. 30:221-227.                                          .

 Hoke, R.A., G.T. Ankley»/A.M. Cotter, T. Goldenstein, P.A. Kosian, G.L. Phipps  and P.M.
                 s              •                    '                                 .
        VanderMeiden. 1994. Evaluation of equilibrium partitioning theory for protecting acute
        toxicity of field-collected sediments contaminated with DDT, DDE, and ODD to the

                                   Draft for £45 1-117

-------
       amphipod Hyalella azteca. Environ. Toxicol. Chem. 13:157-166.

Howard, D.E. and R.D. Evans. 1993.  Acid-volatile siilfide (AVS) in a seasonally anoxic
    *  mesotrophic lake: seasonal and spatial changes in sediment AVS. Environ. Toxicol.
       Chem.  12:1051-1058.

Hug, S.J., H.U. Laubscher, et al.  1997. Iron (III) catalyzed photochemical reduction of
       chromium (VI) by oxalate and citrate in aqueous solutions. Environ. Sci. Technol.
       31:160-170.

Ingersoll, C.G., P.S. Haverland, E.L. Bmnson, T.S. Canfield, F.S. Dwyer, G.E. Henkeand
       N.E. Kemble. 1996. Calculation of sediment effect concentrations. J. Great Lakes Res.
       In Press.                    •      "            .
                          s
Jenne,  E.A. 1968. Controls on Mn, Fe, Co, Ni, Cu, and Zn concentrations in soils and water-
       the'significant role of hydrous Mn and Fe oxides.  In Advances in Chemistry. American
       Chemical Society, Washington, DC, USA, pp. 337-387.

Jenne,  E.A. 1977. Trace element sorption by sediments and soil - sites and processes. In
       Symposium on Molybdenum in the Environment.  Vol. 2. W. Chappell and K.  Petersen,
       Eds. M. Dekker, Inc.,  New York.  (pp. 425-553).
                                     s>        '                     t
Jenne,  E.A., D.M. Di Toro, H.E. Allen and C.S. Zarba. 1986.  An activity-based model for
       developing sediment criteria for metals: a new approach. Chemicals in the environment.
       International Conference, Lisbon, Portugal.

Johnson, C. A. and A.G. Xyla.  1991 . The oxidation of chromium (HI) to chromium (VI) on the
       surface of manganite (gamma-MnOOH). Geochim.Cosmochim.Acta. 55:2861-2866.

Kaczynski, S.E. and R.J. Kleber. 1994. Hydrophobic C18 bound organic complexes of chromium
       and their potential impact on the geochemistry of chromium in natural waters. Environ.
       Sci. Technol. 28:799-804. <'

Kemble, N.E. et al.  1994. Toxicity of metal-contaminated sediments from the upper Clark Fork
                                 Draft for SAB 1-118

-------
       River, Montana, to aquatic invertebrates and fish in laboratory exposures. Environ.
       Toxicol. Chem. 13:1985-1997.

Kemp, P.P. and R.C. Swartz.  1986.  Acute toxicity of interstitial and particle-bound cadmium to
       a marine infaunal amphipod.  Mar. Environ. Res. 26:135-153.

Kraak, M.H.S., D. Lavy, H. Schoon, M. Toussaint, W.H.M. Peeters and N.M. van Straalen.
       1994. Ecotoxicity of mixtures of metals to the zebra mussel Dreissena polymorpha.
       Environ. Toxicol. Chem. 13:109-114.

Landers, A.H., M.B. David and M.J. Mitchel.  1983.  Analysis of organic and inorganic sulfur
       constituents in sediments, soils, and water.  Int. J. Environ. Anal. Chem. 14:245-256.

Leonard, E.N., V.R. Mattson, D.A. Benoit, R.A. Hoke and G.T. Ankley.  1993.  Seasonal
       variation of acid volatile sulfide concentration in sediment cores from three northeastern
       Minnesota lakes. Hydfobiologia. 271:87-95.

Leonard, E.N., V.R. Mattson and G.T.  Ankley. 1995. Horizon-specific oxidation of acid
       volatile sulfide in relation to the toxicity of cadmium spiked into a freshwater sediment.
       Arch. Environ. Contain. Toxicol. 28:78-84.

Leonard, E.N., G.T. Ankley and R.A. Hoke.  1996a.  Evaluation of metals in marine and
       freshwater surficial sediments from the Environmental Monitoring and Assessment
       Program relative to proposed sediment quality criteria for metals. Environ. Toxicol.
       Chem.  15:2221-2232.
  >                                                     '
Leonard, E.N., A.M. Cotter and G.T. Ankley.  1996b. Modified diffusion method for analysis
       of acid volatile sulfides and simultaneously extracted metals in freshwater sediment.
       Environ. Toxicol. Chem.  In Press.

Liber, K., D. Call,,!. Markee, K. Schmude, M. Balcer, F. Whiteman and G.T. Ankley.  1996.
       Effects of acid volatile sulfide on zinc bioavailability and toxicity to benthic
       macroinvertebrates in spiked-sediment field experiments.  Environ. Toxicol. Chem.
       15:2113-2125.                                                            f

                                  Draft for SAB 1-119

-------
Long, E.R. and L.G; Morgan.  1991. The potential for biological effects of sediment-sorbed
       contaminants tested in the national statwand trends program. NOAA Technical
       Memorandum NOS OMA 52.
            t
Long, E.R. et al. 1995. Incidence of-adverse biological effects within ranges of chemical
       concentrations in marine and estuarine sediments.  Environ. Management. 19(l):81-97.

Long, E.R., D.A; Wolfe, KJ. Scott, G.B. Thursby, E.A. Stern, C. Peven and T. Schwartz.
       1995. Magnitude and Extent of  Sediment Toxicity in the Hudson-Raritan Estuary.
       National Oceanic and Atmospheric Administration Technical Memorandum NOS ORCA
       88 NOAA/NOS Office of Ocean Resources, Conservation,  and Assessment, Silver Spring,
       MDUSA.                                                       rj

Long, E.R., G.M. Sloane, R.S. Carr, KJ. Scott, G.B. Thursby and T.L. Wade. 1996.  Sediment
       Toxicity in Boston Harbor: Magnitude and Extent, and Relationships with Chemical
       Toxicants. National Oceanic and Atmospheric Administration Technical Memorandum
       NOS ORCA 88 NOAA/NOS Office of Ocean Resources, Conservation, and Assessment,
       Silver Spring, MD USA. (REMAP)

Long, E.R., L.J. Field and D.D. MacDonald.  1998.  Predicting toxicity in marine sediments
       with numerical sediment quality guidelines.  Environ. Toxicol. Chem.  17(4):714-727.

Luoma, S.N. 1989.  Can we determine the biological availability of sediment-bound trace
       elements? Hydrobiologia 176/177:379-396.

Luoma, S.N. and W. Bryan.  1981. A statistical assessment of the form of trace metals in
       oxidized sediments employing chemical  extractants.  Sci. Total Environ. 17:165-196.

Luoma, S.N. and J.L. Carter.  1993. Understanding the toxicity of contaminants hi sediments:
       beyond the bioassay-based paradigm.  Environ Toxicol. Chem. .12:793-796.

MacDonald, D.D., R.S. Carr, F.D. Calder,  E.R. Long and C.G. Ingersol.  1996. Development
       and evaluation of sediment quality guidelines for Florida coastal waters. Ecotoxicology
     ^  5:253-278.     ..           ,
                                 Draft for SAB 1-120

-------
 Mahony, J.D., D.M. Di Toro, R. Koch, W. Berry and D. Hansen.  1993. Vertical distribution of
       AVS and SEM in bedded sediments, biological implications and the role of metal sulfide
       oxidation kinetics.  Abstract. Annual Meeting. Society of Environmental Toxicology and
       Chemistry Annual Meeting,  p.46.

 Mahony, J.D., D.M. Di Toro, A.M. Gonzalez, M. Curto and M. Dilg.  1996. Partitioning of
       metals to sediment organic carbon. Environ. Toxicol, Chem. 15:2187-2197.

 Masschelyeyn, P.M., J.H.  Pardue, R.D. DeLauna and W.H. Patrick, Jr. 1992. Chromium redox
       chemistry in a lower Mississippi Valley bottomland. Environ. Sci. Technol. 26:1217-1226.

 Mesuere, K. and W. Fish.  1992.  Chromate and oxalate adsorption of goethite. 1 .calibration of
       surface complexation models. Environ. Sci. Technol.  26:2357-2364.

 Mesuere, K. and Wi Fish.  1992.  Chromate and oxalate adsorption of goethite.2.surface
       complexation modeling of competitive adsorption. Environ.  Sci. Technol.  26:2365-2370.

 Meyer, J.S., W. Davison, B. Sundby, J.T. Oris, D.J. Lauren, U. Forstner, J. Hong and D.G.
       Crosby. 1994.  The effects of variable redox potentials, pH and light on bioavailability hi
       dynamic water-sediment environments.  In Bioavailability: Physi-cal, Chemical and
       Biological Interactions.

Morse, J.W., F.J. Millero, J.C. Cornwell and D. Rickard. 1987. The chemistry of the hydrogen
       sulfide and iron sulfide systems hi natural waters.  Earth Sci. Rev. 24:1-42.

Oakley, S.M., K.J. Williamson and  P.O. Nelson.  1980. The geochemical partitioning and
       bioavailability of trace metals hi marine sediments. Water Res. Inst., Oregon State Univ.,
       Corvallis, Oregon.  1-84.    .

Ohtake, H. and Hardoyo.  1992.  New biological method for detoxification and removal of
       hexavalent chromium.  Water Sci. Technol. 25:295-402.
                                                                              -j
Ozretich, R.J. and D.W. Schults. 1998. A comparison of interstitial water isolation methods
       demonstrates centrifugation with aspiration yields reduced losses of organic constituents.

                                  Draft for SAB 1-121

-------
       Chemosphere 36:603-615.                            .

Pankow, J.F. 1979. The dissolution rates and mechanisms of tetragonal ferrous sulfide
       (mackinawite) in anoxic aqueous systems. Ph.D thesis.  California Inst. of Technology,
       Pasadena, Calif.  1-146.         ,

Patterson, R.R., S. Fendorf, et al. 1997.  Reduction of hexavalent chromium by amorphous iron
       sulfide.  Environ. Sci. Technol.  31:2039-2044.          .

Persaud, D., R. Jacqumagi and A. Hayton. 1989.  Development of provincial sediment quality
       guidelines.  Ontario Ministry of the Environment, Toronto, ON, Canada.

Pesch, C.E., D.J. Hansen, W.S. Boothman, WJ. Berry and J.D. Mahony.  1995. The role of
       acid volatile sulfide and interstitial water metal concentrations in determining
       bioavailability of cadmium and nickel from contaminated sediments to the marine
       polychaete Neanthes arenaceodentata. EnyironT Toxicol. Chem. 14:129-14.

Peterson, G.S., G.T. Ankley and E.N. Leonard. 1996. Effect of bioturbation on metal-sulfide
       oxidation in surficial freshwater sediments.  Environ. Toxicol. Chem. 15:2147-2155.

Pettine, M. and F.J. Millero.  1990.  Chromium speciation hi seawater: the probable role of
       hydrogen peroxide.  Limnol. Oceanogr.  730-736.

Pettine, M., I. Barra, et al.  1998. Effect of metals on the reduction of chromium (VI) with
       hydrogen sulfide. Water Research.  32:2807-2813'.
                                                                               A.    t
Phillips, H.O. and K.A. Kraus. 1965. Adsorption on inorganic materials VI. Reaction of
       insoluble sulfides with metal ions in aqueous media.  J. Chromatography. 17:549-557.

Prothro, M.G.  1993. Office of Water Policy and Technical Guidance on Interpretation and
       Implementation of Aquatic Life Metals Criteria.  Letter from M.G. Prothro, Acting
       Assistant Administrator for the Office of Water, to Water Management Division Directors.
       U.S. EPA Regions I - X.  October 1, 1993.           ~                       ,
                                  Draft for SAB 1-122

-------
                                                        .
Rai, D., B.M. Sass, et al.  1987. Chromium (III) hydrolysis constants and solubility of chromium
       (III) hydroxide. Inorg. Chem.  26:345-349.

Richard, F.C. and A.C. Bourg.  1991. Aqueous geochemistry of chromium: a review. Wat.Res.
       25(7):807-816.

Schmieman, E.A., J.N. Petersen, etal. 1997.  Bacterial reduction of chromium. Appl. Biochem.
       Biotech. 63-5:855-864.

Schoonen, M.A.A. and H.L. Barnes.  1988. An approximation of the second dissociation
       constant for H2S. Geochim. Cosmochim. Acta. 52:649-654.

Schroeder, D.C. and G.F.  Lee.  1975. Potential transformations of chromium in natural waters.
       Water, Air and Soil Pollution.  4:355-365.

Sibley, P.K:, G.T. Ankley, A.M. Cotter and E.N. Leonard.  1996.  Predicting chronic toxicity of
       sediments spiked with zinc: an evaluation of the acid volatile sulfide (AVS) model using a
       life-cycle test with the midge Chironomus teutons. Environ. Toxicol. Chem.  15:2102-
       2112.   .        ,"".'•                                                -

                         '                •   \            ]
Silvester, E., L. Charlet, et al.  1995.  Mechanism of chromium (HI) oxidation by nabuserite. J.
       Phys. Chem. 99:16662-16669.

Slotton, D.G. and J.E. Reuter.  1995.  Heavy metals in intact and resuspended sediments of a
       California reservoir, with emphasis on potential bioavailability of copper  and zinc.  Mar.
       Freshwater Res. 46:257-265.

Smillie, R.H., K.  Hunter, et al.  1981. Reduction of chromium (VI) by bacterially produced
       hydrogen sulfide in a marine environment. Wat. Res. ,15:1351-1354.

Spehar, R.L. and J.T.  Fiandt. 1986.  Acute and chronic effects of water quality criteria-based
       metal mixtures  on three aquatic species.  Environ. Toxicol. Chem. 5:917-931.

Sposito, G, J.C.M. De Wit and R.H. Neal. 1988. Selenite adsorption on alluvial soils:  III.

                                  Draft for SAB 1-123

-------
       Chemical modeling.  Soil Sci. Soc. Am. J. 52:947-950.

Stephan, C.E., D.I. Mount, D.J. Hansen, J.H. Gentile, G.A. Chapman and W.A. Brungs. 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.  98pp.

Strobel, C.J., H.W. Buffum, S.J. Benyi, E.A. Petrocelli, D.R. Reifstick and DJ. Keith. 1995.
       Statistical summary EMAP-Estuaries Virginian Province 1992-1993. U.S. EPA/620/R-94/-
       26.  U.S. Environmental Protection Agency, Narragansett, RI. 72pp.

Stumm, W. and J.S.Morgan.  1996.  Aquatic Surface Chemistry, Chemical Equilibria and Rates
       in Natural Waters, Third Edition. John Wiley & Sons, Inc. New York.
                            x
Stumm,'W. and J.S. Morgan.  1981.  Aquatic Surface Chemistry - an Introduction Emphasizing
       Chemical Equilibria in Natural Waters. John Wiley & Sons, New York, NY, USA.
    \                            '
              \        •        ,                                 •
Stumm, W.  1987. Aquatic Surface Chemistry, Chemical Processes at the Particle-Water
       Interface. John Wiley & Sons, New York.

Sullivan, J., J. Ball, E. Brick, S Hausmann,  G. Philarski and D. Sopcich. 1985. Report of the
       technical subcommittee on determination of dredge material suitability for in-water
       disposal. Wisconsin Department of Natural Resources, Madison, WI, USA.

Sunda, W.G. and R.R.L. Guillard. 1976. The relationship between cupric ion activity and the
       toxicity of copper to phytoplankton.  J. Mar. Res. 34:511-529.

Sunda, W. G. and J.M. Lewis. 1978. Effect of complexation by natural organic ligands on the
       toxicity of copper to a unicellular alga, Monochrysis lutheri. Limnol. Oceanogr. 23:870-
       876.
   r            •  .                           t
Sunda, W.G., D.W. Engel and R.M. Thuotte. 1978. Effect of chemical speciation of toxicity of
       cadmium to grass shrimp, Palaemonetes pugio: importance to free cadmium ion. Environ.
       Sci. Technol. 12:409-413.
                                 Draft for SAB 1-124

-------
Swartz, R.C., G.R. Dittsworth, D.W. Schults and J.O. Lamberson.  1985. Sediment toxicity to a
       marine infaunal amphipod: cadmium and its interaction with sewage sludge. Mar.
       Environ. Res. 18:133-153.

Swartz, R.C., D.W. Schultz, T.H. DeWitt, G.R. Dittsworth and J.O. Lamberson. 1990. Toxicity
       of fluoranthene in sediment to amphipods: a test of the equilibrium partitioning approach to
       sediment quality criteria. Environ. Toxicol. Chem. 9:1071-1080.

Swartz, R.C., F.A. Cole, J. O. Lamberson, S.P. Ferraro, D.W. Schults, W.A. DeBen, Henry
       Lee n and R.J. Ozretich. 1994. Sediment toxicity, contamination, and amphipod
       abundance at a DDT- and dieldrin-contaminated site in San Francisco Bay. Environ.
       Toxicol. Chem. 13:949-962.

Takacs, M. J. 1988.  The oxidation of chromium by manganese-oxide: the nature and control of
       the reaction. Michigan State.

Tessier, A.,  P.G.C. Campbell and M. Bisson.  1979.  Sequential extraction procedure for the
       speciation of paniculate trace metals.  Anal. Chem. 51:844-851.

Tessier, A. and P.G.C. Campbell.  1987. Partitioning of trace metals in sediments: relationships
       with bioavailability.  Hydrobiologia 149:43-52.

Tessier, A.,  Y. Couillard, P.G.C. Campbell and J.C.  Auclair.  1993. Modeling Cd partitioning
       in oxic lake sediments and Cd concentrations in the freshwater bivalve Andonata grandis.
       Limnol. Oceanogr. 38:1-17.

Thurman, E.M.1985. Organic Geochemistry of Natural Waters. Kluwer Academic Publishers
       Group, Martinus Nijhoff/Dr. W. Junk Publishers, Boston, MA. pp. 497.

Troup, B.N. 1974. The interaction of iron with phosphate, carbonate, and sulfide in Chesapeake
       Bay interstitial waters: a thermodynamic interpretation. Ph.D thesis. Johns Hopkins
       University, Baltimore, Maryland. 1-114.                             ' -• -*v«.-

U.S. Environmental Protection Agency. 1980.  Ambient Water Quality Criteria for Silver -1980.
                                               '.                            ,
                                  Draft for SAB 1-125

-------
      Office of Water Regulations and Standards.  EPA-440/5-80-071.
                     1                    •             •
U.S. Environmental Protection Agency. 1984, Ambient Water Qualify Criteria for Chromium -
      1984. Office of Water Regulations and Standards.  EPA-440/5-84-029.

U.S. Environmental Protection Agency. 1985a. Appendix B - Response to public comments on
      "Guidelines for deriving numerical national water quality criteria for the protection of
      , aquatic organisms and their uses". July 19, 1985. Fed. Register. 50:30793-30796.

U.S. Environmental Protection Agency. 1985b.  Ambient Water Quality Criteria for Cadmium -
      1984. Office of Water Regulations and Standards. EPA 440/5-85^032.   .
                                                                      i
U.S. Environmental Protection Agency. 1985c.  Ambient Water Quality Criteria for Copper -
      1984. Office of Water Regulations and Standards. EPA 440/5-85-031.

U.S. Environmental Protection Agency. 1985d.  Ambient Water Quality Criteria for Lead -
      1984. Office of Water Regulations and Standards. EPA 440/5-85-027.

U.S. Environmental Protection Agency. J986. Ambient Water Quality Criteria for Nickel  -
      1986. Office of Water Regulations and Standards. EPA 440/5-85-004.

U.S. Environmental Protection Agency. 1987. Ambient Water Quality Criteria for Zinc -1987.
      Office of Water Regulations and Standards. EPA 440/5-85-032.

U.S. Environmental Protection Agency. 1989. Sediment classification methods compendium.
      Watershed Protection Division, U.S. EPA. 280pp.
                                                                       \

U.S. Environmental Protection Agency. 1994a.  Equilibrium partitioning approach to predicting
      metal bioavailability hi sediments and the derivation of sediment quality criteria for metals.
      Vol 1. Briefing Report to the Science Advisory Board.  EPA 822-D-94-002. Office  of
      Water, Office of Research and Development, Washington, DC, USA.

U.S. Environmental Protection Agency. 1994b. Methods for the determination of metals in
      environmental samples. Supplement I. EPA-600-R-94-111. Office of Research and ,
                                 Draft for SAB 1-126 ,

-------
       Development, Washington, DC.

U.S. Environmental Protection Agency. 1994c. Interim guidance on determination and use of
       water-effect ratios for metals. EPA-823-B-94-001. Office of Water, Washington, DC,
       USA.                           .

U.S. Environmental Protection Agency. 1995a. An SAB report: review of the agencies approach
       for developing sediment criteria for five metals. EPA-SAB-EPEC-95-020. Office of
       Water, Washington, DC, USA.

U.S. Environmental Protection Agency. 1995b. Water quality criteria for dissolved metals.
       Federal Register Notice, 60(86):22231-22237.

U.S. Environmental Protection Agency. 1995c. Ambient water quality criteria- saltwater copper
       addendum. April 14, 1995. U.S. Environmental Protection Agency, Office of Water,
       Office of Science and Technology, Washington, D.C.

U.S. Environmental Protection Agency. 1995d. Method 1669: Sampling ambient water for trace
       metals at EPA Water Quality Criteria Uvels. EPA-821-R-95-034. Office of Water,
       Engineering and Analysis Division (4303), Washington, DC 20460.

U.S. Environmental Protection Agency. 1995e. Method 1638: Determination of trace metals in
       ambient waters by inductively coupled plasma-mass spectrometry. EPA-821-R-95-031.
       Office of Water, Engineering and Analysis Division (4303), Washington, DC 20460.

U.S. Environmental Protection Agency. 1995f. Method 1639: Determination of trace metals in
       ambient waters by stabilized temperature graphite furnace atomic absorption. EPA-821-R-
       95-032. Office of Water, Engineering and Analysis Division (4303), Washington, DC
       20460.             .

U.S. Environmental Protection Agency. 1995g. Method 1636: Determination of hexavalent
       chromium by ion chromatography. EPA-821-R-95-029. Office of Water, Engineering and
       Analysis Division (4303), Washington, DC 20460.

U.S. Environmental Protection Agency. 1996. EMAP-Estuaries Virginian Province Data 1990-
       1993. Available from: EMAP Home Page WWW site, http://www.epa.gov/emap  via the
       INTERNET. Accessed 1997 Sept 17.

                                 Draft for SAB 1-127

-------
U.S. Environmental Protection Agency.  1997a.  The Incidence & Severity of Sediment
       Contamination in Surface Waters of the United States. Vol 1: National Sediment Quality
       Survey (EPA 823-R-97-006), Washington, D.C.

U.S. Environmental Protection Agency.  1997b.  The Incidence & Severity of Sediment
       Contamination in Surface Waters of the United States. Yol 2: Data Summaries for Areas
       of Probable Concern (EPA 823-R-97-007), Washington, D.C.

U.S. Environmental Protection Agency.  1997c.  The Incidence & Severity of Sediment
       Contamination in Surface Waters of the United States. Vol 3: National Sediment
       Contaminant Point Source Inventory (EPA 823-R-97-008), Washington, D.C.

U.S. Environmental Protection Agency.  1998a. Technical basis for establishing sediment quality
       criteria for nonionic organic chemicals by using equilibrium partitioning. (In preparation)
*                      •                           •               «
U.S. Environmental Protection Agency.  1998b. Users guide for multi-program implementation of
      , sediment quality criteria. (In preparation)
    ^                                   '•                    -   .         -
    '                   .
U.S. Environmental Protection Agency. 1998c. Guidelines for deriving site-specific sediment
       quality criteria for the protection of benthic organisms. (In preparation)

Vaughan, D.J. and J.R. Craig.  1978.  Mineral chemistry of metal sulfides.  Cambridge, UK,
       Cambridge University Press.  .-f£j$j$&

Wang, W-X, S.B. Griscom and N.S. Fisher. 1997.  Bioavailability of Cr(m) and Cr (VI) to
       marine mussels from solute and paniculate pathways. Environ. Sci. Technol. 31:603-511.

Wang, Y.T. and H. Shen.  1997.  Modelling Cr (VI) reduction by pure bacterial cultures.  Wat.
       Res. 31:727-732.             .

Wells, AJF. 1962.  Structural inorganic chemistry.  London, Oxford University Press.
                                            /
Wittbrodt, P.R. and C.D. Palmer.  1995.  Reduction of Cr (VI) in the presence of excess soil
       fulvic acid. Environ. Sci. Technol. 29:255-263.
Wolfe, D.A., S.B. Bricker, E.R. Long, K.J. Scott and G.B. Thursby. 1994.  Biological effects
       of toxic contaminants in sediments from Long Island Sound and environs. National

                                  Draft for SAB 1-128

-------
      Oceanic and Atmospheric Administration Technical Memorandum NOS ORCA 80
      NOAA/NOS Office of Ocean Resources Conservation and Assessment, Silver Spring, MD
      USA.
                                                    X
Zamuda, C.D. and W.G. Sunda.  1982.  Bidavailability of dissolved copper to the American
      oyster Crassostrea virginiea: Importance of chemical speciation. Mar. Biol. 66:77-82.

Zouboulis, A.I., K.A. Kydros, et al. 1995. Removal of hexavalent chromium anions from
      solutions by pyrite fines.  Wat. Res.  29:1755-1760.
                                 Draft for SAB 1-129

-------
           APPENDIX A
Glossary of Abbreviations and Equations
       Draft for SAB'1-130

-------
ACR        Acute-Chronic Ratio
Ag         Silver
AVS        Acid Volatile Sulfide                                  ,
ASTM      American Society for Testing and Materials
Cd         Cadmium
Cd          Freely dissolved interstitial water concentration of contaminant
r           Total interstitial water concentration of contaminant
Cs          Concentration of contaminant in sediment
Cte         Concentration of contaminant in sediment on an organic carbon basis
CCC        Criteria Continuous Concentration
CFR        Code of Federal Regulations
CLOGP     Computer program for generating partition coefficients
CMC       Criteria Maximum Concentration
CV         Coefficient of Variation
CWA       Clean Water Act                    .
DOC        Dissolved Organic Carbon
EDTA      Ethlyene  diamine tetra-acetic acid
EMAP      Environmental Monitoring and Assessment Program
ESG        Equilibrium Partitioning Sediment Guidelines
            Organic carbon-normalized Equilibrium Partitioning Sediment Guidelines
            Fraction of organic carbon in sediment
         1   Equilibrium partitioning
            Final Acute Value
            Final Chronic Value
            activity of Fe2* (mol/L)
[Fe2+]       concentration of Fe2* (mol/L)
[FeS(s)]     concentration of iron sulfide (mol/L)
[FeS(s)]j   '  initial iron sulfide concentration in the sediment (mol/L)
FeS         Iron monosulfide
FPV        Final Plant Value                  ' .      '      ..
FRV        Final Residue Value
GC/EC      Gas Chromatography/Electrpn Capture
GC/MS     Gas Chromatography/Mass Spectrometry
GFAA      Graphite Furnace Atomic Absorption
EqP
FAV
FCV
                                Draft for SAB 1-131

-------
Km
IWGTU     Interstitial Water Guidelines Toxic Unit
                         •\
IWTU       Interstitial Water Toxic Unit
             solubility product for FeS(s)[(mol/L)2]
             solubility product for MS(s) [(mol/L)2]
             Organic carbon: water partition coefficient
             Octanol: water partition coefficient
Kp          Sediment: water partition coefficient
Kgp          Solubility product constant  ':
LCSO        Concentration estimated to be lethal to 50 percent of the test organisms within a
             specified time period.
L           Liter
{M2*}       divalent metal activity (mol/L)                              *-
[M2*]        concentration of M2* (mol/L)
IM]A        concentraton of added metal (mol/L)
[MS(s)J      concentration of solid-phase metal sulfide (mol/L)
m1 or cam   Cubicmeter                                       /
ttg          Microgram                               '   ,
A*m          Micrometer
/imole       Micrbmole
mg          Milligram
mg/1        . Milligram per liter                                  •   •
ml          Milliliter
mm          Millimeter
NA          Not AppHcable, Not Available
ND          Not Determined, Not Detected
ng           Nanogram         .
Ni           Nickel  •    \  '
NOAA       National Oceanographic and Atmospheric Administration
NOEC       No Observed Effect Concentration
NST         National Status and Trends monitoring program
NTA        Nitrilotriacetic acid      ...
Pb          Lead
pH          Negative logarithm of the effective hydrogen ion concentration
OEC        Observed Effect Concentration
                                 Draft for SAB 1-132

-------
POC
ppb
ppm
Ppt
REMAP
{s2-}
[S2!
[SEM]
[SEMI*
SAB
SD
SLC
SEM
SOP
STORET
TDS
TOC
TU
TVS
U.S. EPA
WQC
Zn
OJJ2+
[2Fe(aq)]
[SM(aq>]
[2S(aq)]
Paniculate Organic Carbon
Parts per billion
Parts per million
Parts per trillion                          "A
Regional Environmental Monitoring and Assessment Program
activity of S2" (mol/L)
concentration of S2* (mol/L)
simultaneously extracted metal concentration (pmol/g)
simultaneously extracted Cd concentration (pmol/g)
   *       •                  -   •   .
simultaneously extracted Cu concentration (pmol/g)
simultaneously extracted Ni concentration Qimol/g)
simultaneously extracted Pb concentration (^mol/g)
simultaneously extracted Zn concentration (pmol/g)
U.S. EPA Science Advisory Board
Standard Deviation
Screening Level Concentration
Simultaneously Extracted Metals  ,.
Standard Operating Procedure
EPA's computerized water quality data base
Total Dissolved Solids
Total Organic Carbon
Toxic Unit
Total Volatile Solids
United States Environmental Protection Agency
Water Quality Criteria
Zinc                    ,
{Fe2+}/[LFe(aq)]
(M2+}/[Sm(aq)]
activity coefficient of Fe2*
activity coefficient of M2+
activity coefficient of S2"
concentration of total dissolved Fe(H) (mol/L)
concentration of total dissolved M(II) (mol/L)
concentration of total dissolved S(Q) (mol/L)
                                 Draft for SAB 1-133

-------
           APPENDIX B
Solubility Relationships for Metal Sulfides
         DrttfforSAB 1-134

-------
             Consider the following situation: a quantity of FeS is titrated with a metal that
forms a more insoluble sulfide. We analyze the result using an equilibrium model of the M-
(n)-Fe(n)-S(-n) system.  The mass action laws for the metal and iron sulfides are
                            - KMS                                              (B-l)
where [M2+], [Fe2+] and [S21 are the molar concentrations; YMi,, Yn» andYSj- are the
activity coefficients; and Km and KFe8 are the sulfide solubility products.  The mass balance
equations for total M(H), Fe(II) and S(-H) are   /
                                                                \

         o'1ll»4ft/2*]  + [MS(j)]  =  [Afl^                                          (B-3)
          a'1
          a'1 ,3-[S2-] * [MS(s)]  + rFeS(s)] = [FeS(s)].                             (B-5)
where
                                                                                 (B-6)
                                                                                 (B-7)
                                  Draft for SAB 1-135

-------
             - = [S21/[ES(aq)]
                                                                       (B-8)
are the ratios of the divalent species concentrations to the total dissolved M(H), Fe(n), and S(-
n) concentrations, [2M(aq)], [SFe(aq)], and [SS(aq)L respectively, [MS(s)] and [FeS(s)] are
the concentrations of solid-phase metal and iron sulfides at equilibrium. [FeS(s)]f is the initial,
iron sulfide concentration in the sediment, and [M]A is the concentration of added metal.

       The solution of these five equations can be obtained as follows.  The mass balance
Equations B-3 and B-4 for M(n) and FE(n) can be solved for [MS(s)J and [FeS(s)] and
substituted in the mass balance Equation B-5 for S(II):                    '
—a
                                                                                 
-------
••       Hence, the leading term in Equation B-10 must be small relative to [M]A and can safely be

           ignored.



                 The metal activity can now be found from the solubility equilibrium Equation B-l:
           so that
           where
           and
                                      Y.a-[Sa1
                                                       «-i   v      «-i  v   *~l
                                                       « Fe2tJSkFeS ^  °  M^^MS
                                                                                          (B-13)
                                             DrtfforSABl-m

-------
Equation A-13 can be expressed as
                                P.  a      S
                           Fe*+ *  PM*—
(B-16)
The magnitude of the term in parentheses can be estimated as follows. 'The first term in the
denominator is always greater than or equal to 1, $&*>. 1, because it is the reciprocal of two
terms both of which are less than or equal to 1, Equation B-14. they axe aFc2+ ,<_ 1, which is
the ratio of die divalent to total aqueous concentration, and yFe2+ <_ 1, which is an activity
coefficient. The second term in the denominator cannot be negative, PM^KM/KRS > 0» since
all of its terms are positive.  Thus, the denominator of the expression in parentheses is always
greater than 1, Pft2+ + PM2+KMs^ns >  *•  .Therefore, the expression in parentheses is always
less than 1. Hence, the magnitude of the ratio of metal activity to total added metal is
bounded from above by ratio of the sulfide solubility products:
          {Mea*}/[M]
(B-17)
This results applies if [FeS]; > [M]A so that excess [FeS(s)] is present.

      If sufficient metal is added to exhaust the initial quantity of iron sulfide, then [FeS(s)]
= 0.  Hence, the iron sulfide mass action equation (B-2) is invalid and the above equation no
longer applies.  Instead, the only solid-phase sulfide is metal sulfide and
          [MS] = [FeS].
(B-18)
so that, from the metal mass balance equation
                                  Draft for SAB 1-138

-------
i
          tills completes the derivation of Equations 2-8 and 2-9.
                                          Draft for SAB 1-139

-------
               APPENDIX C
Late Michigan EMAP Sediment Monitoring Database
             Draft for SAB MAQ

-------
Concentrations of SEM, AVS, TOC, and IWCTU for cadmium, copper, lead, nickel, and zinc in 46 surficial samples from Lake Michigan
1
Sample TOC
(*)
1 0.18
2 4.63
3 3.36
4 4.89
5 0.92
6 4.37
7 5.27
8 0.08
9 4.27
10 2.11
11 1.89
12 0.41
13 2.87
14 3.68
, 15 0.28
16 0.07
17 3.51
18 0.40
19 1.73
20 0.69
21 2.51
22 1.17
23 0.13
\ 24 1.03
r 25 0.63
26 0.30
27 0.29
28 0.21
29 0.11
30 0.05
31 0.27
32 4.95
33 ' 0.54
34 6.75
35 0.18
36 0.15
37 0.56
38 0.10
39 0.06
40 2.68
41 0.16
42 1.80
43 1.29
44 0.05
45 0.14
46 0.57
Source: Columns
SEM
AVS
SEM-
(umol/g) (umol/g) AVS
0.53
3.46
2.78
3.55
0.14
2.82
1.20
0.17
1.47
0.25
1.12
0.74
1.17
1.56
1.32
0.17
0.75
0.97
1.74
0.70
0.19
0.59
0.21
0.62
0.13
0.15
0.25
0.12
0.20
0.04
0.85
1.17
0.44
1.37
0.26
0.06
0.17
0.22
0.06
5.83
0.16
0.56
1.02
0.06
0.16
0.66
0.03
0.35
0.06
0.05
0.03
1.13
0.13
0.03
4.49
* 0.03
0.03
0.07
0.18
0.03
0.44
0.05
0.08
0.03
0.15
0.03
0.05
0.03
0.03
0.03
0.20
0.03
0.03
0.03
0.06
0.03
0.03
1.66
0.12
0.09
0.03
0.05
0.05
0.12
0.03
0.03
0.07
O.O3
2.25
0.03
0.05
0.03
for Sample, TOC,
0.51
3.11
2.72
3.50
0.12
1.69
' . 1.07
0.15
1-3.02
0.23
1.10
0.67
0.99
1.54
0.88
0.12
0.67
0.95
1.59
0.68
0.14
0.57
0.19
0.60
^0.07
0.13 '
0.23
0.10
0.14
0.02
0.83
-0.49
0.32
1.28
0.24
0.01
0.12
0.10
0.04
5.81
0.09
0.54
-1.23
0.04
0.11,,
0.64



IWCTU
Cadmium Copper Lead Nickel
-
0.029
0.018
0.018
0.0002
0.024
0.029
0.115
0.050
-
-
0.0002
-
0.0002
0.0002
'-
0.018
-
0.079
- '
-
-
-
-
.
- -
.
0.0002
0.0002
-
-
0.012
-
0.018
-
-
-
-
-
0.003
. •
0.006
0.0002
-
-•
-
.
0.003
0.308
0.266
0.034
0.049
0.003
0.003
0.034
-
-
0.070
-
0.003
0.119
- •
0.060
-•
0.013
-

-
-
-
.
-
-
0.155
0.003
-
-
0.036
-
0.041
• -
•
-
-
-
0.119
-
0.003
0.028
-

.
SEM, AVS, SEM-AVS and IWCTU
-
0.00004
.0.002
0.0004
0.0008
0.0002
0.0001
0.001
0.0008
-
-'
0.002
- •
0.0004
0.0002
-
0.0008
-
6.0008

-
-
-
-
.
-
.
0.0001
0.0004
-
-
0.0004
-
0.0002
-
.
-
-
-
0.001
.
0.0006
0.002
-
-
. -
taken directly
.
0.005
0.003
0.003
0.006
0.004
0.006
0.006
0.004
-
-
0.0005
.
0.006 .
0.004
•
0.008 .
-
0.010
-
-

.
-
.
- '
. .
0.011
0.007
- '
x
0.002
.
0.017
-
- .
.
'-
-
0.0005
-
0.008
0.0005
-.
-
-
% Survival
Zinc
.
0.003
0.029
0.006
0.032
0.020
0.020
0.055
0.026
.'
- '
0.001
.
0.015
0.050
.
0.058
.
0.020
-
-
-
-
-
.
.
_
0.0003
0.0003
.
-
0.020
-
0.012
-
-
.
- .
-
0.020
.
0.015
0.044
-
-
.
from Leonard ct al.,
Sum Hyalella
azleca
.
0.040
0.360
0.293
0.073
0.097
0.058
0.180
0.115
-
-
0.074
.
0.025
0.173
.
0.145
-
0.123
-
.
-
.
-
_
.
.
0.167
0.011
.'
.
0.070
.
0.088
.
.
.
-
-
0.144
.
0.033
0.075
-
-

1996*.
92.5
90
92.5
100
0
97.5
92.5
95
95 ,
77.5
97.5
_
97.5
96.5
90
100
100
95
97.5
97.5
75
97.5
57.5
72.5
95
.
35
75 '
80
97.5
97.5
97.5
100
95
95
95
_ .
60
97.5
90
62.5
75
100
82.5
-
70
Column
Chironoauu
teutons
40
90
90
97.5
90
100
100
87.5
100
87.5
100
.
97.5
92.5
87.5
100
100
100
97.5
97.5
923
100
65
57.5
90
_
35
72.5
82.5
100
97.5
95
100
90
100
92.5
-•
55
100
95
65
95
55
72.5
-
67.5
for survival
from fienonal communication with Leonard. 1998.
a AVS LOD=0.05 urn S/g
  Insufficient pore-water volume for metals analysis
  Cadmium LOD-0.01 ug/L (0.0002 IWCTU)
d Copper LOD-0.2 ug/L (0.0003 IWCTU)
e Lead LOD=0.1 ug/L (0.0001 IWCTU)
f Nickel LOD=0.5 ug/L (0.0005 IWCTU)
                                              Draft for SAB 1-141

-------
                                                           fl&
          APPENDIX D
Saltwater Sediment Monitoring Database
        Draft for SAB 1-142

-------
                                APPENDIX D
Concentrations of SEM, AVS. Toxicity and TQC for EMAP, NOAA NS &T and REMAP Database!
STUDY*
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
SEM
umol/g
.289
1.500
.066
.134
.266
.266
1.292
.347
.750
.212
.497
.624
.032
.988
.604
.031
1.597
1.065
.189
.018
.079
.421
.798
.903
1.202
.159
.246
.687
.699
1.663
.083
.740
.878
.044
.910
.567
.734
2.171
3.423
.197
.162
2.803 .
.472
2.079
AVS
umol/
1.400
.742
.029
.028
3.740
1.080
1.230
.087
.948
.283
.490'
13.400
.024
81.100
3.340
.331
72.400
8.480
6.460
.034
.976
3.210
68.000
3.150
67.700
3.310
4.870
2.420
- .430
116.000
1.300
.976
1.220
.025
3.430
.621
25.000
5.610
138.000
.892
3.590
11.900
12.500
26.600
SEM-AVS
ugmol/g
-1.111
.758
.037
.106
-3.474
-.814
.062
.260
-.198
-.071
.007
-12.776
.008
-80.112
-2.736
-.300
-70.803
-7.415
-6.271
-.016
-.897
' -2.789
-67.202
-2.247
•66.498 .
. -3.151 '
-4.624
-1.733
.269
-114.337
-1.217
-.236
-.342
.019
-2.520
-.054
-24.266
-3.439
-134.577
-.695
-3.428
-9.097
-12.028
-24.521
SURVIVAL*
100.
98.
99.
103.
99.
102.
107.
102.
99.
108.
. 103.
113.
101.
101.
107.
98.
102.
93.
103.
99.
97.
111.
104.
99.
105.
104.
106.
93.
91.
100.
»-.
101.
98. '
106.
104.
104.
107.
102.
100.
107.
. 82.
101.
101.
94.
SIGNIFICANCE"
0
0
0
0
0 .
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o ,.
0
0
0
0
0
0
0
0
0
.0
0
0
0
0 ,
0
0
0 .
0
TOC
.60
2.68
.17
.14
.49
.56
1.80
.30
.95
.37
1.00
1.58
.11
3.36
1.38
.09
4.19
3.17
J2
.15
.14
.49
2.84
2.85
2.28 ,
.51
.71
1.70
2.05
4.12
.14
2.30
2.84
.15
3.00
.76
2.21
2.57
4.14
.37
.81
2.36
2.77
3.18
                           Draft for SAB 1-143

-------
STUDY"
            SEM
                                           SURVIVAL*
SIGNIFICANCE' TOC
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
.445
2.228
.847
1.402
1.425 .
.263
2.936
.394
3.074
2.555
.452
.173
£78
.209
5.411
1.298
1.039
.960
7.369,
1.380
4.259
8.229
3.535
2.543
2.124
.188
.229
1.820
3.468
1.622
.693
.294
.178
.223
.239
.801
.751
.299
.341
.205
2.415
.632
1.516
3.249
.462
.043
.050
UJIIWJ
.056
15.100
17.300
52.700
22.300
' .079
29.600
.031
10.400
.402
.480
.201
.257
3.460
17.800
.228
.705
12.900
3.460
2.270
54.600
68.000
61.800
35.600
35.600
.836
.692
.227
14.600
6.080
1.200
.026
.074
.087
1.120
5.120
.090
.090
.174
.611.
4.050
28.200
52.700
12.300
6.140
.024
.025
™ 	 J!KSZ!i£..
.389
-12.872
. -16.453
-51.298
-20.875
, .184
-26.664
.363
-7.326
2.153
-.028
. -.028
.321
-3.251
-12.389
1.070
.334
-11.940
3.909
-.890
-50.341
-59.771
-58.265
-331057
-33.476
-.648
-.463
. 1.593
-11.132
-4.458
-.507
.268
.104
.136
-.881
-4.319
.661
.209
.167
-.406
-1.635
-27.568
-51.184
-9.051
-5.678
.019
.025
	 5_ 	 J
106.
103.
99.
109.
88.
84.
100.
87.
104.
96.
100.
98.
101.
96.
100.
100.
102.
94.
87.
97.
76.
43.
99.
\
33.
0.
108.
95.
104.
102.
102.
99.
95.
81.
104.
88.
92.
102.
104.
105.
95.
100.
88.
85. "•
103.
108.
100.
102.
J6_ 	
0
0
o
0
0

0
0
o
0-
0

o
0'
o
0
0

0
0
_
0
0
.
o
0
0
0
o
o
o
o
o
o
o
o
o
• 0
o
o
0
o
0
o
0
*)f\
•Zil
2.92
t 4o
^.Jp
t 1t\

-------
,m
STUDY*

EMAP-VA
EMAP-VA -
EMAP-VA ^
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
EMAP-VA
NOAA- U
NOAA-U
NOAA- LI
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA- LI
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA- LI
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA- LI
NOAA-U
NOAA- LI
NOAA- LI
NOAA-U
NOAA- LI
NOAA- LI
NOAA- LI
SBM
umol/g_
1.177
.624
,.799
,.020
.088
2.220
.813
.851
.701
1.113
.601
1.505
.701
.717
2.163
.616
2.368
1.278
2.253
.865
.950
1.113
1.026
1.446
2.777
.211
2.665
2.813
1.235
2.198
3.624
3.594
1.342
2.462
.964
.332
2.311
.623
.896
.544
.641
.355
.222
2.262
1.307
1.963
2.785 ,
AVS
umol/
3.460
6.210
29.700
.259
4.150
59.600
.381
.029
3.600
3.510
6.440
18.730
5.630
13.090
65.310
6.940
19.990
4.710
59.590 .
3.880
16.520
14.950
.850
12.480
29.720
.090
78.900
35.050
2.080
14.690
21.800
27.410
37.970
46.450
1.000
4.010
79.890
6.610
16.370
2.170
2.060
1.390
4.180
39.960
.380
, 51.820
61.020
SEM-AVS SURVIVAL* SIGNIFICANCE' TOC
uemol/g.
-2.283
-5.586
-28.901
-.239
-4.062
-57.380
.432
.822
-2.899
-2.397
-5.839
-17.225
-4.930
-12.373
-63.147
•6.324
-17.622
-3.432
-57.337
-3.015
-15.570
-13.837
.176
-11.034
-26.943
.121
-76.235
-32.237
-.844 '
-12.492
-18.176
-23.816
-36.628 (
-43.988
. -.036
-3.678
-77.579
-5.987
-15.475
-1.626
-1.419
-1.035
-3.958
-37.698
.927
-49.857
-58.235
% 	 ___*_
100.
104.
400.
96.
100.
74.
93.
87.
100.
96.
96.
93.
93.
93.
92.
92.
91.
91.
91.
91.
90.
89.
88.
88.
87.
87.
87.
86.
84.
84..
83.
82.
.82.
**•
81.
81.
81.
80.
80.
79.
79.
79.
77.
77.
76.
76.
76.

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0 •
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o
0
0
0
0
.
.
.
.

.
1.
- 1.

1.83
2.25
4.10
.30
.25
2.18
.98
.57
.74
1.12
1.43
2.56
.77
2.05
.3.22
.81
. 3f02
1.81
2.51
1.32
1.52
2.00
1.63
2.05
2.81
.54
3.33
3.83
1.58
2.80
2.48
2.59
1.85
3.18
1.60
1.29
3.69
.67
1.11
.27
1.56
.64
.45
2.67
1.56
3.46
3.81
                                         Draft for SAB 1-145

-------
STUDY* SEM AVS SEM-AVS
	 _ — JHJ»Vg._ 	 umol/ ugmol/g.
NOAA-LI
NOAA-LI
NOAA-LI
NOAA-U
NOAA-LI
NOAA-U
NOAA-U
NOAA-U
NOAA-LI
NOAA-LI
NOAA-LI
NOAA-U
NOAA-U
NOAA-LI
NOAA-LI
NOAA-U
NOAA-LI
NOAA-U
NOAA-U
NOAA-U
NOAA-U
NOAA-LI
NOAA-U
NOAA-U
NOAA-BO
NOAA-BO
NOAA-BO
NOAA-BO
NOAA- BO
NOAA- BO
NOAA-BO
NOAA-BO
NOAA-BO
NOAA-BO
NNOAA-BO
NOAA-BO
NOAA- BO
NOAA-BO
NOAA- BO
NOAA-BO
NOAA-BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA- BO
NOAA-BO
4.333
1.927
.004
3.831
.808
1.783
2.622
.597
1.181
1.862
2.726
2.102
2.471
1.870
1.607
4.942
2.705
2.087
1.514
2.629
3.194
.872
1.080
.123
2.914
2.218
2.609
3.650
1.634
1.267
2.892
2.511
.661
2.458
1.872
.959
2.480
.784
.943
1.683
1.753
2.447
1.839
1.296
1.697
1.390
2.310
16.060
3.710
24.580
9.250
.960
40.630
61.840
1.090
3.730
50.390
62.760
33.630
7.220
17.120
, 17.810
100.800 .
83.010
26.730
30.880
32.050
. 35.390
25.810
11.300
5.310
2.893
2.369
43.959
101.984
5.237
3.256
80.584
2.241
13.490
23.077
48.062
53.288
7.599
22.486
8.831
42.399
17.697
10.958
68.306
56.838
9.089
43.801
51.857
-11.747
-1.783
-24.576
-5.419
-.152
-38.847
-59.218
-.493,
-2.549
-48.528
-60.034
-31.528
-4.749
-15.250
-16.203
-95.858
-80.305
-24.643
-29.366
-29.421
-32.196
-24.938
-10.220
-5.187
.021
-.151
-41.350
-98.334
-3.603
-1.989
-77.692
.270
-12.829
-20.619
•46.190
-52.329
-5.119
-21.702
.-7.888
-40.716
-15.944
-8.511
-66.467
-55.542
-7.392
-42.411
**9.547
SURVIVAL* SIGNIFICANCE' TOC
% %
75. i. ,„«
75.
74.
73.
71.
70.
70.
69. ,
68.
67.
67.
64.
63.
61.
59.
54.
53.
47.
42.
39.
37.
34.
16.
10.
8.
15.
26.
29.
36.
52.
83.
86.
87.
87.
89.
90.
90.
91.
91.
92.
94.
94.
95.
96.
97.
97.
97.
1 i £n
*• Ji*OU
1. * 1 87
*« » . ^.o/
1. JtA
3.08
11 10
• i.iy
It  1.5J •
1 4 A1
*• 'i.yi
1 ' tt
*• .22
1A t*r
3.05
I t fid
*• 2.99
1 11A.
*• J./4
1 1 B1
A. I.o3
' 1 1 T5
*• I./*
'' 1 1 «
* • 1 .3;9
0 fiOR
v . . w.yo
0 2 19
w A. & A
0 1 00
^ &.W
0 31^
^* . <9«£J
0 3 25
w ^.^7
0 299
A.J7
0 A 4^
^^ "r.^J
O • 1 ftfi
^^ X.Oo
0 1 7ft
*. *o
0 3 4.1
*• p j «^i
0 • 1 41
.— •"*
0 ' 4 45
^ •»»*^rf
0 ' 254
^ +rftt^
0 . 3.05
0 2.68
0 3.27
0 3J5
Draft for SAB 1-146

-------
STUDY1

NOAA- BO
NOAA-BO
NOAA- BO
NOAA- BO
NOAA-BO
NOAA- BO
NOAA-BO
NOAA-HR
NOAA-HR
NOAA-HR
NOAA- HR
NOAA-HR
NOAA- HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA- HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA- HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA-HR
NOAA- HR
NOAA- B A
REMAP-BA
SEM
umol/g_
.399
2.481
1.736
.958
9.192
1.525
.678
5.037
4.202
1.174
1.855 ,
3.092
2.997
2.581
2.869
5.442 .
2.618
5.061
2.376
6.998
4.480
4.662
5.896
3.103
1.662
3.512
.273
.335
1.664
2.674
5.532
4.029
4.614
3.379
4.240
4.303
5.209
4.801
4.697
2.600
1.013
1.527
.505
3.341
3.449
.270
.341
AVS
^ URIOI/
3.899
19.604
148.969
18.622
120.622
81.842 •
5.679
69.320
21.980
27.540
14.170
51.770
79.710
61.050
28.080
25.900
1.080
12.240
4.390
63.450
20.780
23.720
51.580
59.780
7.230
25.840
.050
.036
18.760
3.630
29.210
18.440
20.530
30.120
19.320
22.570
14.570
35.370
54.710
56.730
10.160
15.130
.630
43.920
37.860
.950
.156
SEM-AVS
ugmol/g_
-3.500
-17.123
-147.233
-17.664
-111.430
-80.317
-5.001
-64.283
-17.778
-26.366
-12.315
•48.678
-76.713
-58.469
-25.211
-20.458
1.538
-7.179
-2.014
-56.452
-16.300
-19.058
•45.684
-56.677
-5.568
-22.328
.223
.299
-17.096
-.956
-23.678
-14.411
-15.916
-26.741
-15.080
-18.267
-9.361
-30.569
-50.013
-54.130
-9.147
. -13.603
-.125
^10.579
-34.411
-.680
.185
SURVIVAL*
•*
99.
99.
99.
99.
100.
102.
103.
0.
41.
11.
18.
101.
112.
119.
81.
95.
109.
97.
108.
0.
20.
14.
2.
77.
19.
0.
91.
93.
69.
3.
96.
51.
91.
88.
101.
102.
101.
70.
38.
37.
. 29.
68.
105.
86.
76.
96.
M.
SIGNIFICANCE' TOC
%
0
0
0
0
0
o
0
.1.
1.
1.
1.
0
0
0
. 0
0
0
o
0
1.
1.
1. .
• 1.
1.
1.
1.
o .
0
1.
1.
0
' 1.
0
0
0
0
0
1.
1.
1.
1.
1.
0
0
1.
0
0


.80
3.31
2.94
1.77
4.61
2.96
1.45
5.02
3.47
1.88
4.44
3.86
3.09
2.86
2.50
2.20
2.67
2.98
2.49
1.98
2.98
3.19
4.78
3.99
2.61
4.44
.07
.07
.69
1.00
3.18
2.20
1.94
2.80
3.15
3.02
3.21
2.98
3.47
1.47
.77
.95
.25
2.55
3.63
.26
.06
Dn^ for SAB 1-147

-------
STUDY-
j, '•^••^•••. ^.L
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-BA
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
SEM
umol/&
.888
.722
.362
2.138
3.008
.151
.115
.543
.103
.167
,073
.294
.120
.109
.185
.120 .
.347
.120
- 2.275
.344
.258
..119
.258
.494
.109
. .266
v 327
.230
2.026
14.550
3.332
3.763
.357
.524
.244
1.247 -'
2.478
1.744
.131
.846
4.399
3.884
.673
3.150
.270
.162
2.880
AVS
umol/
12.971
4.948
.936
3.295
3.941
.555
.156
.156
.156
.932
.156
.156
.156
.156
.156
.156
.156
.156
16.592
.012
.343
.156
.156
.156 '
.156
.156
.393
6.400
47.793
389.857
243.322
201.687
10.923 .
3.974
4.502
48.130
47.376
.156
1.184
.927 .
116.954
237.650
21.769
43.975
4.491
.873
153.755
SEM-AVS
	 wgmgl/g.
-12.083
-4.226
-.574
-1.157
-.933
-.404
-.041
.387
-.053
-.765
-.083
.138
-.036
-.047
.029
-.036
.191
-.036
-14.317
.332 7
-.085
-.037
' .102
.338
-.047
.110
-.066
-6.170
-45.767
-375.307
-239.990
-197.924
-10.566
-3.450
-4.258
-46.883
-44.898
1.588
-1.053
-.081
-112.555
-233.766
-21.096
-40.825
-4.221
-.711
-150.875
SURVIVAL*
%
92-
85.
98.
95.
95.
96.
99.
94.
85.
97.
; 99.
91.
84.
92.
90.
88.
89.
81.
69.
91.
94.
84.
91.
86.
89.
86.
93.
83.
51.
0.
37.
79.
95.
; 98.
84.
91.
36.
69.
94. .
73.
93.
89.
77.
91.
91.
98.
92.
SIGNIFICANCE'
	 _%
0
0
0
0
0
0
0
0
0
0
0
0
6
0
0
0
0
0
1.
0
0
0
/
0
0
0
0
0
0
1.
1.
1. -
1.
0
0
0
0
1.
1.
0
1.
0
0
I.
0
0
0
0
TOC

4.05
.40
.26
.43
.18
.15
.08
.07
.05
.16
.05
.34
.83
.92
4.48
.83
1.26 .
62
• *WA
1.81
• 3 a<
J.O^
.77
223
*••<&«>
.88
2.10
4.07
1.06
.29
.19
' .77
1 55
I.^A
.83
.97
.26
.35
.27
.54
1.12
1.14
.21
1.58
6.55
8.45
4.11
5.47
.74
1.40
7.70
Draft for SAB 1-148

-------
i
STUDY-

REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB
REMAP-JB

REMAP-JB
REMAP-JB
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-LS
REMAP-NB .
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB

REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
SEM
umol/g^
.323
.413
.377
.099
1.100

.209
.213
.954
2.759
.711
1.915
2.186
2.480
.606
3.289
3.241
.616
1.506
2.485
1.894
3.149
.632
1.057
.638
1.087
3.711
2.990
8.894
1.277
3.925
5.632
6.809
7.645
4.012
3.905
.942
3.515
2.216
3.323
3.391
3.443

2.466
2.294
5.768
1.013
2.479
.554
AVS
urool/
1.684
3.056
3.056
.686
58.945

1.466
.780
1.542
6.498
10.240
12.596
17.605
23323
"2.501
91.773
56.100
1.070
26.201
28.248
25.394
64.643
1.310
4.647
.218
.312
17.184
59.256
60.816
23.266
42.727
114.770
135.354
150.012
43.663
26.229
6.531
7.134
11.243
7.573
4.820
3.982

20.273
11.046
5.028
11.079
25.687
2.634
SEM-AVS
ugmol/g_
-1.361
-2.643
-2.679
-.587
-57.845
(
-1S57
-.567 '
-.588
-3.739
-9.529
-10.681
-15.419
-21.043
-1.895
-88.484
-52.859
-.454
-24.695
-25.763
-23.500
-61.494
-.678
-3.590
.420
.775
-13.473
-56.266
-51.922
-21.989
-38.802
-109.138
-128.545
-142.367
-39.651
-22.324
-5.589
-3.619
•9.027
-4.250
-1.429 '
-.539

-17.807
-8.752 .
.740
-10.066
-23.208
-2.080
SURVIVAL*
_ % „._.._ 	
, 93.
94.
92.
93.
96.

93.
95.
83. ,
96.
97.
97.
95.
99.
,98.
95.
97.
95.
96.
96.
93.
93.
87.
90.
92.
90.
88.
80.
85.
92.
90.
86.
91.
92.
86.
89.
84.
• '87.
86.
85.
83.
95.
f
82.
84.
75.
90.
83.
84.
SIGNIFICANCE' TOC
%
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
b
0
0
0
0
0
0
0
0
0
0
0 K
0
0
0
0
0
0
0

0
0
1.
0
'• • o
. 0


.20
1.20
1.30
.75
3.86

.58
.69
.26
.45
.56
.21
.27
.32
.25
.77
1.14
.15
.95
?25
.98
.90 ,
131
2.44
3.52
7.36
3.99
5.24
3.63
3.18
3.85
4.29
4.36
6.04
3.73
3.93
.67
.75
1.22
1.25
1.05
.88

1.40
.95
1.77
.76
.99
.60
                                         Draft for SAB 1-149

-------
STUDY"
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-NB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB

REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-RB
REMAP-UH
REMAP-UH
REMAP-UH
SEM
5.222
5.116
14.791
4.917
.398
4.855
3.290
5.822
9.167
6.214
.794
4.985
5.280
2.268
6.678
2.833
.333
.756
.582
1.012
1.596
.326
2.709
5.485
f
, 3.596
5.329
.337
.986
.856
5.364
1.706
.371
.193
.869
1.288
1.650
2.422
.512
4.198
5.081
6.095
8.471
3.370
1.198
2.127
1.360
1.197
AVS
utnol/
22.617
7.352
109.780
.530
.218
9.606
10.105
51.460
93.563
42.415
2.651
43.663
1.934 ,
6.300
17.559
45.222
22.315
1.216
.821
.567
.447
.156
3.120
14.666

19.503
4.321
2.901
.156
.156
39.700
23.515
4.210 .
.156
19.617
.593
.624
.156
.156
4.086
,36.490
5.957
8.078
17.247
.156
'12.446
1.790
3.373
SEM-AVS
-17.395
-2.236
-94.989
4.387
.180
-4.751
-6.815
-45.638
-84.396
-36.201
-1.857
-38.678
3.346
, -4.032
-10.881
. -42.389
-21.982
-.460
-.239
.445
1.149
,170
-.411
-9.181

-15.907
1.008
-2.564
.830
.700
-34.336
-21.809
-3.839
.037
-18.748
.695
1.026
2.266
.356
.112
-31.409
.138
.393
-13.877
1.042
-10.319
-.430
-2.176
SURVIVAL*
83.
9.
8.
89.
94.
83.
60.
41.
25.
68.
93.
53.
\ 83.
16.
77.
54.
93.
92.
**.
94.
95.
93.
70.
92.

62.
91.
97.
96.
96.
91.
93.
91.
92.
85.
92.
91.
98.
93.
90.
89.
4.
. 91.
94.
94.
83.
99.
92.
- SIGNIFICANCE1 TOC
0
1.
1.
0
0
0
1.
1.
' 1. . '
1.
.0
1.
0
1.
1.
1.
0
0
0
0
0
0
1.
0

1.
0
0
0
0
0
0
0
0
0
0
0
0
o
o
0
1.
0
o
0
0
0
0
*
1.48
1.45
9.15
3.10
2.42
2.62
5.70
2.22
6.48
3.24S
2.36
3.90
6.10
1.99
15.20
2.02
1.23
.33
.30
.30
:n
.08
•42
2.29

.88
.97
.53
.12
.51
1.17
3.21
3.54
2.52
2.39
2.44
2.68
2.60
.42
2.63
2.08
3.03
5.30
3.91
1.03
3.43
1.26
5.85
Draft for SAB 1-150

-------
STUDY"
SEM
AVS   SEM-AVS
SURVIVAL1*
SIGNIFICANCE' TOC
REMAP-UH 1.975 17.136 -15.161
REMAP-UH 2.829 25.189 -22.360
REMAP-UH 2.830 56.401 -53.571
REMAP-UH 1.385 44.588 -43.203
REMAP-UH 1.519 11.549 ' -10.030
REMAP-UH 3.186 86.235 -83.049 •
, REMAP-UH 2.086 11.713 -9.627
REMAP-UH 1.799 12.631 -10.832
REMAP-UH .930 10.093 -9.163
REMAP-UH .459 .156 .303
REMAP-UH .889 2.623 -1.734
REMAP-UH .833 2.464 -1.631
REMAP-UH 1.317 15.563 -14.246
REMAP-UH 2.480 32.123 -29.643
REMAP-UH .626 9.949 -9.323
REMAP-UH 1.500 5.427 -3.927
REMAP-UH , .723 1.341 -.618
^ j
REMAP-UH 4.158 13.504 -9.346
REMAP-UH 2.241 27.788 -25.547
REMAP-UH 2.907 29.285 -26.378
REMAP-UH .852 1.591 -.739
REMAP-UH 2.294 53.955 -51.661
REMAP-UH 2.995 33.995 -31.000
REMAP-UH 2.981 44.910 -41.929
REMAP-UH .677 10.323 -9.646
a) Sources: EMAP-VA is U.S. EPA, 1996 •
NOAA-LI is Wolfe ct al., 1994
NOAA-BO is Long et al., 1996
NOAA-HR is Long et al., 1995
REMAP is Adams et al., 1996
b) Conclusion of signifigance varies' for three databases.
EMAP significance based on percent survival of control
NOAA significance based on percent survival kss than 80%
.REMAP significance based on percent survival less than 80%
c) Significance: 0 - No significant toxicity
1 - Significant toxicity
45.
84.
96.
88.
82.
93.
82.
37.
89.
98.
95.
86.
88.
87.
97.
89.
89.

96.
70.
95.
93.
15.
88.
94.
91. '

'







••

1.
0
0
0
0
0
0
1.
0
0
0
0
0
0
0
0
0

0
1.
0
0
1.
0
0
0











2.33
.91
1.21
1.03
1.06
1.39
.79
1.06
.43
.13
.21
4.96
2.56
3.06
2.58
2.71
3.89

, 4.78
2.66
5.15
2.03
4.37
3,55
2.97
3.32







,



                                Draft for SAB 1-151

-------

-------
i
         Sediment Assessment Presentation Materials

-------
i

-------
i
The Technical Basis of the Use of Acid Volatile Sulfide (AYS) and Interstitial
                 Water Normalizations in Sediment Guidelines
 .                                   '            "                 \

                                        by
                                              Walter J. Berry
                             U.S.EPA, Atlantic Ecology Division, Narragansett RI

                                           Dominic M. Di Toro
                                        HydroQual, Inc., Mahwah NJ

                                             David J. Hansen
                             Great Lakes Environmental Center, Traverse City MI
                                               March 1999
           INTRODUCTION                                .     ,

                 The U.S. EPA Science Advisory Board (SAB) has endorsed the use of the equilibrium
           partitioning  (EqP) approach as the foundation for the Agency's sediment guidelines for both
           nonionicorganics and divalent metals (U.S.EPAj 1990,1992,1995). The EqP approach was chosen
           because it was the only available approach that could provide numerical guidelines, applicable across
           sediments, which had a firm theoretical basis (Figure 1).  The usefulness of the sediment guideline
           for metals (then referred to as the Sediment Quality Criterion) had several limitations, however. One
           limitation was that it only applied to five metals: cadmium, copper, lead, nickel, and zinc. Another
           limitation was that it was a "one-way criterion", which is to say that it could predict that some
           sediments would not be toxic, but could not predict that other sediments would be toxic.
                 This SAB review is intended to examine some modifications to the metals  guidelines
           previously reviewed by the SAB which extend the usefulness of the EqP approach to assessing the
           potential effects of metals in sediments. These modifications extend the use of acid volatile sulfide
           (AYS) and interstitial water normalization to include chromium and silver, and may eliminate the
           "one-way" limitation by incorporating organic carbon into the assessment of sediments contaminated
           with metals.

                 The  technical basis of the use of acid volatile sulfide (AVS) and interstitial  water
           normalizations in sediment guidelines has been extensively reviewed elsewhere. Ankley et al:
           (1994) described the AVS and interstitial water approach and Ankley et al. (1996) reviewed all of
           the available data.  Ankley et al. (1996) was the first of a large collection of papers relating to the
           use of AVS and interstitial water normalization in sediment assessment which appeared in the sanie

                                                    2-2       .

-------
issue of Environmental Toxicology and Chemistry (December, 1996, Vpl 15, No. 12). In this
introduction we will first review the published literature on EqP as it relates to metals. Then we will
summarize the available data from acute laboratory sediment toxicity tests with spiked sediments
and with field sediments which can be used to test the utility of AVS and IWTU normalizations in
predicting sediment toxicity.  Then the available chronic data will be summarized.  The  data
pertaining to the bioaccumulation of metals in relation to AVS normalization will be reviewed.
Some considerations of the use of AVS normalization of field sediments will be discussed. Finally,
the guidelines statements will be presented.

LITERATURE REVIEW

       Di Toro et al. (1990) showed that the toxicity of cadmium-spiked marine sediments was
linked to metals/AVS ratios and interstitial water (TW) metals concentrations. Since then several
studies using fresh and salt water sediments spiked with cadmium, copper, lead, nickel and zinc
(Berry et al., 1996; Casas and Crecelius, 1994; Green et al., 1993; Di Toro et ah,  1992; Carlson et
al., 1991) have demonstrated the utility of these parameters in causally linking toxicity to metals in
sediments. Kemp and Swartz (1998) maintained constant TW concentrations in cadmium-spiked
sediments modified by varying quantities of organic carbon and found that mortality was correlated
with IW concentration but not total  sediment concentration.  The utility of AVS and IW
normalizations has also been demonstrated in studies conducted at field sites contaminated with
copper (Ankley et al., 1993) and a mixture of cadmium and nickel (Pesch et al., 1995; Di Toro et al.,
1992; Ankley et al., 1991).  Two colonization experiments with cadmium-spiked sediments, one
conducted in a freshwater lake (Hare et al., 1994)  and a seawater colonization test conducted in the
laboratory (Hansen et al., 1996a) also support the use of this approach for predicting the toxicity of
these metals in sediments. The success of this approach for predicting the bioavailability of these
metals in  sediments is in direct contrast to the lack of success  in using dry weight metals
concentrations for this purpose (Di Toro et al., 1990; Di Toro et al., 1992; Luoma, 1989).

TECHNICAL BASIS

       The theoretical foundation for EqP-based AVS predictions of metal toxicity is that the
sulfides of cadmium, copper, nickel, lead, and zinc all have lower sulfide solubility product constants
than do the sulfides of iron and manganese, which are formed naturally in sediments as a product
of the bacterial oxidation of organic matter (Goldhaber and Kaplan, 1974). As a result, these metals
will displace manganese and iron whenever they are present together with manganese and iron
monosulfides (Di Toro et al., 1992). Because the solubility product constants of these sulfides are
small, sediments with an excess of AVS will have very low metal activity in the IW and no toxicity
due to these metals should be observed in the sediments.

       The results of the studies cited above were consistent with the following predictions based
on EqP theory: 1) when sediments have an excess of AVS over metals, sediments will not be toxic,
and little or no metal will be present in the IW; and 2) when sediments have an excess of metals
over AVS, AVS binding potential will be exceeded and metals will be present in the IW or available
to bind with other sediment phases (i.e. total organic carbon) (Di Torq et .al., 1990).  Nontoxic
sediments with a small excess of metals over AVS may have low IW concentrations, less than those
                                          2-3

-------
known to be toxic in water-only tests, suggesting the importance of additional metal binding phases
in sediments (Green etal., 1993; Ankley et al, 1993; Gonzales et al., 1992).  The appropriate
fraction of metals to use for AVS normalization is referred to as^simultaneously extracted metal
(SEM).  This is the metal which is extracted in the cold acid used in the AVS procedure.  This
fraction is appropriate because some metals form sulfides which are not fully labile in the short time
required for the AVS extraction (e.g. nickel and zinc) (Di Toro et al., 1992).  If a more rigorous
extraction were used to increase the fraction of metal extracted which did not also capture the
additional sulfide extracted; then the sulfide associated with the additional metal release would not
be quantified. This would result in an erroneously high metal to AVS relationship (Di Toro et al.,
1992).

       An analysis of a simple chemical equilibrium model for the system M(H), Fe (IT), S(-H)
where M(II) are the divalent metals that form sulfides shows that

          {M}/[MJT > KMs/KFeS                           '                       ,

Where {M} is the activity of M in the IW, [M]T is the total metal concentration, KMS is the solubility
product for the metal sulfide, MS, and KFeS is the solubility product of FeS. The ratio KMS/K^ is
10'56,10'60, lO'103,10'u°, 10'186 for M = Ni, Zn, Cd, Pb, and Cu, respectively (Berry et al., 1996).
If the metals are present in excess of the sulfides (SEM-AVS >0.0), and there are no other sediment
phases capable of binding the metals (e.g., dissolved Organic carbon (DOC) or total organic carbon
(TOC)), then metal will be present in the IW and the sediment may be toxic.

       Toxicity predictions based on sulfide binding for sediments contaminated with mixtures of
metals which form insoluble sulfides would use the sum of the molar concentration of SEM for the
divalent metals present (i.e., Cd, Cu, Ni, Pb, Zn) for comparison with the molar concentration of
AVS in the sediment.  If the sum of the SEM is greater than that of AVS, metals may occur in the
IW in sufficient concentrations to be toxic.  If the toxicity of the cationic metals in IW is assumed
to be additive (Spehar and Fiandt, 1986), it should be possible to predict the toxicity of the sediments
in the same way as in the individual metal experiments, using the sum of the interstitial water toxic
units (IWTU). The divalent metals should appear  in the IW in reverse order of the solubilities of
their sulfides (Di Toro et al., 1990). Thus, nickel should appear first in the IW in sediments with
SEM-AVS slightly greater  than  0.0, followed by  zinc, cadmium,  lead, and copper as the
concentration of metals increases relative to that of AVS.   This predicted outcome was observed
by Berry etal. (1996).

ACUTE TOXICITY DATA

       Di Toro et al. (1990) found that toxicity of cadmium in sediments increased with increasing
concentration when the cadmium concentration was expressed on a dry weight basis, but that the
response was sediment specific (Figure 2). This meant that it was not possible to predict the toxicity
of cadmium across  sediments using dry  weight  normalization.   This sediment  specificity
undoubtedly  contributes to the scatter seen when  mortality is plotted against dry  weight
concentration in a data set which includes a number of spiked metals tests and field sites where
metals were thought to be the cause of observed toxicity (Figure 3 from Hansen et al., 1996b as is
presented in the Metals Mixtures ESG document). Note that in Figure 3 the sediment concentrations

                 .   •                     2-4

-------
that caused little or no mortality overlap with those concentrations which caused 100% mortality for
three orders of magnitude, and there is no theoretical basis for predicting where toxicity might occur.
                         **                         i
       When mortality was plotted against the concentration of cadmium in the interstitial water,
however, Di Toro al. (1990) found that the results were similar across sediments (Figure 4).  The
larger data set of lab and field data seen in Figure 5 further validate the use of IWTUs. In this data
set little or no toxicity was seen in sediments with less than 0.5 IWTU, while mortality increased
with increasing IWTU in sediments with greater than 0.5 IWTU.

       Similarly, Di Toro  et al. (1990) found that mortality was not sediment specific when
cadmium was expressed on an SEM/AVS basis (Figure 6). (SEM/AVS ratios were used in earlier
papers but have been supplanted by SEM- AVS differences, especially in field sediments; see Hansen
et al., 1996b and the Metals Mixtures ESG document). In the larger data base when there was an
excess  of sulfide (SEM-AVS < 0.0) the sediments were not toxic, but when the metals were in
excess, many of the sediments were toxic (Figure 7). Note that many of the sediments with large
excesses of metals were not toxic. This is the reason that the SEM-AVS guideline is considered a
"one-way" guideline, which is more useful in the prediction of lack of toxicity than it is in the
prediction of toxicity. This is still very useful, because the vast majority of field sediments have an
excess  of AVS. Furthermore,  it should be pointed out that the ability of AVS normalization to
predict toxicity, although not as good as its ability to predict lack of toxicity, is comparable to the
ability of the empirically-derived methods to predict toxicity.

CHRONIC  DATA                                                ,

       The data from the available life-cycle and colonization exposures also support the use of
AVS and interstitial water normalizations in the prediction of the lack of biological effects of metals
in sediments. Data are available for a number of metals from both freshwater and saltwater sediment
tests. These data are summarized in Table 1. Note that all sediments in which effects were observed
had an excess of metal over AVS.

BIOACCUMULATION
        I                              '          .
       In a perfect world organisms would not take up metals from sediments with SEM - AVS <
0.0. Unfortunately, data exist which show that bioaccumulation of metals from sediments is often
more highly correlated with  dry weight concentration than it is with an AVS-normalized
concentration. One such data set is shown in Figure 7. In some of the early experiments it was
thought that the apparent accumulation of metal in sediments where SEM-AVS < 0.0 may have been
an artifact of the use of very high and increasing concentrations of metals in spiked sediments.
However, a recent experiment, in which the metal concentration was held constant and the AVS was
varied, seems to show conclusively that this is not the case. In this experiment the accumulation of
metal was more closely related to dry weight concentration than AVS-normalized concentration
regardless of whether the metal concentration or AVS was varied (Lee et al., 1998). It would appear
then, that a guideline using AVS normalization would not be protective ofbioaccumulation of metals
from sediments.  It must be noted, however, that although metal uptake has been shown  from
sediments with an excess of AVS, no biological effects of these sediments have been shown. If there
                                          2-5

-------
1
 was an effect of metals uptake from these sediments on the lower levels of the food chain, then it
 might have been expected to show up in the colonization and life-cycle tests described above.
 Furthermore, there is little evidence of bipconcentration of these metals up the food chain.
 Therefore, the fact that metals can be taken up from sediments with an excess of AVS means that
 the guideline must be qualified, but this finding does not negate the usefulness of the guideline.

 USE OF AVS WITH FIELD SEDIMENTS
                           N         '                   •        • . '         '
       There are some considerations relative to the use of AVS in predicting biological effects in
 field sediments which should be noted here.  1) AVS can break down in aerobic storage, but it is
 relatively stable, especially if frozen in glass. 2) AVS varies with season, being lowest in the winter
 and early spring, and highest in the late summer. Since most sampling is done in the warmer months
 there may be a concern about sampling when AVS is at apeak. If sediments which have a slight
 excess of AVS over metals are found or expected sampling can be done in the winter or early spring.
 3) AVS concentration varies with depth, being higher at depth and  lowest at the surface.  Thus,
 attention must be paid to the depth of sampling. Sampling should be at the depth of concern (depth
 of dredging, for example). The top 2.0 cm is recommended for evaluation of "in-place" sediments.
 4) AVS and interstitial water normalization can provide insight into the causes of sediment toxicity
 in field sediments with mixtures of contaminants, but cannot by itself predict the toxicity of those
 sediments because other toxicants may make these sediments toxic even if the sediments are not
 toxic due to metals (Figure 9).

 STATEMENT OF GUIDELINE

       Based on the considerable evidence summarized above the following ESG for metals were
. proposed in the 1995 draft document:

 Solid phase guideline:

 S[SEM]i - AVS <0.0                 ,

 (Where S[SEM]f = molar sum of cadmium, copper, lead, nickel, and zinc)

 Interstitial water guideline:
           (Where S([IW]j-= interstitial water concentration of cadmium, copper, lead, nickel, and zinc; and
           [FCV]i = Final Chronic Value, from water quality criteria)

           If either of these guidelines are met, sediments should not be toxic due to metals (cadmium, copper,
           lead, nickel and zinc).                                     -•- '
                                                    2-6

-------
                                     References

Ankley.G.T. 1996. Evaluationofmetal/acid-vplatilesulfiderelationshipsinthepredictiohofmetal
bioaccumulation by benthic macroinvertebrates. Environ: Toxicol. Chem.  15:2138-2146.

Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and J.L. Bennett. 1993. Predicting the acute
toxicity of copper in freshwater sediments: evaluation of the role of acid-volatile sulfide. Environ.
Toxicol. Chem. 12:315-320.

Ankley, G.T., G.L. Phipps, E.N. Leonard, D.A. Benoit, V.R. Mattson, P.A. Kosian, A.M. Cotter,
J.R. Dierkes, D.J. Hansen and J.D. Mahony.  1991. Acid-volatile sulfide as a factor mediating
cadmium and nickel bioavailability in contaminated sediments. Environ. Toxicol. Chem, 10:1299-
1307.

Ankley, G.T., N.A. Thomas, D.M. Di Toro, D J. Hansen, J.D. Mahony, WJ. Berry, R.C. Swartz and
R.A. Hoke. 1994. Assessing potential bioavailability of metals in sediments: A proposed approach
Environ. Mgt.  18:331-337.

Berry W.J., D.J. Hansen, J.D. Mahony, D.L. Robson,  B.P. Shipley, B. Rogers and J.M. Corbin.
1996. Predicting the toxicity of metals-contaminated  sediments using acid volatile sulfide and
interstitial water normalization in laboratory-spiked sediments. Environ. Toxicol. Chem. 15:2067-
2079!                                                           .

Carlson, A.R., G.L. Phipps, V.R. Mattson, P. A. Kosian and A.M. Cotter. 1991. The role of acid-
volatile sulfide in  determining cadmium bioavailability and toxicity in freshwater sediments.
Environ. Toxicol Chem.  10:1309-1319.

Casas, A.M. and E.A. Crecelius.  1994. Relationship between acid volatile sulfide and the toxicity
of zinc, lead and copper in marine sediments. Environ' Toxicol.  Chem. 13:529-536.

Di Toro., D.M., J.D. Mahony, DJ. Hansen, KJ. Scott, A.R. Carlson and G.T. Ankley. 1992. Acid
volatile sulfide predicts the acute toxicity of cadmium and nickel in sediments. Environ. Sci.
Technol 26:96-101.                        ,

Di Toro, D.M., J.D. Mahony, DJ. Hansen, KJ. Scott, M:B. Hicks, S.M. Mays and M.S. Redmond.
1990^ Toxicity of cadmium in sediments: The role of acid volatile sulfide. Environ. Toxicol. Chem.
9:1489-1504.                            '

Goldhaber, M. B. and I.R. Kaplan, 1974. The sulfur cycle. In E.D. Goldberg, ed., The Sea, Vol. 5-
Marine Chemistry. John Wiley and Sons, New York, NY, USA, pp. 569-655.

Gonzales, A.M., J.D. Mahony and D.M. Di Toro.  1992 The role of organic carbon in the toxicity
of anoxic sediments contaminated witti copper and other metals: An experimental study. Abstracts.
13th Annual Meeting, Soc: Environ. Toxicol. Chem., Cincinnati, OH, USA,  Nov. 8-12,1992, p.
162.                                                              '
                                          2-7

-------
i
 Green, A.S:, G.T. Chandler, and E.R. Blood.  1993. Aqueous-, pore-water-, and sediment-phase
/ cadmium: toxicity relationships for a meiobenthiccopepod. Environ. Toxicol. Chem. 12:1497-1506.

 Hansen, D.J., J.D. Mahony, W. J. Berry, S.  Benyi, J. Corbin, S. Pratt and M.B. Able.  1996a.
 Chronic effect of cadmium in sediments on colonization by benthic marine organisms: An evaluation
 of the role of interstitial cadmium and acid volatile sulfide in biological availability.  Environ,
 Toxicol. Chem 15:2136-2137.

 Hansen, D.J., W.J. Berry, J.D. Mahony, W.S. Boothman, D.M. Di Toro, D.L. Robson, G.T. Ankley,
 D. Ma, Q. Yan and C.E. Pesch.  1996b.  Predicting the toxicity of metals-contaminated field
 sediments using interstitial concentration of metals and acid volatile sulfide normalizations. Environ.
 Toxicol. Chem 15:2080-2094.

 Hansen, D.J., J.D. Mahony,1 W. J. Berry, S.  Benyi, J. Corbin, S. Pratt and M.B. Able.  1996a.
 Chronic effect of cadmium in sediments on colonization by benthic marine organisms: An evaluation
 of the role of interstitial cadmium and acid volatile sulfide in biological availability.  Environ.
 Toxicol. Chem 15:2136-2137.

 Hare, L., R. Carignan and M.A. Huerta-Diaz. 1994. A field experimental study of metal toxicity and
 accumulation by benthic invertebrates; implications for the acid volatile sulfide (AVS) model.
 Limnol. Oceanogr. 39:1653-1668:

 Kemp, P.P., and R.C. Swartz. 1988. Acute toxicity of interstitial and particle-bound cadmium to
 a marine infaunal amphipod. Mar. Environ. Res. 26:135-153.

 Lee, B.-G., H.-S. Jeon, S.N. Luoma, J.-S. Yi, C.-H. Koh.  1998. Effects of AVS (Acid Volatile
 Sulfide) on the bioaccumulation of Cd, Ni, and Zn in bivalves and polychaetes. Abstract: 19th
 Annual Meeting of the Society of Environmental Toxicology and Chemistry. Charlotte, N.C.

 Luoma, S.N. 1989. Can we determine the biological availability of sediment-bound trace elements?
 Hydrobiohgia. 176/177:379-396.

 Pesch, C.E., D.J. Hansen, W.S. Boothman, W.J. Berry and J.D. Mahony. 1995. The role of acid-
 volatile sulfide and interstitial water metal concentrations in determining bioavailability of cadmium
 and nickel from  contaminated sediments  to the marine pplychaete, Neanthes arenaceodentata.
 Environ.  Toxicol. Chem.  14:129-141.
                                                                                  \

. Spehar, R.L. and J.T. Fiandt. 1986. Acute and chronic effects of water quality criteria-based metal
 mixtures on three aquatic species. Environ. Toxicol. Chem. 5:917-931.

 U.S. Environmental Protection Agency. 1990. An SAB evaluation of the equilibrium partitioning
 (EqP) approach  for assessing sediment quality.  ' EPA-SAB-EPEC-90-006.  Office of Water,
 Washington, DC.

 U.S. Environmental Protection Agency.  1992.  An SAB report. Review of sediment  criteria

                                          2-8                                • .

-------
development methodology for non-ionic organic contaminants. EPA-SAB-EPEC-93-002. Office
of Water, Washington, DC.

U.S. Environmental Protection Agency. 1995. An SAB report: Review of the Agencies approach
to developing sediment quality criteria for five metals. EPA-SAB-EPEC-95-020. Office of Water,
Washington, DC.
                                       2-9

-------
i
                                  Figure Legends

Figure 1: Conceptual models of chemical exposure.

Figure 2: Amphipod mortality vs. dry weight cadmium concentration in three cadmium-spiked
sediments. From Di Toro etal. (1990).

Figure 3: Mortality vs. Total metal or SEM (dry weight) in field and lab sediments in which
metals were the probable cause of toxicity. From Metals Mixtures ESG document.

Figure 4: Amphipod mortality vs. interstitial water cadmium concentration in three cadmium-
spiked sediments. From Di Toro et al. (1990).

Figure 5: Mortality vs. interstitial water toxic units in field and lab sediments in which metals
were the probable cause of toxicity. From Metals Mixtures ESG document.

Figure 6: Amphipod mortality vs. SEM/AVS in three cadmium-spiked sediments. From Di Toro
etal.(1990).                    .                          ,

Figure 7: Mortality vs. SEM - AYS in field and lab sediments in which metals were the probable
      cause of toxicity.  From Metals Mixtures ESG document.

Figure 8: Cadmium tissue concentration vs. SEM/AVS in polychaetes exposed to cadmium-
      spiked sediments. FromPeschetal. (1995).

Figure 9: Mortality vs. SEM-AVS in field sediments contaminated with a mixture of toxicants.
                                                  2-10

-------
                                                            _fll.
CD
 t&
 O
        U>
        t»»
              S
     P-B  s'B-
     •   sr  vo-fi
   A
            «£
            &s
            2" P
            S *• s?1

            ^^
                                   n

                                                       I
                               ?
                               x
        a
s 0
"•I
-   &
                 vo
                 vo
                 -J
                               00
                                              K)
                                              oo
   U)

   4x
             j:
             5
                                            00 -
                                            **  I
                                               i
                                              I
         ^^
                                                         Q
EL
                              .
                          vo J5
                                                      f
                           I

                           1
                                                                  o
                                                               0)
                             2-11

-------
   O
   o
CO
                                                  O
                                                  O
o
CD
                                                  Q>
                                                 CD

                                                 C/)

                                                 O
                                                 O
                                                  I
                                                  o
                                                  m
                                                  X
                                                  o
                                                  (/)
                                                  C

                                                  CD
                           2-12

-------
tt
o
                MORTALITY  vs  SEDIMENT CADMIUM
                       DRY WEIGHT NORMALIZATION
             (A) INITIAL EXPERIMENTS
                • LI SOUND
     80


     60


     40


     20
(B) JOINT EXPERIMENT
   • LI SOUND
   • MIXTURE
   O NINIGRET POND

                                     uL
        10
          100
1000
                                                10000
                                                  1000PO
                SEDIMENT CADMIUM (ug Cd/gm dry wt)

       Figure 1. (A) f oxicity test results for Long Island Sound sediments (AmpeVsca). (B) Toxicity test
       results for Ninigret Pond (Rhepcaynius hudsonf), Long Island Sound and the 50/50% mixture of
       the two sediments (Ampelisca). Cadmium concentrations on a sediment dry weight basis.
                                2-13

-------
                     Mortality (%)
 S
 SL


 (D
 F+
 SL

 o

 CO
 m
(D"
                                              z  C
                       2-14

-------
                                     •as
                                            i
                MORTALITY (%)
»l
  3
  N
  a
iti
8-
r-|
« ~
¥ i r
         §
         o
         o
         0
      o  2
         P
         o
         o
           o
           o
       S

       >
       O
     3   8

     *   §


     to

     o   J*
     Q.   O


     *.fe
         O
         O
         o
         o
         o
         P
         o
         o
         o
         o
         o
                                     O
                                   CO
                                   m
                                   a
                                   CO
                M 01 N
           -   T- ,> «X »N X



           .   I I If I
                                     I
                                     m
                                     33
                                     o
                                     >
                                     o

                                     C
                  2-15

-------
•m
                                    Mortality (%)
                 i
                 i
                 o
                                      2-16

-------
                 MORTALITY (%}
§ri
1*1
U
?i
2. <
5'

S
o
2
o
IS
r §.
i3  -a
  E
 * £
 frff
 §:
          o
          b
      m
      o
      m  p

      -•  s
      o
           o
           o
o
a

\   ^
c   o
3   b
o   o


I
         o
         p
         o
         o
                                     >

                                     2   o
                                    o  >

                                   •
                                       CO
0)
c

-II

5
m


o
3)
                                     N
                                     >
                                       m
                                       o
                                       m


                                       H

                                       O
                                       >
                                       D
                   2-17

-------
                    Mortality (%)
V)
m
                                 '6<3P  6
                      2-18

-------
                    jxmol Cd/g Tissue
       §
     P
     b
            i iniif ^  i i uui|

              A


              f.
 (0

 a

 3

 §
m

I
    o
    o
                 "H"  ~'~

                 .ce0:

                   00

                 a a

                    o
               a
               CO  CO
         o
         o-


         I
         S
   o
   o
   o
               o
3
               CO
                                               O


                                               D
CO


*
m
o

o>
rn
o
                                               m

                                               I
                        2-19

-------
C0
rn
o
o1
            % Mortality
        ro
        o
o>
•o
00
o
o
o
        T
  O
  o
T
T
             21! CO

             O
             t


             o
             X
                       W
                       m
            a •
            o
                  a M

                  2.1 i
                    CD Q)
                  ft)


                  !|l

                  !»i
                  CO C 5
                                   •   DT
                                 B>

                      (D
                  0) ±. ?
                    3 3"

                    e? o
                      o
                      3

                      5"
                  2-20

-------
2-21

-------
                    Predicting the Toxicity of Metals in Sediments
                                          by .         :

       Dominic M. Di Toro tt, David J. Hansentff, Joy A., McGrathV and Walter J. Berry*

                              tHydroQual, Inc. Mahwah, NJ
        ^Environmental Engineering and Science Dept. Manhattan College, Bronx NY
  §US EPA National Health and Environmental Effects Research Laboratory, Narragansett, RI.
                    *Great Lake Environmental Center, Traverse City MI

                                     March 1999

Introduction

The SEM-AVS method for evaluating the toxicity of metals [1,2] has proven to be quite successful
at predicting the lack of toxicity in spiked and field contaminated sediments [3,4]. However, it does
not appear to be able to predict very well the onset of toxicity in a sequence of sediments spiked with
increasing concentrations of metals, or for sets of field contaminated sediments.  In fact, in a recent
article [5] it was claimed mat the empirically derived methods  [6] are better at making  this
prediction. The purpose of the note is to introduce a modification of the SEM - AYS procedure that
significantly improves the prediction of mortality. Additionally, the inability of the empirically
derived methods to make this prediction is demonstrated.
Theory
The Equilibrium Partitioning (EqP) model gives the prescription for the development of sediment
concentrations that predict the toxicity or lack of toxicity in sediments. The sediment concentration
C'se,, that corresponds to a measured LC50 in a water only exposure of the test organism is
                                         2-22

-------
Stt
                                                                            (1)
  where C*^ is the sediment concentration ( g/g dry wt), £ (Ukg) is the partition coefficient between
  pore water and sediment solids, and C'w is the LC50 concentration ( g/L)[7J.  For application to
  metals that react with AVS to form insoluble metal sulfides, this formula becomes [1]
                             Csw= AVS + KpCy            .   •           (2).
  where AVS is the sediment concentration of acid volatile sulfides. The formula simply states that
  since AVS dan bind the metal as essentially insoluble sulfides, the concentration of metal in a
  sediment that will cause toxicity is at least as great as the AVS that is present. The sediment metal
  concentration that should be used is the SEM concentration since any metal that is bound so sHongly
  that IN hydrochloric acid cannot dissolve it is not likely to be bioavailable [2], Of course, this
  argument is just speculation, which is why so much effort has been expended to demonstrate
 experimentally that this is actually the case [2-4, 8]. Therefore eq.(2) becomes

                            SEM= AVS + KPCW                                 (3)
 The basis for the SEM/AVS method is to observe that if the seeond.term in eq.(3) is neglected then
 the critical concentration is                                                         .
                            SEM= AVS
                                         /
 or the criteria for toxicity or lack of toxicity is
                                                  (4)
or, equivalently
                           SEM/AVS = 1
                           SEM - AVS =
                                                  (5)

                                                  (6)
-The Mure of the either the ratio condition (eq.5) or the difference condition (eq.6) to predict toxicity
is due to the neglect of the partitioning term K, C*w. Note that ignoring the term does not affect the
prediction of lack of toxicity since it makes the condition conservative (i.e. smaller concentrations
of SEM are at the boundary of toxicity and no toxicity).
                                          2-23

-------
                                                                        V
The key to improving the prediction of toxicity is to approximate the partitioning term rather than
ignoring it. In anaerobic sediments, the organic carbon fraction is an important partitioning phase
and partition coefficients for certain metals at certain pHs have been measured [9].  This suggests
that the partition coefficient K? in eq.(3) can be expressed using the organic carbon based partition
coefficient, KQC, together with the fraction organic carbon in the sediment, f^.
Using this expression in eq.(3) yields

                           SEM=  AVS + KocfocCy                             (8)
                                                                    •
Then moving the known terms to the left-hand side of the equation yields      .

                           (SEM-AVS)/foc^KocCw    ,                      (9)

If we knew both KOC and C'w we could use eq.(9) to predict toxicity. The method evaluated below
uses (SEM - AVSy f^ as the predictor of toxicity and evaluates the critical concentrations (the
right hand side of eq.9) based on observed SEM, AVS, ^ and toxicity data.
Data Sources             '•'.-.

Data from toxicity tests using both laboratory-spiked and field-collected sediments were compiled
from the literature. Three sources of laboratory spiked tests using marine sediments were included,
Casas  and Crecelius  [10] and   Berry et al. [3, 11].   Data  reported included  total metals,
simultaneously extracted metals (SEM), acid volatile sulfide (AVS), organic carbon fraction (f^ and
10 day mortality. In the study by Casas and Crecelius [10], the toxicity of zinc, lead and copper
were tested on the marine polychaete Capitetta capitata. In Berry et al. [3] the toxicity of cadmium,
                                          2-24

-------
copper, lead, nickel, zinc and a mixture of four metals (cadmium, copper, nickel and zinc) were
tested on the marine amphipod Ampelisca abdita. In Berry et al. [11] the toxicity of silver to A.
abdita was tested. .Three sources for metal contaminated field sediments were included, Hansen et
al. [4], Kemble et al. [12], and Call et al.  [13].  Data reported included total metals, SBM, AVS,
fraction organic carbon and mortality.. The metals included in total SEM were cadmium, copper,
nickel, lead and zinc.  In Hansen et al. [4], data were reported for five saltwater and four freshwater
locations.  For the freshwater stations, organic  carbon and total metals data were not provided.
Organic carbon data for one location, Keweenaw Watershed, were obtained separately [Berry,
personal communication]. Ten day bioassays were conducted using A.  abdita and a freshwater
amphipod, Hyaletta azteca. In Kemble et al. [12] 14-day bioassays were conducted on Chironomus
ripdrius, a freshwater midge. Data included from Call et al. [13] were the control freshwater
sediment data with 10-day mortality to the midge, Chironomus tentans.

Methods

Laboratory spiked and field-contaminated sediment data were grouped together and analyzed as one
                                                                              .1
data set. Mortality data were compared  against the SEM-AVS difference and this difference
normalized to the fraction organic carbon, (SEM-AVS)/^. For each comparison two bounds were
computed for the SEM - AVS comparison and the (SEM - AVS)/ f^: a lower bound concentration
equivalent to a 95 percent chance that the mortality observed would be less than 50 percent and an
upper bound concentration equivalent to a 95 percent chance that the observed mortality would be
greater man 50 percent. The lower bound limit was computed by evaluating the fraction of correct
                   /      i                               •
classification starting from the lowest abscissa value. When the fraction correct dropped to below
95%, the 95* percentile was interpolated.  The same procedure was applied to obtain the upper
bound. Upper and lower bounds were calculated for both SEM-AVS difference and the organic
carbon normalized difference.

Using the  same data set, individual total metal  concentrations were divided by the effects range
median (ERM) values of Long et al. [6] for that metal or metals and the mean quotient value was
                                         2-25

-------
 determined. This was compared to observed mortality. Upper and lower bounds were calculated
 using the same method as described above.
 Results
 Mortality in the laboratory spiked and field-contaminated sediment tests were organism independent
 when plotted against the SEM-AVS difference (Figure  1, top panel).  The lower and upper
 boundaries determined for the SEM-AVS difference were  1.7 and 190 mol/g respectively and are
 shown on the top panel of Figure 1. A line indicating SO percent mortality is also shown. These
 boundaries can be used as predictive limits. For SEM-AVS values lower than 1.7 mol/g, there is a
 95 percent chance that the observed mortality will be less than 50 percent Similarly, for SEM-AVS
 values greater than 190 mol/g, there is a 95 percent chance that the observed mortality will be greater
                                                               v
 than 50 percent.  The uncertainty falls in the range of SEM-AVS values between the upper and lower
 bounds, which in this case are approximately two orders of magnitude wide. When normalized to
 the organic carbon content of the sediment, the range of uncertainty is narrowed to approximately
 one order of magnitude (Figure 1, bottom panel). The lower and upper boundaries are 86 and 2000
'  mol/goe respectively.                 .

 This same analysis is shown on a metal specific basis for cadmium, copper, nickel, lead and zinc in
. Figures 2 through 6 together with the upper and lower boundaries determined above. It appears that
 for  each metal, the boundaries could be adjusted slightly to encompass a narrower range in
 uncertainty.  However, these metals all appear to be behaving similarly in this analysis. In fact, this
 analysis was tested using the four metal mixture experiment (cadmium, copper, lead and nickel) and
 the area of uncertainty fell within the boundaries (Figure 7) suggesting that these metals behave
 similarly and it is appropriate to analyze them together.                                    ,
                                                s  '                            .      '
  A metal which appears to be more toxic than the other five metals is silver (Figure 8).  Mortality was
  observed at much lower SEM-AVS values. Upper and lower boundaries could not be computed for
  silver due to lack of data at low SEM-AVS values that had an absence of mortality.
                                           2-26

-------
Long et al. [5] have suggested that sediment quality guidelines (SQGs) based on dry weight
normalizations, were equally if not more accurate, in predicting the non-toxicity or toxicity of
sediment associated metals than AVS-norraalized SEM concentrations. It was concluded that using
the mean ERM quotient (metal concentration/ERM) was more effective at predicting whether or not
a sample was toxic compared to SEM/AVS ratios. The data set from Hansen et al. [4] was used.
                                                                   «             /  t

We reexamine this question using the complete data set described above, which includes spiked as
well as field contaminated sediments. Upper and lower bounds corresponding to 95% correct
predictions for the mean ERM quotient were determined as described above. The results are shown
in Figure 9. The lower and upper bounds were 0.24 and 270 respectively. The range of uncertainty
                                                                      \
is three orders of magnitude.  By contrast, the range of uncertainty for the SEM-AVS difference was
two orders of magnitude, indicating that the AYS normalized  SEM concentrations were more
accurate at predicting the absence or presence of metal-associated toxicity. However, normalizing
these concentrations to the sediment organic carbon content lowered the range of uncertainty to close
to one order of magnitude, therefore providing the most reliable method of predicting toxicity or
non-toxicity of sediment associated metals.

Conclusions
                                                                                i
The use of the organic  carbon normalized SEM-AVS as a predictor of toxicity reduces the
uncertainty of the prediction from three orders of magnitude using average ERM ratios, and two
orders of magnitude using SEM-AVS, to one order of magnitude using (SEM-AVS)/ foe-  There
appears to be no basis for the claim that average ERM ratios are preferable.

References
1.     Di Toro, D.M., et al., Toxicity of Cadmium in Sediments: The Role of.Acid Volatile Sulfide.
Environ..Tox. Chem., 1990. 9: p. 1487-1502.
                                         2-27

-------
i*.
           2.     Di Toro, D.M., et al, Acid volatile sulfide predicts the acute toxicity of cadmium and nickel
           in sediments. Environ. Sci. Tech., 1991.26(1): p. 96401.

           3.     Berry, W.J., et al, Predicting the toxicity of metals-spiked laboratory sediments using acid
           volatile sulfide and interstitial water normalizations. Environ. Tox. Chem., 1996.15(12): p. 2067-
           2079.             .

           4.     Hansen, D.J., etal, Predicting the toxicity of metals-contaminated field sediments using
           interstitial concentrations of metal and acid volatile sulfide normalizations. Environ. Tox. Chem.,
           1996.15(12): p. 2080-2094.                                             .

           5.     Long, E.R., LJ. Field, and D.D. MacDonald, Predicting toxicity in marine sediments with
           numerical sediment quality guidelines. Environ. Toxicol. Chem., 1998.17(4): p. 714-727.

           6.     Long, E.R., et al,  Incidence of adverse biological effects .within ranges of chemical
           concentrations in marine and estuarine sediments. Environ. Management, 1995.19(1): p. 81-97.

           7.    . Di Toro, D.M., et al, Technical basis for the equilibrium partitioning method for
           establishing sediment quality criteria. 1991.12.

           8.     Di Toro,  D.M.,  et  al,  Technical Basis for the Equilibrium Partitioning Method for
           Establishing Sediment Quality Criteria. Environmental Toxicology and Chemistry, 1991.11(12):
           p. 1541-1583.
                                 s
           9.     Mahony, J.D., et al., Partitioning of metals to sediment organic carbon. Environ. Toxicol.
           Chem., 1996.15(12): p.  2187-2197.

          . 10.    Gasas, A.M. and E.A. Crecelius, Relationship between acid volatile sulfide and the toxicity
           of zinc, lead and copper in marine sediments. Environ. Tox.icol.  Chem, 1994.13: p. 529-536.

           11.    Berry, W., et al, Predicting the toxicity of sediments spiked with silver. Environ. Toxicol. •
           Chem., 1999.18: p. 40-48.

           12.    Kemble, N.E., et al, Toxicity of metal-contaminated sediments from the upper Clark Fork
           River, Montana, to aquatic invertebrates and fish in laboratory exposures. Environ. Toxicol. Chem.,
           1994.13: p. 1985-1997.

           13.    Call, D. J., et al., Silver toxicity to Chrionomus tentans in two freshwater sediments. Environ.
           Toxicol. Chem., 1999.18: p. 30-39.
                                                    2-28

-------
g
i
IZU


100

80

60
40

20
0


20


o Field Amphlpod
• V Field Midge
•Lab Amphlpod
DLabPolychaete

•
•
«V j
- o *.*^
o V,°8.°J
1
1
	 1 . . ......1 	 1 1

1
1
tex*pmm
1 • • O .*>
0 «
O •
o a
0 *
---V0 vo_
•
o •
V ° 0
ov *V ° o
>• • o .
• O _
o V
•  0
%
V °l
_• . o


•1 	 '"1 •
'
o DD m»o
0 *<
o
U
.° 0
0
_
'
£j
•
0 . ,
•> o
V .01
' v ".£
o •v+i
CT^fc -flD •
I
1


1
>•§•••••••• •
>.* /*o
•
5
•
'
r • ''
.
•
i
V
***

•
_

       0.1      1       10      100     1000    10000    100000  1000000

                         SEM-AVS/f oc (umol/goc)

       Figure 1. Mortality versus [SEM]-[AVS] difference (top panel) and
       this difference normalized to organic carbon (bottom panel) on a
                           sum total SEM basis.
                            2-29

-------
1
     120
     100
      80
      60
      40
      20
                                CADMIUM
     •20
P Field Amphlpod
V Reid Midge
• Lab Amphipod
OLabPolychaete
i
     100
      80
      60
      40
      20
       0.001     0.01      0.1      1       10     100     1000    10000

                            SEM-AVS (umol/g)
                                     4. -Q-
                                         8
                                                 • *    •
       0.01     0.1      1      10      100     1000    10000    100000

                          SEM-AVS/foc (umol/goc)

        Figure 2. Mortality versus [SEMHAVS] difference (top panel) and
     this difference normalized to organic carbon (bottom panel) for cadmium.
                            2-30

-------
                            COPPER
120

100


80

s*
C, 60
£•
•"«•
f «o

20


0
.90
— I I mini i i IIIHII — i i IIITIII-
o Field Amphlpod
• V Field Midge

• Lab Amphlpod
PLabPolychaete
•
_'

o
v °

V
b
o
o


'
1 .» 900*0
o o
o
o
0

^.v.-!.
.
V V
1 v

o




•


.

m


• .
.

_


  0.1
  0.001    0.01     0.1
                10      100     1000    10000
                        SEM-AVS (umol/g)
120
100



80

60
40

20
)
0
-20
	 '" ' * •"•••i 	
. ; .



-


o
V
.V
o
o


0 D D «0
O
o
• o

o
-5 °^

o
v ^

0°


• •• • •
o
.

•
'
I_ , '


.



10     100    1000    10000   100000  1000000
                     SEM-AVS/foc (umol/goc)

   Figure 3. Mortality versus [SEM]-[AVS] difference (top panel) and
this difference normalized to organic carbon (bottom panel) for copper.
                        2-31

-------
S
S
 S
120
                               100
                               80
                               60
                               40
                               20
                                                          NICKEL
                              •20
o Field Amphlpod
• V Field Midge
• Lab Amphlpod
DLabPolychaete
•
V
V
0°
'•
• • 1 •
0
	 o_ _
° 0
o
o
• •
• •.
0

'
 0.001    0.01      0.1       1       10      100     1000    10000

                       SEM-AVS (umol/g)
WO
100
80
60

40
20
0
A/I
\
-
.

V
V
0°

•
o

- _
-------
                                  LEAD
120

100


80
60

40

20

0

	 	 *' • " ""' " ""' .••••••I
o Field Amphlpod
• V Field Midge
. • Lab Amphlpod
DLabPolychaete

"

•
V

V .
o


1*

O
•

'
_ -

-._____
V • Q






-
.


-

-


'


£
       0.0001    0.001     0.01     0.1
                             SEM-AVS(umol/g)
               10      100     1000
1ZU
100
80
60


40.
20
0

-20


.
• • '
1 • i
"
^ — _ _» -1

• •
-
\
o
. .


D


1 _ _ _ -J

"
D •



.
• £ . .
9
.
_


•
'




       0.01      0.1
10     100     1000    10000   100000
                          SEM-AVS/foc (umol/goc)
        Figure 5. Mortality versus [SEM]-[AVS] difference (top panel) and
       this difference normalized to organic carbon (bottom panel) for lead.
                             2-33

-------
s
s
t
                         i
120
                              100
                               80
                              .60
                               40
                               20
                             ZINC
       O Field Amphipod
       v Field Midge
       * Lab Amphipod
       DLabPolychaete
                              .20'  	
120
                              100
                               80
                               .
                               40
                               20
                              •20
                                0.1
                        V

                    O
                    CD   O O
•    O
                                0.001     0.01     0.1
                                 10  .    100     1000    10000
                       SEM-AVS (umol/g)
                                                        Wc?o
                  10     100     1000   "10000    100000  1000000

                     SEM-AVS/foc (umol/goc)
                                 Figure 6. Mortality versus [SEM]-[AVS] difference (top panel) and
                               this difference normalized to organic carbon (bottom panel) for zinc.
                                                     2-34

-------
     120
     100
      80
t.    60
      40
      20
                      MIXTURE - LAB AMPHIPOD TEST
     -20
     120
     100
     80
     60
     40
      20
     -20
       0.1
      0.001    0.01     0.1
               10     100    1000    10000
                            SEM-AVS (umol/g)
10     100    1000    10000   100000  1000000
                         SEM-AVS/foc (umol/goc)
        Figure 7. Mortality versus [SEM]-[AVS] difference (top panel) and
        this difference normalized to organic carbon (bottom panel) for an
            amphipod bioassay test spiked with a four metal mixture.
                           2-35

-------
t
                  SILVER -LAB AMPHIPOD TEST




g
1




120
100
80

60
40
20
0

-20

- ' • • • 1
.
.
'• •' ,
•
-
•
-



• •

P

'
_
•
X
•



•
-
• ' •
-
-
-
-


                              0.0001   0.001   . 0.01     0.1
                                        10      100    1000
t
                      SEM-AVS(umol/g)
120
                             100
                              80
                        C-    60
                              40
                              20
                             •20
S
  0.01     0.1      1       10      100     1000    10000    100000

                    SEM-AVS/foc (umol/goc)

   Figure 8. Mortality versus [SEMHAVS] difference (top panel) and
 this difference normalized to organic carbon (bottom panel) for silver.
                                                   2-36

-------
I
     120
     100
      80
      60
      40
      20
     •20
           vvw
                                                                                I
                  y*v
                                                                          t
0.001    0.01     0.1
                                      10
100    1000    10000
                            Average Metal/ERM
        Figure 9. Mortality versus the average Total Metal/ERM quotient
             for field data (o) and laboratory spiked data (v).
                                                                                 I
                               2-37

-------
                      Chromium Chemistry in Sediments

                                        by

                     Dominic M Di Tore ^ and John D. Mahony*

                           *HydroQual; Inc; Mahwah, NJ
       * Environmental Engineering and Science Dept. Manhattan College, Bronx NY

                                    Marchl999

Introduction

       The purpose of this section is to present a review of the  aqueous chemistry of
chromium as it influences its toxicity in sediments. The speciation chemistry of chromium is
presented in Fig.l, and Eh- pH diagram [1]. The principle species that exist at the indicated
Eh (redox potential) and pH is listed.   Chromium exists in.two oxidation states in natural
systems.  Oxidized Cr(VTj species are present as oxyanions CrO42" and HCrCV. The reduced
CrCffl) species are Cr3"1" and the hydroxide complexes CrOH2+, Cr(OH)2+, Cr(OH)3°  and
Cr(OH>4". Although chromium sulfide minerals are known to exit, e.g. Cr2S3(s) [2] and 0384
(Brezinaite) [3] they hydrolyze to chromium hydroxide upon exposure to water [4].

       Cr(IH) concentrations are regulated by an insoluble hydroxide Cr(OH)3(s) in the pH
range from 6 to 11 (the shaded region in Fig.l). The actual concentrations can be seen in
Fig.2 from [5],  Solubility limits the dissolved concentration to approximately 10~7 M in the
pH range  of approximately 6-7  to 10.  The water quality criterion  for Cr(III) is
approximately 10"6 M (74  g Cr/L) [6]. Therefore in the pH range of approximately 6-7 to
10 we expect no toxicity since the interstitial water toxic unit concentration is approximately
10"7 M/10"6 M = 0:1. By. contrast to Cr(DI), the oxidized Cr(VI) species are soluble and are
also toxic  The EPA chronic water quality criteria for Cr(VI) are  11 and 50  gCr/L  in
freshwater and saltwater, respectively.
                                      2-38

-------
       Reduction of Cr(VI) to Cr(HI)

       The reduction of Cr(VI) to CXIH) can occur only in reducing environments (Fig.l).
The reduction can occur by bacterial action [7-10], by ferrous iron [11-13], by organic matter
[14-16], photochemically [17], in soils [18], in aquifer material [19], by hydrogen sulfide [20,
21], pyrite fines [22] and amorphous iron sulfide [23]. We have also examined the reduction
of chromate using  amorphous iron  sulfide in neutral solutions in support of the effort to
develop an EqP based sediment criteria for chromium by the EPA.

       For the data presented in Fig.3, the stoichiometry of the reaction appears to be
       CrO42' + FeS + 3H2O -> Cr(pH)3(s) + FeOOH(s) + S° + 2OJT
(1)
i.e. one mole of chromate -reduced per mole of FeS initially present.  We have also seen
titrations that correspond to complete oxidation of the sulfide to sulfate when sediments are
                           i    .            >.                        .
used. The important fact, from the point of view of sediment criteria, is that if any FeS is
present in sediments, then all the chromate would have been reduced to chromium hydroxide
by the reduction reaction (eq.l).                                        '
             »    -                                                        *'     ,
       Oxidation of Cr(HI) to Cr(VI)

       The reverse reaction: the oxidation of CrQH) to Cr(VI), is of direct concern because
Cr(VI) is soluble and toxic.  The oxidation rate of Cr(in) to Cr(yi) with oxygen as the only
oxidant is quite slow with half-lives of months reported (see [24] for a review). Oxidation
can occur rapidly (in minutes) with hydrogen peroxide as the oxidant [24]  leading to the
suggestion that the photochemical production of HiOj  in surface waters can be producing
Cr(VI). For sediments, however, the likely oxidant is manganese dioxide, which has been
shown to oxidize Cr(ni) to Cr(VI) rapidly [25-29].
                                                                          . ^

       At acid pHs the stoichiometry of the reaction appears to be [30]
                                      2-39

-------
                                                                        (2)
which goes to completion very quickly (minutes to hours); We have modeled, the reported
data for pH = 4.5 assuming that the oxidation rate is 1/2 order with respect-to Cr3"1", which
suggests a surface mass transfer limitation, and saturation kinetics for MnOa.
                          •                            •               T        -
The kinetic equation for the concentration, Cr, of Cr(in) is
                                                                  .      (3)
                     dt          MnOi+K^
where Cr and MnOz are both functions of time, k is the reaction rate constant and KMDOI is
the half saturation constant. The concentration of MnOa is determined by mass balance based
on the stoichiometric eq.(2). The data for experiments at varying MnOi(s) concentrations are
shown in Fig.4.  Both the Cr(VI) and Mn(H) produced by the oxidation are shown.  The
Mn(n) concentrations are multiplied by 2/3 from eq.(2) so that their concentrations should be
the same as the Cr(VT). Other data are available for pH = 3-5 which indicates that the
oxidation rate increases with pH [29].                           .

       The situation at neutral pHs is less clean . Since Cr(in) forms in insoluble Cr(OH)3(s),
the issue is whether chromium hydroxide can be oxidized by MnC>2(s). Takacs (1988) reports
a slow rate of oxidation from a single experiment  However Johnson and Xyla [27] report
that the rate is independent of pH.  It is clear that further experimentation is required to settle
this important issue.
            '  /
                      i •       '  '      ,        -   •  •         •   .    '   •
       SorptionofCr(III)andCr(VI)

       The  sorptipn  of  Cr(in)  and Cr(VI)  are  important reactions  that  limit  the
bioavailability of chrome to organisms in both the water column and sediment. In addition,
complexed forms of chrome can diffuse from the pore water of the sediment to the overlying
                                      2-40

-------
 water.  The sorption of Cr(VI) to hydrous iron oxide has been studied extensively and the
 surface complexation model has been shown .to apply [31-33].
                                            i
                                                      (             •
       The modeling of the sorption of Cr(in) is complicated by the precipitation kinetics of
 Cr(OH)3. Studies of Cr(in) sorption to silica have been carried out at low pHs [34, 35]. We
 have analyzed a set of sorption data generated at neutral pH [36] with a partitioning model.
 Varying amounts of DOC and suspended matter were added to solutions of Cr(ni) at initial
 concentrations  of Cr(ni)r = 1.0, 0.5, 0.2 and 0.1 mg/L.  Dissolved Cr(IH) and Cr(VI)
 concentrations  were measured.  No Cr(Vi) was detected, indicating that no significant
 oxidation occurred..  However, the concentration  of dissolved Cr(ni) varied systematically
 with increasing DOC and suspended solids. The results of a linear partitioning model which
 considers the species: Cr3*, Cr=DOC, and Gr~SS, and CrCOH)3(s) is shown in Fig.5.

 The linear partitioning between Cr3* and Cr(OH)3(s) is modeling the initial  stages of
 precipitation of Cr(OH)3(s).  Fig.5A (DOC = 9.4 mg/L) and Fig.5B  (DOC  = 24 mg/L)
 present the data  and the model results. The total dissolved  chrome is plotted versus  the
 suspended solids (SS) concentration. Each line represents a different initial concentration of
 Cr(m)T, which is plotted, on the left edge of the plot as a star for visual reference.  Fig.5C
 presents the concentrations of species computed from the model for one case: Cr(in)T =  1.0
 mg/L and DOC = 24 mg/L.
 Chromium Cycle                 ,

 A summary of the chromium cycle in natural waters is presented in Fig.6 [1]. The cycling
,from Cr(VI) to CrfHT) in sediments is illustrated, as is the oxidation of CrfHI) to Cr(Vl) by
 manganese dioxide MnOj.  It is this latter reaction that is the only source of concern.
 Preliminary experiments that we have performed indicate that the reaction is extremely slow,
 and is limited by the solubility of Cr(OH)3(s).  Therefore we do not expect that this is a
 source of concern.
                                       2-41

-------
      The formation  of Cr(OH)3(s) lowers that dissolved concentration at  SS=0 below
Cr(HI)T. Note that if the normal solubility product formulation were applied, [Cr3*] -
1(T7  M  and  since [Cr=DOC]  = KODOC [Cr3][DOC], the dissolved  Cr(HI)  = [Cr3*] +
[Cr=DOC]  would also be  constant.   The fact that dissolved  chrome is varying is  a
consequence of the system not having reached equilibrium with respect to Cr(OH)3(s). As
shown in Fig.5C, most of the dissolved chrome is complexed to DOC and, therefore, is
presumably, not  bioavailable [37].   The results of this modeling  exercise  are sorption
constants of Cr(m) to DOC and SS.
                                     2-42

-------
                                   References
 1.     Richard, F.C. and A.C. Bourg, Aqueous Geochemistry of Chromium: A Review. Wat.
 Res., 1991.25(7): p. 807-816.

 2.     Wells, A.F., Structural Inorganic  Chemistry. 3rd ed.  1962, London: Oxford
 University Press. 1055.

 3.     Vaughan, D J. and J.R. Craig, Mineral Chemistry of Metal Sulfides. 1978, Cambridge,
 UK: Cambridge Univerity Press.

 4.     Durrant, P.J. and B. Durrant, Introduction to Advanced Inorganic Chemistry. 1970,
 London: Longman.

 5.     Rai,  D., B.M.  Sass, and  D.A. Moore, Chromium(IlI) hydrolysis  constants and
 solubility ofchromium(UI) hydroxide, biorg. Chem., 1987.26: p. 345-349.
                                    s..         -     -
 6.     EPA, Ambient Aquatic Life Water Quality Criteria for Chromium  EPA 440/5-84-
 029.1984, Washington4>C, 20460: US Environmental Protection Agency. 99.

 7.     Ohtake, H. and Hardoyo, New Biological Method for Detoxification and Removal of
 Hexavalent Chromium.  Water Science and Technology, 1992. 25: p.  395-402.
                                                  i
 8.     Schmieman,  E.A., et al., Bacterial reduction of chromium.  Applied Biochemistry
 and Biotechnology, 1997. 63-5: p.  855-864.

 9.     Blake, R.C., et al., Chemical transformation of toxic metals by a pseudomonas strain
from a toxic waste site.  Environmental Toxicology and Chemistry, 1993. 12: p. 1365-1376.

 10.    Wang, Y.T.  and H. Shen,  Modelling cr(VI) reduction by pure bacterial cultures.
 Water Research, 1997..31: p. 727-732.

 11.    Eary, L.E. and D. Rai, Kinetics ofchromate reduction by ferrous ions derived from
 hematite and biotite at 25 C. Am. J. Sci., 1989.289: p. 180-213.

 12.    Buerge, I.J. and S.J. Hug, Kinetics andph dependence of chromium(VI) reduction by
 iron(H).  Environmental Science & Technology, 1997. 3.1: p.  1426-1432.
                             **-•_,•'
 13.    Fendorf, S.E. and  G.C. Li,   Kinetics of chromate reduction by ferrous  iron.
 Environmental Science & Technology, 1996. 30: p.' 1614-1617.

 14.    Wittbrodt, P.R.  and C.D. Palmer,  Reduction ofCr(VI) in the presence of excess soil
 Julvicacid.  Environmental Science & technology, 1995. 29: p. 255-263,
                                     2-43

-------
15.    Elbvitz, M.S. and W. Fish, Reaox interactions of cr(VI) and substituted phenols:
products and mechanism: Environmental Science & Technology, 1995. 29: p. 1933-1943.

16.    Deng, B.L. and A.T. Stone,  Surface-catalyzed chromium(yi) reduction: reactivity
comparisons of different organic reductants and different oxide surfaces.  Environmental
Science & Technology, 1996. 30: p. 2484-2494.

17.    Hug, S.J., H.U. Laubscher, and B.R. James,   Iron(IU)  catalyzed photochemical
reduction  of chromium(Vl) by oxalate  and citrate in aqueous solutions.  Environmental
Science & Technology, 1997. 31: p. 160-170.

18.  "• Eaiy, L.E. and D. Rai,  Chromate reduction by subsurface  soils under acidic
conditions. Soil Sci. Soc. Am. J., 1991. 55: p. 676-683.

19.    Anderson, L.D., D.B. Kent, and J.A. Davis,  Batch, experiments characterizing the
reduction ofcr(VI) using suboxic material from a mildly reducing sand and gravel aquifer.
Environmental Science & Technology, 1994.28: p.  178-185.

20.    Smillie, R.H., K. Hunter, and M. Hunter, Reducion of Oiromium(VT) by bacteriatty
produced hydrogen sulphide in a marine environment. WatRes., 1981.15: p.  1351-1354.

21.    Pettine, M., et al, Effect of metals on the reduction of chromium (VI) with hydrogen
sulfide. Water Research, 1998.  32: p.  2807-2813.

22.    Zouboulis, A.I., K.A. Kydros, and K.A. Matis,  Removal of hexavaleht chromium
onions from solutions by pyrite fines. Water Research^ 1995. 29: p. 1755-1760.

23.    Patterson, R.R., S. Fendorf, and M. Fendorf,  Reduction ofhexavalent chromium by
amorphous iron sulfide. Environmental Science & Technology, 1997. 31: p.  2039-2044.

24.    Pettine, M. and F.J. Millero, Chromium speciation in seawater: The probable role of
hydrogen peroxide. Limnol. Oceanogr., 1990.: p. 730-736.

25.    Schroeder, D.C. and G.F. Lee,  Potential transformations of chromium in natural
waters. Water, Air, and Soil Pollution, 1975.4: p. 355-365.

26.    Eary, L.E. and D. Rai, Kinetics of Chromium(in) oxidation to  Chromium(VI) by
reaction with Manganese Dioxide. Environ. Sci. Tech., 1987.21: p. 1187-1193.

27.    Johnson, C.A. and A.G. Xyla, The oxidation ofchromium(III) to chromium(VI) on the
surface of manganite (gamma-MnOOH). Geochim. Cosmochim. Acta,  1991. 55: p. 2861-
2866.                                               :  .  .

28.    Fendorf, S.E.   and  R.J. Zasoski,   Chromiumflll) Oxidation  by delta-MnO2  .1.
 Characterization. Environmental Science & Technology, 1992. 26: p. 79-85.
                                      2-44

-------
29.    Silvester, E., L. Charlet, and A. Manceau, Mechanism ofChromium(III) oxidation by
Na-Buserite. J. Phys. Chem., 1995.99: p. 16662-16669;

30.    Takacs, M.J., The Oxidation of Chromium by Manganese-Oxide: The nature and
Controls of the Reaction. 1988, Michigan State:

31,    Dzombak, D.A. and  F.M.M. Morel,  Surface Complexation Modeling.   Hydrous
Ferric Oxide. 1990, New York, NY: John Wiley & Sons. 1-393.

32.    Mesuere, K. and W. Fish,  Chromate and Oxalate Adsorption on Goethite ./.
Calibration of Surface Complexation Models.  Environmental Science & Technology, 1992.
26: p.  2357-2364.
            .                              •        '            -

33.    Mesuere, K. and W. Fish,  Chromate and Oxalate Adsorption on Goethite .2. Surface
Complexation Modeling of Competitive Adsorption.  Environmental Science & Technology,
1992.  26: p. 2365-2370.

34.    Fendorf, S.E., et aL,  Mechanisms of chromium(III) sorption  on silica J.  cr(Iff)
surface structure derived by  extended x-ray absorption fine  structure spectroscopy.
Environmental Science & Technology, 1994. 28: p. 284-289.

35.    Fendorf, S.EJ and D.L. Sparks, Mechanisms of chromium(ffl) sorption on  silica ,2.
effect of reaction conditions. Environmental Science & Technology, 1994. 28: p. 290-297.

36.  ,  Masscheleyn, P.H., et aL, Chromium redox Chemistry in a Lower Mississippi Valley
Bottomland Hardwood Wetland. Environ. Sci. Tech., 1992.26: p. 1217-1226.

37.    Campbell, P.G.C., Interactions between Trace Metals and Aquatic Organisms:  A
Critique of the Free-ion Activity Model, in Metal Speciation and Bioavattability in Aquatic
Systems, A. Tessier and D.R. Turner, Editor. 1995, John Wiley & Sons:
                                     2-45

-------
o
 O)
-1

-2

-3

-4

-5

-6

-7

•8

-9
                                     PH
                                          8   10    12   14
             Fig.1. Stability fields for chromium species [Richard and Bourg, 1991]
                                                              WQC
           h O    6-8d
             V  18-22 d
           .O
             D    134 d
O O
         Cr(OH)J
           3    4    56   78    9    10   11   12   13   14   15

                                       PH
Ftg.2 (A) Stability fields for chromium species. (B) Solubility of chromium hydroxide
[Rai, SassetaL, 1987]
                                   2-46

-------
                 2000
                               ,  i	;	r	1
                        Initial AYS » 475 umoI/L
                               soo        1000      , 1500
                                  Initial CrO42~ (jimol/L)
             2000
           • Fig. 3. Titration of AVS with chromate. The final chromate
            concentration in solution is plotted vs the initial chromate concentration.
            Line corresponds to a 1:1 stoichipmetry for chromate reduction by FeS.
        •  •  O  4.6 uM
        '  T  V  9.1 uM
        "  +O  19 uM
•  O  29 uM
   T  290 uM
              10    20    30    40    50
                    Time (min)
   10    20    30    40    50

         Time (min)
Fig. 4. Oxidation of Cr(III) by MnO2. Initial concentrations of MnO2 indicated in the legend. Open
symbols are Cr(VI) concentrations, closed symbols are 2/3 Mn(II) concentrations. The lines are
model computations.
                                   2-47

-------
          DOC - 9.4 mg/L
    1.00
    OJD1
         0  200 400 600 800 1000
 DOC = 24 mg/L
                                   i.ooy
                                   0.1 o*
                                   0.01
0  200 400 600 BOO 1000

Suspended Sot Mi (ra«/L)
Speclation
                                                                      0  200 400 600 800 1000
Fig.5. Model of CiflU) sorptionto suspended solids and DOC. Varying initial concentrations of Cr(in>r (see the
stars plotted on the y-axis). Plots of total dissolved chrome vs total suspended solids. Alternating filled and hatched
symbols represent the data for each Cr(ni)r. (A) DOC=9.4 mg/L. (B) DOC = 24 mg/L. (C) An example of the
computed speciation for me DOC = 24 mg/L and CrQtth = 1.0 mg/L case.
                                       input

                                     + FeQD or
                               f    -org. matter
                          CrCVD
.— ^Jj
weak
adsorption
V
Ktt%
dfflwfcp /
\ V
CKVD
K.
+MnQz
X
adsorption
or
precipitation
I
settling
CrOnJto
sedimentation
1
V
Cr(IID
*"— ** diffusion — ^
^x
, + organic
matter
\ ^
v/
CrCDO-org
t
1
(f^bsibn.
CrOID-org
+ dissolvedx/
organlcs
                    Fig.6. Chromium cycling in the aquatic environment [Richard and Bourg, 1991]
                                                2-48

-------
2-49

-------
            The Addition of Chromium to the Metals Mixtures ISG

                                         by.
                                    Walter J. Berry
                                 Warren S. Boothman
                   U.S.EPA, Atlantic Ecology Division, Narragansett RI

                                   David J. Hansen
                   Great Lakes Environmental Center, Traverse Gity MI
                                     March 1999
       The usefulness of the measurement of acid volatile sulfide (AVS) as a tool to predict
biological effects is limited by the small number of metals for which it is currently applicable.
Therefore, it is of interest to expand the use of AVS to include predictions of biological effects for
as many metals as possible. One likely expansion candidate is chromium. AVS should be useful in
the prediction of biological effects from  chromium in sediments because (1) chromium III is
sparingly soluble and therefore nontoxic in typical freshwater and marine  sediments that are
anaerobic, contain measurable concentrations of AVS, and have interstitial water pHs ranging from
about 6.5 to 11.5; (2) once formed, chromium ffl does not oxidize to chromium VI; (3) soluble and
toxic chromium VI can only occur in sediments with no detectable AVS; and (4) only chromium VI.
will occur dissolved in interstitial water.  (See chromium section of the Metals Mixtures ESG
document.) The data available for assessing the usefulness of these predictions are discussed below.

       Toxicity tests were conducted with the ampbipod (Ampelisca abdita) exposed to sediment
spiked with chromium VI as potassium dichromate (Berry and Boothman, 1999).  The authors
examined the lexicological implications of the reduction of toxic chromium VI to insoluble nontoxic
chromium m (Kaczynski  and Kleber,  1994) in anaerobic sediments.  This  reaction is
environmentally significant for the reasons described above. The results of these tests are reported
in detail here because the results are unpublished, and because so little other data are available from
experiments with sediments spiked with chromium.

       Ten day toxicity tests with A abdita were conducted by Berry and Boothman (1999) using
methods  described by Berry et al. (1996).  Sandy sediments from  saltwater Ninigret Pond, RI  .
containing 1.7 ,umole AVS/g and about 0.15% TOC were spiked to achieve nominal chromium
concentrations of 11 to 520 //g/g.  The pH of the spiking solution was adjusted with sodium

                                         2-50

-------
hydroxide to 7.6 prior to addition to the sediment. Initial sediment interstitial water pH ranged
between 7.3 in the control and 8.2 in the highest concentration. A modification of the Cranston and
Murray (1978) method was used to determine concentrations of total chromium and chromium VI
in interstitial water. Chromium VI in simultaneously extracted metals (SEM) was determined using
the Wang et al. (1997) modification of the method of Cranston and Murray (1978).

       In the sediments that were not toxic to A. abdita, all of the chromium in the SEM was present
as chromium III; concentration ranged from 0.24 to 2.08 yumoles/g. Sediments with 3.73 and 6.54
pinoles total chromium/g were  lethal (Table 1, Figure  1). In SEM of the two toxic sediments,
concentrations of chromium VI  were 1.01 and 4.08 //moles/g. Average measured total chromium
concentrations were not appreciably different from nominal concentrations. AVS concentrations in
nontoxic sediments ranged from  0.64 to 1.42 jumoles/ g sediment, but AVS was not detected in toxic
sediments.  The, mean concentrations of AVS exceeded those of SEM in two of four nontoxic
chromium-spiked sediments (day 0-10), and three of four nontoxic sediments (after 10 days).
Chromium VI was absent from the interstitial waters of nontoxic sediments. IWTUs of chromium
VI explained observed toxicity: those sediments in which chromium was not detectable in the
interstitial water were not toxic, those with greater than one toxic unit were toxic.

       The results of these tests  were consistent with other chromium experiments performed with
amphipods at the U.S.EPA laboratory in Narragansett, RI. (W.Berry, unpublished data). Grill was
not toxic in water-only tests in which the test solutions were buffered to ambient sea water pH, but
were toxic if the test solutions  were not buffered (pH  < 6.1, Crffl > 40,000  fig/L). Similarly,
sediments spiked with pH-adjusted CrIH solutions were not toxic at concentrations up to 3,000 u£/g
dry weight.   (Sediments with  30,000 ug/g dry weight were  toxic.  These sediments were
approximately 1/3 chromium HI precipitate by volume.)

       Further confirmatory research is still needed.  Planned testing includes experiments  with
sandy and muddy saltwater sediments spiked with chromium m and an experiment with a muddy
saltwater sediment spiked with  chromium VI.  The published water-only research must also be
examined further. Most water-only tests with chromium ffl indicate a lack of toxicity when testing
is done within normal interstitial water pH ranges at concentrations below chromium III solubility.
Some of these data, however, are difficult to evaluate and some may indicate the presence of toxicity
at normal pHs.

       Taken together the results summarized above indicate that chromium in sediments is not
toxic if AVS is present and that chromium IWTUs can be used to identify nontoxic sediments.
Conversely, sediments containing measurable concentrations of chromium in SEM when AVS is not
present of when IWTUs are of toxicological significance might be toxic. This approach does not
apply to unique sediments having interstitial water pHs of less than about 6.5 or greater than about
11.5 that may occur in water bodies  with low water column pHs or near acid mine drainage.
                                         2-51

-------
                                     References
Beny WJ. and W.S. Boothman. 1999. Results of toxicity tests vriihAmpelisca abdiia exposed to
chromium-spiked sediments. Memorandum to the record. February 17,1999.

Berry, W.J., D.J. Hansen, J.D. Mahony, D.L. Robson, D.M. Di Tore, B.P. Shipley, B. Rogers and
J.M. Corbin. 1996. Predicting the toxicity of metals-spiked laboratory sediments using acid-volatile
sulfide and interstitial water normalization. Environ. Toxicol. Chem.. 15:2067-2079.

Cranston, R.E. and J.W. Murray. 1978 The determination of chromium species in natural waters.
Anal.  Chim. Acta. 99:275-282.

Kaczynski, S.E. and R.J. Kleber. 1994. Hydrophobic CIS bound organic complexes of chromium
and their potential impact on the geochemistry of chromium in natural waters. Environ. Sci. Technol.
28:799-804.                                               ;

Masscheleyn, P. H., J.H. Pardue, R.D.  DeLauna and W.H. Patrick, Jr. 1992. Chromium redox
chemistry in a lower Mississippi Valley bottomland. Environ Sci Technol. 26:1217-1226.

Wang, W-X, S.B. Griscom and N.S. Fisher. 1997. Bioavailability of Cr (III) and Cr (VI) to marine
mussels from solute and particulate pathways. Environ. Sci. Technol. 31:603-511.
                                         2-52

-------
                                  Figure Legend
Figure 1. Results of a toxicity test with the amphipod toxicity test with the amphipod Ampelisca
abdita exposed to sediments spiked with chromium VI.  Top panel: Percent amphipod mortality vs
SEM-AVS.  Middle panel: Percent amphipod mortality vs. CrVI interstitial water toxic units.
Bottom panel: Interstitial water toxic units (IWTU) vs. SEM-AVS.

                                        2-53

-------
 IA
00.
              K-*   44.   to   *-   o   o   r>
              °   2   5   b   fe   H   %
                                            q-
 bO    H-
 4*.    i->-
 O    O
              OS   O*   JO   H-    O

              y«   f-j   b   io    In
              4*   W   00   00    (^1
                        P   P   f-»   I—'   N-
                        O\   KO.  l>>   Ul   4k,
                        *•   K»   o   H-   G
              Os   U>   >-   O   o   ^*   ^'
              In   ^   ^U   i*>      4^   OS   D!   3   S
             i^
             en
             OO
             OJ
             ON
             OO
             vo •
                  K)
^
_   IA    IA    IA   -IA    IA
g  . O   O   O   O   O

«   S   S   S   8   S
                  Ui
                                 o   0,0
                                                          00
                                       «.
                                       OQ
                                •5. co
                                t«



                                 > a
                                 •^ rtj
                                 K) ?J
                                  I
                                                         *!
 *
 I
 2,
 P
 ff
 I
                                                  I
                                                  t
 CL
 X




I
 Ci
 g
 a
 I

I
 a
                                                                    o
                     2-54

-------
       Chromium Sediment Experiment
            Mortality vs. SEM-AVS
   100
 o
 *  50
     -5
               0            5
                   SEM-AVS
       10
            Mortality vs. CR6 IWTU
1 100-
o
iS(H
J>  0
    0.001   0.01
               0.1     1     10
                IWTU (2,200 UG/L)
100
1000
             IWTU vs. SEM-AVS
     1000
      100-
       10-

      0.1 -
  c£ 0^01 -
     0.001
§
M
         -2
                    2      4
                     SEM-AVS
       8
                        2-55

-------
               The Addition of Silver to the Metals Mixtures ESG

                                          by   ^
                                    Walter!. Deny
                   U.S.EPA, Atlantic Ecology Division, Narragansett RI
                                     March 1999
       Silver forms a highly insoluble sulfide. This fact makes it a natural candidate for inclusion
into a sediment guideline which uses acid volatile sulfide (AYS) normalization.  The need for
normalization and the technical basis for the prediction of the effects of silver in sediments using
AVS and interstitial water normalizations has been recently described by Berry et al. (1999), and so
will only briefly be described here.

       Previous experiments conducted with freshwater sediments spiked with silver have shown
that, when expressed on a dry weight basis, the toxicity of silver is sediment-specific and dependent
on the form of silver added (e.g. AgNO3, Ag2S: Figure Ib).  Berry et al. (1999) assessed the
usefulness  of silver  interstitial water toxic units (IWTU) and acid  volatile  sulfide  (AVS)
concentrations in predicting the biological effects of silver species across sediments, regardless of
the species of silver present. Two saltwater sediments were spiked with a series of concentrations
of silver. The amphipod, Ampelisca abdita, was then exposed to the sediments in ten-day toxicity
tests.  Amphipod mortality was sediment-specific when expressed on a dry weight basis, but not
when based on IWTU or simultaneously extracted metal (SEM) - AVS (Figure 2). Sediments with
an excess of AVS relative to SEM had IWTU <0.5, and were generally not toxic. Sediments with
an excess of SEM relative to AVS had silver IWTU >0.5, but no measurable AVS, and were
generally toxic.  Sediments with measurable AVS were not toxic. Re-analysis of the previously
published data from the freshwater sediments spiked with silver showed mortality to be correlated
with nominal SEM-AVS and with silver IWTU (Figures la and figure 3). Taken together, these
results support the use of AVS and silver IWTUs in predicting the toxicity of silver in sediments.

       Silver is slightly different from the other metals originally included in the Metals Mixtures
ESG (cadmium, copper, lead, nickel, and zinc) because it is essentially monpvalent in nature.  For
this reason one half of the silver concentration is what is used for comparison to AVS. Another
difference is that silver forms a sulfide that is not soluble in the normal AVS extraction. These
differences affect the way that AVS normalization is used for silver.

       Consider the case of a sediment that is contaminated with silver and other sulfide-forming
metals. The possible contingencies are summarized in Table 1. A mole of sulfide will be bound for

                                         2-56

-------
every two moles of silver (the 1:2 ratio is due to the fact that silver is monovalent) present because
the sulfide solubility product of silver is lower than that of other metals (Lide, 1995). If sulfide
exceeds the sum of the [Ag]/2 plus the other SEM metals (cadmium; copper, lead, nickel, and zinc),
measurable AVS will exceed measurable SEM; therefore, these metals should not be present in
lexicologically significant concentrations in the interstitial water, and the sediment should not-be
acutely toxic due to these metals.  If the sum of the [Ag]/2 plus the other SEM metals exceeds the
sulfide in the sediment, measurable SEM will exceed measurable AYS, and the sediment may be
acutely toxic due to these metals. Sediments which do not have lexicologically significant amounts
of these metals in the interstitial water should not be toxic due to these metals, even if SEM exceeds
AVS in the sediment.  However, any sediment in which SEM exceeds AVS should be looked at
carefully. Sediments which do have lexicologically significant amounts of metals in the interstitial
water are certainly potentially loxic due lo Ihese metals (Berry el al., 1996).

      Theoretically, almost all of Ihe silver in the sediment will be bound to sulfide in any sediment
in which there is an excess of sulfide over [Ag]/2. This is because of the extremely low solubility
of silver sulfide (Lide, 1995). However, if the sum of the [Ag]/2 plus the other SEM metals exceeds
the sulfide in the sediment, some of the other SEM metals may be present in the interstitial water and
the sediment may be toxic due to these metals. In these sediments the silver may be contributing to
the overall toxicity of metals in the sediment, by tying up sediment sulfide which might otherwise
bind the  other SEM metals. This is apparently what happened in some of the Call et al. (1999)
sedimenls, in which zinc and copper were released into interstitial water due to the addition of silver
to the sediment.

      The release of metals into Ihe interstitial water in relation to sulfide solubility is not peculiar
to silver. Berry et al. (1996) found that metals appeared in the interstitial water in the order of their
K,,,, with copper appearing last in their experiments. What is unusual about silver, however, is that
the solubility of silver sulfide in the AVS extraction is so low that any sediment with an excess of
[Ag]/2 over sulfide will have no measurable AVS present. Thus, any sediment with measurable
AVS should not have silver in the interstitial water, and should not be acutely toxic because of silver.
       It is important to remember that the data presented here apply primarily to acute mortality,
and may not address all effects due to chronic exposure or bioaccumulalion (Luoma et al., 1995).
For example, ,Hook and Fisher (1997) demonstrated in preliminary experiments that silver may
affect  copepod reproduction at  environmental  concentrations. However, taken together, the
freshwater and saltwater sediment results indicate that silverAVS relationships and TWTU can
provide insight into the role of silver in the possible toxicity of sediments. From the point of view
of AVS and SEM measurements, these results would indicate that silver can be included along with
cadmium, copper, lead, nickel, and zinc in sediment assessment. If the sum of the SEM for these
metals is less than AVS in a sediment, the sediment should not be acutely toxic due to these metals.
Furthermore, even in sediments which have an excess of metal over sulfide, as long as there is
measurable AVS any observed acute mortality should not be due directly lo silver in Ihe inlerslitial
water.
                                          2-57

-------
                                    References
Beny, W.J., D.J. Hansen, J.D. Mahony, D.L. Robson, D.M. Di Toro, B.P. Shipley, B. Rogers and
J.M.Corbin. 1996. Predicting the toxicity of metals-spiked laboratory sediments using acid-volatile
sulfide and interstitial water normalization.  Environ. Toxicol. Chem. 15:2067-2079.

Berry, W.J., M. G. Cantwell, P.A. Edwards, J.S. Serbst, DJ. Hansen 1999. Predicting toxicity of
sediments spiked with silver. Environ. Toxicol. Chem. 18:40-48.

Call DJ, Markee TP, Brooke LT, Polkinghorne CN, Geiger DL.  1999. Bioavailability and toxicity
of silver to Chironomus tentans in water and sediments. Environ Toxicol Chem, 18:30-39.

Hook S, Fisher N.  1997.  Sublethal response of zooplankton to silver: the importance of exposure
route.  Abstract. Eighteenth Annual Meeting of the Society of Environmental Toxicology and
Chemistry. San Francisco, Ca. November 16-20,1997.

Lide, D.R.., ed. 1995. CRC Handbook of Chemistry and Physics, 76* ed CRC, Boca Raton, FL,
USA, pp 8-85.

Luoma SN, Ho YB,  Bryan GW. 1995. Fate, bioavailability  and toxicity of silver in estuarine
environments. Mar Pott Bull 31:44-54.

Rodgers JA Jr, Deaver E, Rodgers PL.  1997.  Partitioning  and  effects of silver in amended
freshwater sediments. Ecotoxicol Environ Saf 37:1-9.
                                         2-58

-------
                                  Figure Legends
Figure 1. Percentage mortality of the axnpbipodAmpelisca abdita as a function of dry weight silver
concentration (A), ([Ag]/2) - AVS (B), interstitial water toxic units (IWTU) (C), and measured AVS
(D) in two saltwater sediments spiked with silver.  Nin = Ninigret Pond sediment. Pojac = Pojac
Point sediment.  Sediments below the dashed line at 24% mortality are not considered toxic.
Vertical dashed lines at SEM-AVS = 0 (B) and IWTU = 0.5 (C) indicate predicted break points in
toxicity.  Data points believed to be the result of interstitial water ammonia are included but
highlighted and not connected by lines in A. AVS detection limit is indicated as "ND" in D. (After
Berry etal., 1999)

Figure 2. Percentage mortality of the amphipod Hyaletta azteca as a function of nominal ([Ag]/2)-
AVS (A) and dry weight silver concentration (B) in four freshwater sediments spiked with silver.
(From Rodgers et at, 1996). Sediments below the dashed line at 24% mortality are not considered
toxic. A vertical dashed line at SEM-AVS = 0 (A) indicates the predicted break point in toxicity.

Figure 3. Percentage mortality of the midge Chironomus tentans as a function of nominal
 ([Ag]/2) - AVS (A), IWTU (B) and dry weight silver concentration (C) in a freshwater sediment
spiked with silver. (From Call et al., 1999). Sediments below the dashed line at 24% mortality are
not considered toxic.  Vertical dashed lines at SEM-AVS = 0 (A) and IWTU = 0.5 (B) indicate
predicted break points in toxicity.                      .
                                         2-59

-------
Table 1. Contingency table for predicting toxicity due to metals from measurements of silver,
other SEM metals, and AVS in laboratory-spiked and field sediments. (From Berry et al., 1999)
Nominal Metal and AVS in
Sediment
Measured Metal and AVS
in Sediment
Prediction of Acute
Toxicity
SilverOnly
' [Agj/2AVS
AVS > Detection limit
and
(SEM-AVS) < 0.0
AVS < Detection limit
and
(SEM-AVS) > 0.0
Sediment not acutely toxic
due to silver. No metals
detectable in interstitial
water.
Sediment may be acutely
toxic due to silver (but not if
IWTU<0.5).
Metals Mixtures
([Ag]/2+[Cd]+[Cu]+[Ni]
+[Pb]+[Zn]) < AVS
([Ag]/2+[Cd]+[Cu]+[Ni]
+[Pb]+[Zn])>AVS
but
[Ag]/2AVS
AVS > Detection limit
and
(SEM-AVS) < 0.0
AVS > Detection limit
and
(SEM-AVS) > 0.0
AVS < Detection limit
and
(SEM-AVS) > 0.0
Sediment not acutely toxic
due to these metals. No
metals detectable in
interstitial water.
Sediment may be acutely
toxic due to these metals, but
, not silver directly (but not if
sum of IWTU < 0.5).
Sediment may be acutely
toxic due to silver and /or the
other metals (but not if sum
of IWTU < 0.5).
                                         2-60

-------
M


£
?.
0
&
<£

100

80-

60

40-
20
0-









• Sed1
• Sed3
ASed6
• Sed7

i
— mm-
	 Am A 	 m A - — — ,

•»
* . •
f

*
I 	 	 	 	 	 	 	 _.
-^
f »I . ._..--,- -H 	
-2     ,10      1      2
         [Ag/2]-AVS (^Moles/g)
  o.i
    1         10
Dry wt silver (^g/g)
                              100
1000
              2-61

-------
         10      100     1000    10000    100000
              Silver (ng/g Dry Wt)
80-
1 6°-
0 40-
5$ 20-
0.



^Nin
+ Pojad
X PojacS

4 ,*
J V
- 1
1
A
A ,



-30 •   -20
                    20
                  [Ag/2]-AVS
0.01
0.1         1
 IWTU(LC50
                     10
                               40
c

*
o
^
100 -,

80-
40.
20-
0.







4. Pojad
X PojacS

X


I \ .
I
I
I
A 4-
tt x i i *i
100
uu-
80-
60-
40-
20-
A





ANin
4 Pojad
X PojacS


^
(5
.
5
4

ND
»

I
k ,

•
. •• ' •
• ^ ^
	 A_i 	 — — ' 	
0.01
0.1
                      10
                AVS (nMoles/g)
                   2-62
100

-------
   A  100
        80
   £
        60

       40

       20-F
        -10   0    10   20    30   40
               (Ag/2]-AVS(iiM6les/g)
50
  B   100
      80
  =5   60
      40
      20+
       04
       0.001    0.01     0.1      1       10
         Interstitial Water Toxic Units (IWTU)
 C  100
     80
*
I   604
5   404
     204
      0.01   0.1    1.    10   100  KXKMOOOO
               Silver (Ug/g Dry Wt)
                   2-63

-------
                               IMPORTANT NOTE
Section 3 (A Model of the Acute Toxicity of Metals) of the Integrated Approach to Assessing the
Bioavailability and Toxicity of Metals in Surface Waters and Sediments will be mailed to reviewers
on March IS, 1999 as an addendum to this document

-------

-------
                                     DRAFT
                                       - DRAFT -
                A BIOTIC LIGAND MODEL OF THE ACUTE TOXICITY OF METALS
                UI. APPLICATION TO FISH AND DAPHNIA EXPOSURE TO SILVER
                                          by
                       PAUL PAQUIN*, DOMINIC Di ToRO*3, ROBERT S ANTOREC,
                             DEVANSHI TRIVEDI* AND BENJAMIN Wu*
                       A HydroQual Inc., 1 Lethbridge Plaza, Mahwah, NJ 07430
                      B Environmental Engineering Department, Manhattan College
                         4513 Manhattan College Parkway, Bronx, NY 10471
                c HydroQual Inc., 4914 West Genesee Street, Suite 119, Camillus, NY 13031
S
                                      March 23,1999

-------
Abstract- A Biotic Ligand Model of the Acute Toxicity of Metals. III. Application to Fish and
Daphnia Exposure to Silver. When silver is discharged to a water body, speciation and complexation
reactions control its distribution among various organic and inorganic complexes and as ionic silver. It is
the distribution of the silver among these forms, as well as the other water quality characteristics of the
system, that control its bioavailability. The bioavailability can be evaluated in the context of a biotic
ligand modeling framework.  In this mechanistically based framework, the biotic ligand, the organism
tissue at the site of action, is represented in the same way as any other ligand in solution. It has a
characteristic binding site density and a conditional stability constant for each species that it reacts with.
The biotic ligand model simultaneously accounts for the speciation and complexation of dissolved silver
and competitive binding of silver and other cations at the site of action. The organism LC50 corresponds
to the point where silver accumulation at the biotic ligand reaches a critical level. This paper describes
the version of the biotic ligand model developed for silver, as applied to several data sets for fish and
invertebrates. The model is first used to analyze fish gill silver accumulation data over ranges of
hardness, DOC, chloride, pH and alkalinity. It is then used to account for the variation in acute toxicity
of silver to both fish and invertebrates in bioassays where hardness, DOC and chloride levels were
varied systematically. The biotic ligand model has a number of potential applications including use in
prediction of water effect ratios (WERs) from site water chemistry,.development of water quality criteria
for metals, and more generally, for ecological risk assessment purposes.

Key Words-   Silver   Biotic ligand model  Bioavailability    Speciation   Toxicity
                                              3-60

-------
            INTRODUCTION

            It is generally accepted that total metal concentration is not a good measure of exposure that can be
            directly related to biological effects, in either the water column or the sediment.  This acceptance arises
            from repeated demonstrations that consideration of the form of a chemical stressor and its
            bioavailability are pre-requisites for the successful prediction of effects (Di Toro et al., 1991; Ankley et
.;     '     al., 1994 and 1996; Allen and Hansen, 1996).  The importance of considering bioavailability in
            assessing ecological impacts has been recognized by both regulatory authorities and the scientific
;,,..          community (Prothro,  1993; Renner, 1997; Bergman and Dorward-King, 1997).  As a result, much
            research has been carried out during recent years to obtain an improved understanding of
it                                                                      -
TV          bioavailability and the mechanisms of toxicity, and in developing modeling frameworks that are
            appropriate for use in assessing the environmental fate and effects of metals, including silver, in aquatic
I           systems.

            This last in a series of three papers describes the biotic Hgand model (BLM), a generalized framework
            for assessing the bioavailability and acute toxicity of metals in aquatic systems, as developed and
            applied for silver. A version of the model is presented that is developed to predict the acute toxicity of
            silver to both fish and Daphnia, The  model framework as structured for silver is first described.  It is
            then used to  analyze silver accumulation data from experiments performed with fish in laboratory and
            natural waters.  This analysis will illustrate how the model predicts changes in silver accumulation at
            the gill, the site of action of acute toxicity in fish, that result from changes in water quality
            characteristics.  The framework is then applied to several bioassay data sets to refine the  parameter
            values used in the model to account for the effects of variation in water quality (e.g., hardness, DOC,
            chloride, pH and alkalinity) on the acute toxicity of silver to fish.  Finally, the application of the model
            to an invertebrate, Daphnia magna, is presented.
    I
                                                          3-61

-------
                                                                                                    :i  '.'
BACKGROUND

        Description of Model

The conceptual framework for the biotic ligand model is an adaptation of the Gill Surface Interaction
Model (GSIM) originally proposed by Pagenkopf (1983) and applied by him to several metals. A
detailed description of the generalized biotic ligand model framework, consisting of chemical speciation
and acute toxicity sub-models, and the technical basis of the model is described elsewhere (Di Toro et
al., 1999). Briefly, the model is based on the premise that toxicity is not simply related to the total metal
concentration in solution.  Rather, metal speciation and complexation reactions and interaction of the
metal and other cations at the site of action of toxicity must be considered as well. The role of
complexation is critical, since formation of both organic and inorganic metal complexes renders a
significant fraction of the total metal non-bioavailable.

The version of the biotic ligand model developed for silver is illustrated on Figure 1.  The dissolved
silver exists in solution as the free silver ion (Ag+) and as a variety of organic and inorganic metal
complexes (Figure 1, left side). The free silver ion typically represents a relatively minor fraction of the
total silver in solution.  However, it is this free metal species that is hypothesized to control the degree of
biological effects in the free ion activity model (FIAM) of toxicity, either by direct interaction at the site
of action, or indirectly, via its role in the formation of other silver complexes (Morel, 1983; Campbell,
1995). These other complexes form as a result of reactions of free silver with the other organic and
inorganic ligands present in solution, and collectively, they represent the predominant form of silver in
solution. With limited exceptions (e.g., it will be shown that AgCl appears to be toxic to fathead
    \
minnows), they are typically not considered to be bioavailable.

The metal speciation computations needed for the chemical speciation sub-model could be performed
using any of a number of alternative chemical equilibrium models.  For example, the CHEMICAL
Equilibria in Soils and Solutions model, CHESS (Santore and Driscoll, 1995), or EPA's MINTEQA2
model (Allison et al., 1991; Brown and Allison, 1987) or any one of a number of programs (Westall et
al., 1976) are available for use. Though Pagenkopf (1983) recognized the ameliorating effect of organic
matter on toxicity, the effect of organic matter was neglected in this early attempt to model metal-gill
                                             3-62

-------
interactions, since it was to be applied to data obtained using water that was low in organic matter
content. However, dissolved organic matter is frequently judged to be important, and in such instances,
the Windermere Humic Aqueous Model (WHAM; Tipping, 1994 and 1996) offers a relatively detailed
framework for evaluating metal-organic matter interactions. Given that it has been tested against an
impressive number of published data sets and with a wide variety of ambient waters, the Environmental
Chemistry Workgroup at the Pellston Workshop on Reassessment of Metals Criteria for Aquatic Life
Protection, recommended that Model V, WHAM Version 1.0 (Tipping, 1994), be used to represent
metal-organic matter interactions (Kramer et al., 1997). The chemical speciation sub-model employed
herein represents a synthesis of the advanced representation of metal-organic matter interactions of
WHAM with the chemical equilibrium framework of CHESS. However, since the requisite WHAM
stability constants have not been previously determined for silver, and a suitably designed experimental
data set is not currently available for use in performing this evaluation, it was necessary to perform a
preliminary evaluation of the requisite inputs as part of the calibration process to be described.

The remaining component of the biotic ligand model is the acute toxicity sub-model.  For fish, the site of
action of acute toxicity is the gill (McDonald et al., 1989). Silver toxicity is caused by binding of silver
at physiologically active, functionally important sites on the surface membrane of the gill. This
interaction of free silver and other bioavailable silver species at the gill results in an iono-regulatory
disturbance that manifests itself in the form of an acutely toxic effect. Specifically, in the case of
freshwater fish such as rainbow trout, silver toxicity results from disruption of the iono-regulatory
processes that control the active transport of ions such as Na+, Cl" and Ca2+ across the gill (Morgan et al.,
1996; Wood et al., 1996a and 1999). Janes and Playle (1995) proposed a conceptual model that could be
used to predict  the degree of silver binding to gills of rainbow trout (Figure 1, right side) and conducted
studies to investigate the factors that control the degree of accumulation.  Given that the acute toxicity of
silver to fish was understood to be related to accumulation of silver at the gill, this conceptual approach
seemed well suited for use in a model that could be used to predict the acute toxicity of silver to fish.  All
that was needed was a way to relate  the degree of accumulation to the organism response. As described
herein, this remaining step is accomplished by application of the model to bioassay test results.

In the context of the biotic ligand model, binding of the free metal ion at the site of action - the
membrane at the gill surface in the case of fish - is analogous to formation of a metal complex in
                                                      3-63
I

-------
solution, where the tissue at the site of action is viewed as a "biotic ligand." The metal-biotic ligand
interaction is represented in the same way as any other reaction of a metal with an organic or inorganic
ligand.  Both the binding site density of the biotic ligand, which is analogous to the concentration of an
inorganic ligand in solution, and the conditional stability constant for the metal-biotic ligand complex
and for any other cations that interact with the biotic ligand, must be specified.

Experimental results for use in developing initial estimates of model parameter values needed to predict
metal accumulation on the total fish gill, including sites that may or may not be active physiologically,
are currently available for a number of metals, including copper and cadmium (Playle et al., 1993a and
1993b) and silver (Janes and Playle, 1995). The results of these tests provide an indication of the levels
of complexing ligands and competing cations that will affect accumulation of metal at the gill. With
respect to copper, direct evidence also exists that demonstrates a relationship between the short term gill
uptake of copper and acute toxicity (MacRae, 1994; reviewed by Santore et al., 1999). The situation is
less well defined for silver, however, with recent experimental results showing that the level of silver
accumulation on the gill varies over time and that static equilibrium conditions are not necessarily
achieved (Wood et al., 1999), Moreover, only a relatively small fraction of the measured total
accumulation of silver at the gill is associated with binding to physiologically active sites (Wood et al.,
1999). As a result of these factors it has been difficult, at least in the case of silver, to establish empirical
evidence of a direct relationship between total gill silver accumulation and acute effects.  Given the
difficulties associated with establishing a definitive relationship between silver accumulation and
toxicity, this is still an area of active research. It is useful, therefore, to review the results of some recent
studies prior to a review of the model being developed for silver.
     s
        Relationship of Silver Accumulation to Acute Toxicity

Several studies have shown that when juvenile fathead minnows and rainbow trout are exposed to
copper, there is a relatively rapid increase above background levels of copper bound to the gill (Playle et
al., 1992; MacRae, 1994). This short term initial increase, which takes place over a time scale of a few
hours to a day, has been shown to be related to survival (MacRae, 1994; Hollis et al., 1997). Based on
these data, Santore et al. (1999) estimated that an increase in gill copper accumulation of 10 nmol/g,,,
above background gill copper levels, was associated with 50% mortality in juvenile rainbow trout.
                                              3-64

-------
I
During the initial stages of development of the biotic ligand model for silver it was assumed that the
uptake of silver by fish gills followed a similar pattern to that of copper, with a rapid short term initial
uptake of silver at the surface membrane of the gill, and that this initial uptake reached a plateau. Janes
and Playle (1995) showed that this short term accumulation of silver at the gill was associated with
adverse effects to fish.  They showed that the sodium efflux from the gills of juvenile rainbow trout (1-3
grams) increased as the short term (2- to 3-hour) gill loading of silver increased from about 1 to 15 nmol
Ag/g wet tissue (nmol/gw), when the fish were exposed to silver concentrations spanning the range of
LC50 levels analyzed herein (about 6-40 ug/L) in experiments carried out in low chloride waters.

Several investigators (McGeer and Wood, 1998; Bury et al., 1999a; Wood et al., 1999), in tests with
rainbow trout, reported decreases in Na+ influx to the gill and/or reduced ATPase activity as a function
of predicted free silver, while the relationship of these effects with gill silver levels measured in a
separate set of experiments, following the same experimental protocol, was not clearly defined.
However, the exposure concentrations used in these tests, about 3-4 ug/L, were significantly lower than
the range of concentrations used by Janes and Playle (up to about 50 ug/L of silver) to develop a
relationship between Na efflux from the gill and gill silver burden.  The silver concentrations they used
were also about one half of representative rainbow trout LCSOs measured in  unamended laboratory
waters (Davies et al., 1978;  Bury et al., 1999b). Since the acute toxicity data to be analyzed with the
biotic ligand model includes LC50 results in the range of about 6 to 40 ug/L, the results of Janes and
Playle (1995) are more relevant to consider for purposes of the analyses to be described herein.

Recently reported measurements  of the silver loading on juvenile rainbow trout gills over a 48-hour
period show that the gill loading varies markedly over time and with the exposure level of silver used
(Wood et al., 1999).  The results suggest that an equilibrium gill silver level  is not achieved.  Hence,
there may not be a unique equilibrium gill silver level associated with a given response. In another long
term 48-hour study with 320 gram rainbow trout, the measured Na+/K+ ATPase inhibition response did
not vary with gill silver as chloride was added (McGeer and Wood, 1998; Wood et al., 1999), but the
same response did vary with gill silver in a shorter 6-hour duration study using 8.8 gram juvenile
rainbow trout (Bury et al., 1999a). Thus, it is still possible that the short term (a few hours to a day)
accumulation of silver on the gills of juvenile rainbow trout can be related to acute effects associated
                                                      3-65

-------
 with a more extended (e.g., 72 to 96 hours or longer) exposure period, as it has been with copper
 (MacRae, 1994; Hollis et a!., 1997), Hg (Playle, 1998a) and Cd (Hollis et al., 1997). Although not yet
 confirmed by an independent set of tests, the limited results with silver suggest that the short term
 accumulation of silver may be directly related to binding at physiologically active sites at the gill
 surface, while for longer term exposures, the accumulation of silver occurs at other relatively inert sites
 as well, thereby obscuring a direct relationship between longer term gill Ag level and .effects.

 From a practical perspective, what is actually needed is to be able to predict changes in the toxic
 response that results from changes in the complexation of the metal and competitive effects on binding at
 physiologically active cells at the site of action, the biotic ligand. The reported difficulties in achieving a
 consistent correlation of effects with total gill silver may reflect the relatively minor contribution of
 silver at the active gill sites to the total gill silver load. As currently understood it is the chloride cells,
• those cells that transport the Ag+ across the gill membrane to a point where it can exert its inhibitory
 effect on the Na+/K+ ATPase activity, that correspond to the physiologically active sites on the gill
 (Bury et al, 1999a), and thus to the biotic ligand. It has been estimated that these cells represent only 10
 percent or less of the total gill cell population (Bury et al., 1999a).  Unfortunately, it is not possible to
 measure the concentration of silver associated with these cells directly, independent of other silver that is
 bound to the gill. Even so, if the total gill silver varies in a similar way to the silver bound to these
 functional sites, and/or it can be related to an acute toxicity end point, then a direct measure of the silver
 at these specific sites may not be required.
                                                                           •»

 In the interest of expediency, given the preceding difficulties, the approach followed in developing the
 BLM is to predict the acute toxicity  of metals,by use of a parameter that corresponds to a predicted
 equilibrium biotic ligand concentration. This approach provides a generalized computational framework
 for evaluating the effect on metal toxicity of both the formation of metal complexes and competitive
 binding at a biotic ligand that represents the true site of action of toxicity. The degree to which these
 effects (e.g., the "hardness correction") differ from one metal or organism to another can be evaluated by
 calibrating the model directly to toxicity data. At the same time, as the understanding of the underlying
                                                                                            (
 mechanism of toxicity is elucidated  in the future, it is expected that this modeling framework will be
 amenable to future refinements.
                                               3-66

-------
V
    I
             Given the difficulties associated with directly relating gill silver accumulation levels to toxicity, the
             calibration of the biotic ligand model for fish, as presented herein, is based on both the laboratory silver
             accumulation test results of Janes and Playle (1995) and acute toxicity data as well. Development of the
             model using toxicity data is critical, since the ultimate use of the model will be to predict toxicity.  Since
             silver accumulation data that are suitable for use in the development of the BLM for Daphnia magna are
             not currently available, when the model is developed for Daphnia, it is calibrated to toxicity data alone.
             This may ultimately be the only practical approach for organisms the size of Daphnia, since the site of
             action is uncertain and sampling and analysis of the relevant body parts, other than perhaps the whole
             body, would be exceedingly difficult if not impossible to do.

             As applied herein, the primary use of the BLM is to predict the organism LC50, the dissolved
             concentration corresponding to 50 percent mortality, as a function of water quality characteristics.  It is
             emphasized that the critical biotic ligand accumulation level used in the model, the lethal accumulation
             at 50% mortality (LA50) that is associated with the dissolved LC50 to b.e predicted, is viewed as being a
             nominal concentration. It does not necessarily correspond to, nor is it intended to be equivalent to, a
             steady state equilibrium  concentration that can actually be measured at the physiologically active sites of
             the biotic ligand.  Rather, the LA50 is simply a quantitative benchmark that can be correlated to the
             dissolved silver LC50 in water. This approach is adopted, rather than simply relying on a prediction of
             free silver concentration, to expedite the prediction of toxicity with consideration given to alternative
             mitigating factors. That is, it provides a way to evaluate not only the concentrations of free metal and
             other metal complexes in solution, but the net effect of competitive binding of the metal and other
             cations in solution at the physiologically active sites of the biotic ligand as well. The degree to which
             mitigating effects such as competitive binding between the metal of interest and other cations varies with
             organism type and metal is addressed in the context of the model calibration process.
                                                           3-67

-------
        Applicability of the Biotic Ligand Model

The current EPA water quality criteria (WQC) for copper (USEPA, 1985) and silver (USEPA, 1980) are
functions of hardness, but are independent of other water quality characteristics. Since it is known that
other water quality characteristics such as pH and organic carbon often affect toxicity as well, it is
frequently necessary to perform extensive bioassay testing to develop water effect ratios (WERs) that can
be used to establish site-specific WQC for metals. The model framework described herein provides a
computational alternative to bioassay testing that can be used to explicitly evaluate these effects.

The biotic ligand model is currently being developed for use in predicting silver LCSOs on the basis of
site-specific water quality characteristics. The relative sensitivities of the organisms to which the model
has been applied thus far are shown on Figure 2 (left panel). This figure shows the cumulative
distribution of Genus Mean Acute Values (GMAVs) upon which the 1987 draft water quality criterion
(WQC) for silver was based (USEPA, 1987). The data from the 1987 draft WQC document are used for
this illustration because they include more data than were available in 1980, when the current acute
criterion was developed (USEPA, 1980). Each data point represents the mean LC50 for a single genus,
and the points are ranked from the most sensitive organisms (low LCSOs) to the least sensitive organisms
(high LC50s). The BLM analyses to be presented for fish will use data for rainbow trout and fathead
minnows, species that have similar intermediate levels of sensitivity to silver (near the 50 percentile on
the GMAV probability distribution; USEPA, 1980 and 1987). BLM results will also be presented for
Daphnia magna, a relatively sensitive invertebrate having an LC50 that is close to the current acute
water quality criterion for silver, a Criterion Maximum Concentration (CMC) of about 1 ug/L.
     \
It should be recognized that the concentration ranges to which the BLM will be applied herein are
significantly greater than typically observed ambient levels of silver in aquatic systems (Figure 2, right
panel).  Development of the BLM at these levels is appropriate for several reasons.  First, it is envisioned
that the BLM will provide a computational alternative to measuring Water Effect Ratios (WERs) for use
in establishing site-specific criteria. Use of the BLM for prediction of fathead minnow LC50s is
consistent with the current routine use of fathead minnows in determination of WERs for silver. Further,
EPA guidelines recommend that the WER be developed using an organism that is sensitive at a level
close to the WQC (USEPA, 1994), and development of the model for Daphnia is consistent with,this
                                             3-68

-------
recommendation. These published guidelines also stipulate that the endpoint of a toxicity determination
must not be below the Criterion Maximum Concentration (CMC) or Criterion Continuous Concentration
(CCC) to which the WER is to be applied. Since the freshwater acute WQC for silver, the CMC, is about
1 ug/L, development of the BLM for application at lower concentrations would not be appropriate for
development of a site-specific freshwater acute criterion.  Although it is envisioned that the applicability
of the biotic ligand model will eventually be extended to lower chronic effect levels, the model is
currently only applicable for use in predicting acute toxicity.

The biotic ligand model framework described herein has thus far been applied to copper (Santore et al.,
1999) and, as presented below, to silver. The principal features that distinguish this model from the
earlier conceptual frameworks that have been proposed to predict metal toxicity are that (1) it is a
working program incorporating a state of the art representation of metal-organic matter complexation,
and (2) it can be used to predict site-specific acute toxicity levels for fish and invertebrates. Studies are
being planned or are in progress to develop the requisite data to evaluate the BLM parameters for other
metals, for both fish and invertebrates. Once these results are available, it is envisioned that the
conceptual framework can be extended to other metals, and in its more general  form, to organisms other
than fish.

BIOTIC LIGAND MODEL CALIBRATION

The biotic ligand model for silver will be applied to a variety of data sets for purposes of obtaining a
preliminary calibration.  As a starting point,  the model was first calibrated using total gill silver
accumulation data generated by Janes and Playle (1995) in the hopes that the resulting parameter values
would to some degree reflect the characteristics of the active sites corresponding to the biotic ligand
itself.  The model was then applied to aquatic toxicity data sets to refine the preliminary parameter
values that were  determined. The results to be presented will be based on the parameter values that were
determined as a result of following this overall approach.

        Analysis of Silver Accumulation Data
                                                                                          *
The first application of the biotic ligand model was to rainbow trout gill silver accumulation data from
                                              3-69

-------
exposure to silver in laboratory water spiked with varying levels of complexing ligands and competing
cations (Janes and Playle, 1995). This set of data is of particular interest because it demonstrates the
protective effects of both silver complexation (by the addition of DOC and chloride) and of competition
of Ag+ with Ca2+ and Na+ for binding sites on the gill. The analysis will also illustrate how the model can
be used to simulate these effects. The model is then applied to gill silver accumulation measurements in
natural waters that were spiked with silver to show that the model can be used to predict similar effects
over a range of natural water characteristics.

Janes and Playle (1995) conducted an elegant set of experiments to demonstrate the effects of
complexation and competition on the accumulation of silver on the gills of juvenile rainbow trout, and to
obtain data to evaluate the required gill site densities and gill binding constants needed in a chemical
equilibrium model that could be used to predict gill silver levels. The first step in the procedure was to
conduct tests where rainbow trout were exposed for two to three hours to 0.07 uM of silver nitrate in the
presence of increasing levels of thiosulfate.  The effect of the thiosulfate additions on silver
accumulation on the gills was monitored. The accumulation results are shown on Figure 3. The first bar
shows the background level of gill silver in the control is less than 1 nmol Ag/gram wet weight of gill
(nmol/gw).  The second bar shows the increase in gill silver to about 6 nmol/g* when the silver is added
in the absence of thiosulfate, and the remaining bars show the effect on gill silver with increasing
amounts  of thiosulfate in the test chamber. Note that at levels in excess of 0.5 uM of thiosulfate, there is
a progressive decrease in gill silver levels. This decrease was interpreted to indicate that the thiosulfate
was forming a complex with the added silver and out-competing the gill for Ag* from binding to the gill.
Since silver thiosulfate (AgS2O3~) is a very strong complex (Log K = 8.8, a value considered to be well
established) this seemed reasonable. Also, since it required about 29 times more thiosulfate than silver
to prevent the accumulation of Ag+ on the trout gills, it was reasoned that the Ag-gill conditional
equilibrium constant must be greater than 8.8. Pursuant to several simplifying assumptions it was then
shown algebraically that the Ag-gill equilibrium constant was <10.3. Based on further computations
with MINEQL+, they determined that the Ag-gill Log K = 10.0.

The preceding approach yielded a gill-Ag binding constant that appears to be quite large, at least with
regard to constants typically associated with binding to organic acid functional groups (Sikora and
Stevenson, 1994; Varshal et al., 1995).  Alternatively, if the Ag is binding to reduced sulfur groups, then
                                              3-70

-------

            it may be reasonable. However, if this were the case, then the Log K for Cu-gill binding (7.4; Playle et
            al., 1993b and Santore et al., 1999) would be expected to be higher as well, for a similar reason. To the
            degree the gill Ag is associated with the silver thiosulfate complex, this would lead to an overestimation
            of the Ag* that is bound to the gill. There is some recent evidence that shows that elevated gill silver
            levels can result from exposure offish to Ag-thiosulfate under conditions where acute toxicity to
            rainbow trout is not observed and where most of the silver is complexed as Ag-thiosulfate. For example,
            in longer term exposures Hogstrand et al. (1996) reported gill silver levels of approximately 30 nmol/gw
            at 98 mg/L Ag after 7 days and Wood et al. (1996b) reported 29 nmol/gw at 30 mg/L Ag after 6 days.
            Although the total silver concentrations in these silver thiosulfate tests were relatively high, the free Ag
            concentration was calculated to be O.003 ug/L by Wood et al. (1996b), a level that appears inconsistent
            with such high gill silver levels. For example, in a parallel set of experiments with 10 ug/L Ag added as
            silver nitrate, the gill silver was 11.5 nmol/gw at a calculated free silver concentration that was 3 orders
           , of magnitude higher, at 3.8 ug/L (Wood et al., 1996a). The inclusion of Ag-thiosulfate in the gill silver
            measurements by Janes and Playle (1995) would have significant implications to the analysis because the
            Ag-gill constant evaluated from the thiosulfate experiments was  subsequently used in the evaluation of
            the remaining conditional equilibrium constants, including a Ag-DOC equilibrium constant of Log K =
p          9.0 - 9.2.

II          Given the possibility that the Ag-thiosulfate was binding to the gills in these experiments, the BLM was
            calibrated without use of the thiosulfate data. The calibration approach used herein also differed from
'".           the approach followed by Janes and Playle (1995) because the representation of metal-organic matter
            interactions developed for WHAM was applied. As a preliminary step, the model was applied to Ag-
            DOC  sorption data obtained with humic and fulvic acids (Sikora and Stevenson, 1988). The initial
 :"          metal-proton exchange constants evaluated from this analysis were about 2.6 for humic samples and 2.2
•;£•          for fulvic samples (Figure 4 shows representative results), with a range of values associated with the
            different samples analyzed. These values were then subsequently refined over the course of the BLM
?           calibration, with the final values set at 2.0 and 1.4 for humic and fulvic acids respectively. Table 1
            summarizes the base case model input parameters that resulted from the model calibration. Departures
            from these values will be noted as appropriate in the discussion of the calibration results that follow.
t            Given the high concentrations of silver used in the experiments by Sikora and Stevenson (1988), and
            recognizing that measured proton exchange constants are typically inversely related to the level of metal

                                                          3-71

-------
tested, these initial estimates were considered to set a lower bound on the metal-proton exchange
constants to use for the silver accumulation test conditions. This effect occurs because of variations in
the strength of the organic matter binding sites. At low silver levels, binding occurs only at the strong
sites, while as the silver concentration increases, more and more weak sites form complexes with the
silver and the effective equilibrium constant decreases. The analysis of the data indicated that when
these initial estimates of proton exchange constants were used, the mitigating effect of DOC on toxicity
was underestimated (as expected), and the values were adjusted to 2.0 and  1.4 for Ag-humic and Ag-
fulvic proton exchange constants, respectively. (Note that as these values are expressed here, this
adjustment increases the degree of DOC complexation.)
The calibration of the biotic ligand model to the rainbow trout gill accumulation data (Janes and Playle,
1995) is summarized on Figure 5, These graphs compare measured juvenile rainbow trout gill Ag levels
following 2-3 hour exposures to silver (the open bar, the control and the single hatched bars for silver
treatments) with the gill-Ag levels computed by the BLM (solid bar = Ag* binding to the gill and the
cross hatched section indicates AgCl binding to the gill). The water chemistry used in these
computations is reported elsewhere (Janes and Playle, 1995) and the constants that were developed from
the calibration are listed in Table 1 and discussed throughout the remainder of this description of the
calibration of the BLM, The upper left panel shows gill silver accumulation levels as a function of the
total dissolved Ag that was added. A gill-Ag Log K of 7.3 was used in conjunction with a gill site
density of 35 nmol/gw to achieve these results.  Generally good agreement is obtained, with both
measured and predicted gill silver levels increasing as the silver concentration increases from less than
the detection limit (the control) to about 54 ug/L (0.5 uM).  The data suggest a leveling off of the .gill
silver concentration at the three highest exposure concentrations, suggesting that the available gill sites
are saturated at about  15 nmol/gw. The slight over prediction of the gill silver concentration that is
evident at the higher exposure levels occurs because a gill site density of 35 nmol/gw is assigned.
Although use of a lower gill site density would provide a better fit of these data (Janes and Playle, 1995,
used an average of 13 nmol/gw), an effort was made to maintain consistency across all data sets analyzed,
and a higher gill site density is needed to predict the observed gill silver levels in the natural water
experiments, reviewed subsequently.

The upper right panel of Figure 5  shows gill silver results obtained with increasing amounts of sodium
                                              3-72
t

-------
            added to water containing about 11.9 ug/L of silver. This set of data illustrates how a competing cation
            can reduce the level of silver on the gill. The first two pairs of bars compare mode! and data in the
            control water, with and without silver addition. With the addition of silver, the gill silver level increases
            to about 10 nmol/gw and model and data (measured) are in good agreement. Increasing the sodium
            concentration by about a factor of 4, from 400 to 1600 uM (about 9 to 37 mg/L) has a negligible effect
            on the results. However, when the  sodium concentration is increased by another factor of 10, to 16,000
            uM (370 mg/L), gill silver levels decrease to about 4 nmol/gw. The model, with the gill-Na binding
            constant set at Log K = 2.3 responds accordingly. The utility of this data set is that it provides a way to
;,           evaluate the gill-Na Log K by defining the concentration range over which this change in gill silver level
            occurs. It should be understood, however, that the values set for these gill binding constants are to some
?*          degree dependent on each other, so the overall calibration process is somewhat less straightforward than
            might at first appear.                                                    •

            Another example of a competing cation that was tested is calcium.  Janes and Playle (1995) reported that
            gill silver levels remained in the range of 5 to 7 nmol/gw at calcium levels as high as 10.6 mM (about 265
            mg/L).  These results indicate that calcium does not compete well with Ag for binding sites on the gill.
f|f          As calibrated (gill-Ca Log K = 2.3), BLM predictions for gill silver are within this range up to about
            2000 uM, and then decrease slightly to about 3.5 nmol/gw at 10.6 mM of calcium.

            The lower  left panel of Figure 5 presents gill Ag results versus chloride concentration. Until recently, it
\           was understood that chloride mitigates the toxicity of silver by forming the AgCl complex, thereby
            reducing the free silver concentration that was considered to be the bioavailable species. Although AgCl
\           was previously thought to be non-toxic, recent data, to be reviewed when the toxicity test results are
            discussed,  suggests that it may actually bind to the gills of fish and be toxic to some species (Erickson et
y*:~          al., 1998; Bury et al., 1999b).  Accumulation of gill-AgCl, in addition to gill-Ag+, is thus included in the
            BLM. The model results indicate these components separately as the solid (gill-Ag) and cross-hatched
;           regions (gill-AgCl) of the bars. The model and data both show excellent agreement up to 1500 uM of
8           chloride, with total gill silver levels of about 10 nmol/gw. When the chloride is increased an additional 6-
            fold, however, the observed and predicted total gill Ag decreases to about 4 to 5 nmol/gw. This decrease
            occurs because most of the silver in solution is now AgCl (the free silver is reduced significantly at this
            level of chloride), and the gill-AgCl Log K of 6.7 is less than the gill-Ag Log K of 7.3, so less Ag binds
                                                          3-73

-------
to the gill. It is also of interest to note that the predicted contribution of AgCl to total gill-Ag increases
with increasing chloride levels, reflecting the shift in the distribution of silver species in the water.

One of the primary difficulties encountered in the analysis of the gill accumulation data in the laboratory
water experiments was that the water had a relatively high DOC concentration, reported to be in the
range of 1.7 to 2.4 mg/L. This was significantly higher than the 0.3 mg/L of DOC in the laboratory RO
water prior to its use in these experiments. The source of this increase in DOC was attributed to mucosal
secretions and excreta from the fish (Playle, 1998b), with the effect possibly enhanced by the relatively
high fish biomass/volume (6-18 grams offish/liter) used in these short term static tests. When the BLM
was initially applied to this data set, the gill silver was consistently over-predicted. Decreasing the Ag-
giil equilibrium constant was not a viable solution since it was desirable to maintain consistency in the
parameter values used in the model for all data sets, and this change was inconsistent with the
comparisons to the other data sets, particularly the toxicity data sets. Since this increase in DOC was
apparently limited to the gill accumulation experiments, perhaps because fish biomass /volume used in
the toxicity experiments was much lower (about 1 to 2 g/L), it was decided that adjusting the
        .•                                      "                       .
characteristics of DOC from this source was reasonable for calibration purposes.  Additionally, although
information on the complexation capacity of this type of DOC was not available, it would be reasonable
to expect that it could have characteristics that differed from those of humic and fulvic acids of a
terrestrial origin. In light of these considerations, the site density of this incremental DOC from the fish
was assigned to be approximately a factor of 4 times the site density of humic acid in order to fit these
gill accumulation data. It was also assumed that the source water background DOC was 10% humic acid
and 90% fulvic acid for this data set and for the other data sets to be analyzed as well. These adjustments
were.included in the BLM results discussed previously for Figure 5  (i.e., the Ag, Na and Cl addition
tests).
Janes and Playle (1995) also conducted studies where the DOC of the water was increased from the
control water DOC of 2.4 mg/L, to about 24 mg/L at the highest treatment level (Figure 5, lower right
panel). The silver concentration was 0.17 uM (18.3 ug/L) in these tests. The data and model results are
compared on the lower right panel of Figure 5. The results provide a clear illustration of how formation
of a silver-organic matter complex reduces Ag accumulation at the gill by reducing the free silver
concentration in solution.                                 •
II
                                              3-74

-------
The DOC that was added in these experiments was isolated from Luther Marsh, Ontario, by tangential
flow ultrafiltration. No other characteristics of this source of organic matter were reported. When the
model was applied using the standard set of humic and fulvic acid characteristics developed for WHAM
for the added DOC, gill silver was overestimated. It was therefore necessary to increase the binding site
density of the Luther Marsh DOC by about a factor of seven to obtain the indicated results for the DOC
experiments. The need to increase the Luther Marsh DOC site density is consistent with results of
analyses with mercury and with mixtures of metals reported by Playle (1998a) where it was necessary to
increase the binding site density of Luther Marsh DOC by as much as a factor often. It would be
preferable if the DOC characteristics did not need to be adjusted for these different sets of experiments,
and these changes were made with due consideration given to this.  However, they were necessitated by
changes in the experimental conditions that were unique to these tests. Fortunately, when the model is
applied to the toxicity data sets, the conditions were such that the contribution of DOC from the test
organisms did not seem to be a significant factor (static renewal test procedures and/or a much lower
mass offish per unit volume of water were used). Also, the DOC that was added, Aldrich humic acid, is
commonly used and relatively well characterized, so a consistent set of DOC related inputs could be
assigned in these critical analyses.

Finally, with regard to pH, Janes and Playle (1995) reported thatpH was not a significant factor over the
range of pH levels tested. Although the  detailed data for the pH experiments were not presented, the
measured gill-Ag was reported to be about 6 to 6.5 nmoVgw over a pH range of 4.5 to 6.5 and a silver
concentration of 0.06 uM (6.5 ug/L).  The gill proton binding constant used in the analyses of Figure 5 is
Log K = 4.3. The modeling analyses of the datasets to be presented herein (Figures 5-10) were not very
sensitive to this parameter over a range of Log K values of 4.3 to 5.4, with the higher value currently
used in the copper BLM model (Santore et al., 1999). The insensitivity of the results occurs because
most tests were conducted at pH > 6.2. A gill proton binding constant of Log K = 4.3 is used here
because it is more consistent with the low sensitivity of the reported results at pH levels as low as 4.5.
Note that with the silver DOM model used in the BLM, pH changes will also affect the  results through
proton-DOC interactions. As pH  increases, DOC binding sites de-protonate and this results in an
increase in Ag:DOC binding. Over the pH range of 6.5 to 8.5, this appears to result in a decrease in
calculated free silver that is qualitatively inconsistent with the limited data considered to date (Sikora
                                             3-75

-------
and Stevenson, 1988). As a result, for purposes of the current version of the model, the inputs are
specified to avoid this effect at pH levels above 6.5. Although this is an area where future refinements to
the model are envisioned, these refinements must await the completion of a carefully designed Ag:DOM
adsorption experiment for use in this task.

The BLM has also been applied in the analysis of a similar set of gill silver accumulation experiments
conducted by Janes and Playle (1995) using a variety of natural water samples. Here, the DOC of the
laboratory control waters was characterized as before, but the natural waters were modeled using
standard WHAM-based BLM inputs for DOC. That is, the percent of the DOC that is characterized as
being humic (10%) and fulvic (90%) material is specified. The remaining Log K values are as
summarized previously in Table 1.  The measured gill silver accumulation and predicted values are
compared on Figure 6. Exposure waters I and 2 are the lab water controls, with and without silver
added, and the remaining exposure waters (3 - 8) are from surface waters in southern Ontario. The water
quality characteristics in these waters exhibited significant variability: DOC = 4.4 - 8.5 mg/L, pH = 7.4 -
8.2, Ca = 500 - 2100 uM, Na = 46 -1045 uM and Cl = 32 -1150 uM.  The measurements, the open
(control) and filled (silver treatments) are shown with the 95% confidence interval. The model results
are indicated as ranges corresponding to the predicted gill Ag with and without giil-AgCI included. The
model performs quite well in simulating the range of variations that were measured using these test
waters. This is important, since the ultimate use of the model will be to predict acute effect levels for
silver in natural waters.
It is worth reiterating at this point in the discussion that the total gill silver concentrations presented
above and predicted with the BLM should not be construed as representing the silver at physiologically
active sites that control the iono-regulatory response of the organism.  The purpose for the analysis of
    \
these data was to obtain an indication of the magnitude of the biotic ligand model input parameters, the
gill binding site density and the gill binding equilibrium constants, that are associated with protective
effects. Since  the short term gill silver uptake studies reviewed thus far were not carried out for purposes
of eliciting an acutely toxic response and none was measured, the data cannot be related to effects.  The
purpose of the next set of analyses of acute toxicity studies will be to make this connection.

        Analysis of Acute Toxicity Data
                                              3-76

-------
             The short term accumulation studies of Janes and Playle (1995) were never directly related to an acute
             toxicity end point.  Further, only a fraction of the measured gill silver is binding to physiologically active
             sites,.and this accumulation level appears to vary significantly over time and with the exposure regime
             (Wood et al., 1999). Thus, the requisite data are not available at this time to empirically relate a
             particular accumulation level to an acutely toxic effect. The biotic ligand model will therefore be applied
             to the analysis of LC50 data from bioassays to establish this connection between the predicted level of
             silver accumulation at the biotic ligand and toxicity (i.e., the LC50).  The practical test of the utility of
             the model will then be to determine if it can be used to predict toxicity levels of silver over a range of
             water quality conditions.
f?? ^
The remaining data sets are analyzed to define the relationship between a predicted biotic ligand
accumulation level and a toxic effect concentration for a range of water quality characteristics.  Since gill
silver levels were not measured in the toxicity tests, the data were analyzed with the underlying
assumption that a single biotic ligand accumulation level of silver is associated with 50 percent mortality
for a particular age or species of fish, or organism type. The analysis was originally conducted in
parallel with the analysis of the accumulation data, so the model parameter values presented previously
are employed here as well. The model will be used to explain how the variation in each of these water
quality parameters affects the bioavailability and hence the toxicity of silver.
   I
The first set of data to be analyzed was a comparative study of the acute toxicity of silver to juvenile
rainbow trout and fathead minnows (Bury et al., 1999b).  As discussed previously, these fish species are
similar in their sensitivity to silver. Figure 7 presents what are considered to be interesting and
previously unexpected results for the effect of chloride. The upper panel presents measured LC50
    s
concentrations for fathead minnows (open bars) and rainbow trout (filled bars) as a function of
increasing chloride concentration (50, 250, 800 and 1500 uM). The LC50 for rainbow trout increases
from about 7 to 25 ug/L over this range in chloride levels, but, surprisingly when first reviewed, the
fathead minnow data are nearly constant, with only a marginal increase in LC50 with chloride
concentration. The increase in rainbow trout LC50 conforms with the notion that bioavailability is
reduced at increasing chloride levels, due to the formation of AgCl and the decrease in free silver. The
fathead minnow results, on the other hand, suggest that AgCl is in fact toxic to this fish species. Erickson
et al. (1998) reported similar results for fathead minnows. Although the data were somewhat limited, a
                                                           3-77

-------
slight increase in toxicity was actually observed with increasing chloride concentration.

The middle panel of Figure 7 shows the predicted free ion concentration at each of the LCSOs. For
rainbow trout, the free ion concentration is approximately constant, but for fathead minnows, since the
LC50 silver concentration is approximately constant, the free ion concentration decreases with increasing
chloride. Hence, these data suggest that for fathead minnows at least, free ion concentration is not a
good uniformly applicable predictor of toxicity.

On the bottom panel of Figure 7, the predicted biotic ligand concentration that exists in association with
the LC5Q is shown. For rainbow trout, the gill-Ag is relatively uniform across the test conditions, at
about 12 to 18 nmol/gw, with a slightly downward trend with increasing chloride concentration. For the
fathead minnow, the bar indicates the two components of Ag on the gill, the lower portion corresponding
to bound Ag+ and the upper portion to bound AgCI.  Since AgCl appears toxic to fathead minnows, and it
has been reported to bind to the gill (e.g., Wood et al., 1999), it is assumed that it contributes to toxicity
as well. Considered in this way, the total gill silver on fathead minnows follows a similar pattern to that
of rainbow trout (without AgCl included in the. gill loading, although it is considered to bind to the gill as
well), with both exhibiting relative uniformity across the range of treatments (about 15 +/. 3 nmol/gw) in
comparison to the water LCSOs shown on the upper panel.  Note that this level of silver at the gill that is
associated with the water LC50 will ultimately be used to set the model parameter that will be used to
predict toxic effect levels.

Figure 8 shows analogous results from this same set of experiments, but in this case the water quality
variable is DOC, added as Aldrich humic acid. Here, the LC50 increases for both the fathead minnows
and rainbow trout, as would be expected. That is, since the added DOC forms an organic complex with
the silver, thereby reducing the free silver, more silver must be added to exert the same level of effect
(50% mortality). The resulting calculated free silver concentrations at the silver LC50 (middle panel) are
somewhat variable, but considerably more consistent than the  LC50 concentrations.  Similarly, the
calculated gill silver concentrations are relatively consistent for both fish species and across all DOC
treatments, at about 15 to 20 nmol/gw.

The final set of experiments by Bury et al. (1999b) considered the effect of calcium (i.e., a cation that
                                              3-78

-------
 I
contributes to hardness) on toxicity of silver (Figure 9). Here, the LC50 concentrations increased
slightly as calcium was increased from 50 to 2000 uM, but not as much as was observed with the
addition of DOC, or Cl as well in the case of rainbow trout. The calculated free silver follows a similar
pattern, increasing with increasing calcium, since the calcium does not affect silver speciation and the
calculated free silver is based on specification of the silver LC50 as the total silver concentration in the
water. Finally, the calculated gill silver level (lower panel) is about 17 to 18 nmol/gw across all of the
calcium treatment levels.  This accumulation level is also relatively consistent with the gill silver levels
presented on Figures 7 and 8 as well.
I
A second independent set of experiments is available where the effect of DOC (added as Aldrich humic
acid) and chloride additions on the toxicity of silver to 28 day old fathead minnows Was measured. The
results are shown on Figure 10. The upper panel shows the three concentrations of DOC (nominally, 0,5
and 10 mg/L) in each of 4 chloride treatments (3, 20,40 and 60 mg/L or about 100 to 2100 uM, similar
to the range tested by Bury et al., 1999b). The measured LC50 concentrations are shown on the middle
panel.  In this case, within each chloride treatment, there was a slight increase in the silver LC50 with
increasing DOC, though less than in the previous set of experiments. In some cases the increase is
masked by the fact that the  measured DOC levels were actually quite similar (e.g., the two highest DOC
treatments at 40 mg/L of chloride were approximately  9 and 10 mg/L, and the silver LCSOs were about
equal). Overall, the LC50 concentrations tended to be  higher than in the previous set of experiments,
ranging from about 20 to 35 ug/L. This trend is qualitatively consistent with the fact that the reported
unamended DOC concentrations in the studies by Bury et al (1999b) were lower than the levels in these
studies (about 0.3 mg/L, versus about 1 to 3 mg/L). As previously observed for fathead minnows,
however, the LC50 did not  increase markedly as the chloride concentration was increased. Finally, the
predicted gill silver levels, when bound AgCl is included, are approximately uniform across all
treatments, at about 15 to 18 nmol/gw. The low DOC controls tend to be slightly higher within each
chloride treatment group, resulting in a slight decrease in gill silver with increasing DOC.  This suggests
that the model may be overstating the mitigating effect of DOC for this set of experiments. However,
considering all of the data sets analyzed collectively, this unaccounted for variation is acceptable.
         USE OF BIOTIC LIGAND MODEL TO PREDICT THE ACUTE TOXICITY OF SILVER
                                                     3-79

-------
At this point a set of model parameters has been evaluated that performs a reasonable job of predicting
the accumulation of silver on fish gills over a range of water quality characteristics (Figures 5 and 6).
The model was then used to evaluate the short term accumulation of gill silver that is associated with the
96-hour silver LC50 concentration for fathead minnows and rainbow trout, again, over a range of water
quality characteristics (Figures 7-10). Based on these results, the predicted gill silver concentration
associated with the silver LC50 in water was relatively consistent, typically in the range of 15 to 20
nmol/gw.  It remains now to make use of these results to predict LCSOs.

The approach for predicting LCSOs is as follows. A critical total gill silver lethal accumulation level at
50 percent mortality, an LA50 of 17 nmol/gw_ is assumed to be associated with the dissolved silver LC50
concentration in water. This gill silver concentration includes the sum of gill-Ag and gill-AgCl for
fathead minnows and only gill-Ag for rainbow trout. A "numerical titration" is next performed with the
BLM. That is, the water chemistry for the sample of interest is specified as an input to the BLM and
silver is incrementally added to this water. At each step, the gill Ag concentration is predicted. The
dissolved silver concentration corresponding to the point where the gill silver reaches the critical gill
concentration (LA50 = 17 nmol/gw) is the predicted LC50. A more detailed description of this procedure
and an example of how it is implemented for copper is available in the companion papers (Di Toro et al.,
1999; Santoreetal., 1999).

The preceding approach has been applied to the LC50 data for fathead minnows and rainbow trout that
were analyzed herein.  The results of the predicted LCSOs are summarized on Figure 11,  which shows the
predicted LC50 versus the measured LC50 for silver. .The diagonal line is the line of perfect agreement
(predicted LC50 = observed LC50) and the dashed line corresponds to plus or minus a factor of two
times the measured LC50. Fathead minnow data are indicated with filled squares and rainbow trout data
with open squares. As shown, the predictions are typically within a factor of 1.5 of the measured LCSOs,
and more generally within a factor of two of the measured values, over a range of LCSOs of about 6 to 35
ug/L. Although some refinement to the BLM calibration could be made, this agreement was considered
quite good, given that LC50 test results will very often vary by a factor of two. Thus, some of the
deviations from the  line of perfect agreement may be attributed to the LC50 measurements themselves.

As discussed previously, it would be useful to develop the BLM for an organism that is sensitive near the
                                             3-80

-------
I
I
 acute water quality criteria for silver.  This has been done for Daphnia magna using a data set similar to
the fathead minnow data set of Figure  10 (Bills et al., 1997).  The DOC, chloride and calcium levels were
varied in this experiment, with the primary factor that mitigated toxicity being the addition of DOC.
(Aldrich humic acid). Briefly, the same model input parameters were used to predict the biotic ligand
silver accumulation levels at the LC50 concentration. A critical biotic ligand silver concentration was
then determined based on these results (LA50 = 2.3 nmol/gw). The numerical titration was then
performed with this value specified as  an input to the BLM and the LCSOs were predicted. The results of
this analysis are presented in conjunction with the preceding results for fish (Figure 11) on Figure 12.
The Daphnia magna results are indicated by the triangle plot symbol. The measured LCSOs ranged over
approximately one order of magnitude, from approximately 0.5 to 5 ug/L. As with the results for fish,
the BLM predictions of LCSOs are typically within a factor of 1.5 of the line of perfect agreement, and
consistently within a factor of two (within the dashed lines).

SUMMARY

In summary, the biotic ligand model for silver seems to work quite well. The calibrated model generally
reproduces the observed gill silver levels over a range of laboratory test conditions, and is also capable of
tracking gill silver levels in tests with natural waters having a significant range in water quality
characteristics as well. As parameterized in these calculations the Ag-Cl component of the predicted
total gill silver loading (Ag-gill + AgCI-gill) was relatively small, but it did improve agreement with the
measured gill silver data for rainbow trout. The model yielded relatively uniform 3-hour gill silver levels
of about 15 to 20 nmol/gw over a range of LCSOs, where AgCl was assumed to be toxic to fathead
minnows and relatively benign to rainbow trout over the range of DOC, chloride and hardness levels
tested. These results indicate that the proposed model framework can explicitly account for variation in
silver toxicity to both fish and daphnia that result not only from changes in hardness, but from site-
specific variations in DOC and chloride as well.

The results presented herein are considered to provide a demonstration of the viability of this general
approach for using the biotic ligand model. To date, the predictions for LCSOs have only been made for
the bioassay data analyzed herein. The model will need to be verified by application to one or more
independent sets of data as they become available. There may be situations where additional organic  or
                                                       3-81

-------
inorganic ligands that form complexes with silver (e.g., sulfide) will be present at high enough levels to
be significant. In such instances a conservative result will be obtained (i.e., the BLM predicted LC50
will be lower than a corresponding measured result).

The analysis illustrates how consideration of metal speciation and bioavailability provides an improved
understanding of how environmental exposure levels of metals are related to acute effects. The modeling
framework that was described has a number of potentially useful applications, including: (1) setting
discharge permit limits, (2) development of site specific WQC via a modified Water Effects Ratio
(WER) procedure, (3) evaluation of fate and toxicity of metals for use in ecological risk assessments, and
(4) development of updated water quality criteria for metals where, in addition to hardness, the refined
WQC might also be a function of TOC, DOC, pH and other variables which affect the speciation,
complexation and toxicity of metals in aquatic systems.

Recently reported results (McGeer and Wood,  1998; Bury et al., 1999a & b; Wood et al., 1999) indicate
that the toxicity of silver to fish may be more closely related to the free ion concentration of silver than
to gill silver levels.  Competitive ion effects seem to be relatively unimportant. Although considered
preliminary at this time, this  finding is consistent with the results presented herein  that the effect of
calcium on silver toxicity to fish is relatively minor in comparison to its effect on the toxicity of other.
metals, such as copper. Even so, the continued use of the BLM framework for silver is warranted for
several reasons.  First, the BLM framework includes a detailed speciation evaluation, and hence the
requisite exposure assessment based on free silver can still be addressed in the context of the BLM
modeling  framework. Second, it remains to be determined if the competitive ion effects that are
included in the biotic ligand model framework for silver are required for other organisms, even if not
required for fish. Finally, the BLM framework provides a computational framework for assessing the
bioavailability and toxicity of metals in general. It therefore makes sense to include silver within what is
considered to be a unified framework for use in making these evaluations.

Acknowledgment- This work was completed with the financial support of the Eastman Kodak Company
and the Silver Council.
                                              3-82

-------
I
                                                 REFERENCES

           Allen, H.E. and D.L. Hahsen, 1996.  "The Importance of Trace Metal Speciation to Water Quality
           Criteria," Water Environment Research, 68:42-54.

           Allison, J.D., D.S. Brown, K.J. Novo-Gradac, March 1991. "MINTEQA2/PRODEFA2, A Geochemical
           Assessment Model for Environmental Systems:-Version 3.0, Users Manual," EPA/600/3-91/021, USEPA
           ERL ORD, Athens, GA.

           Ankley, G.T., W.J. Berry, L.D. De Rosa, D.M. Di Toro, D.J. Hansen and M. Reiley, December 1994.
           "Briefing Report to the EPA Science Advisory Board on the Equilibrium Partitioning Approach to
           Predicting Metal Bioavailability in Sediments and the Derivation of Sediment Quality Criteria for
           Metals," USEPA, Office of Water and Office of Research and Development.

           Ankley, G.T., D.M. Di Toro, DJ. Hansen, W.J. Berry, December, 1996. "Technical Basis and Proposal
           for Deriving Sediment Quality Criteria for Metals, Environmental Toxicology and Chemistry, 15:2056-
           2066.

||         Bergman, H. L. and E. J. Dorward-King, 1997. Reassessment of Metals Criteria for Aquatic Life.
•i '-
           Protection, Pensacola, FL, SETAC Press.
s,
           Bills, T., S.J. Klaine, T.W. LaPoint, G. Cobb, B. Forsythe and M. Wenholz, June 1997.  "Influence of
j          Water Quality Parameters on Acute Silver Nitrate Toxicity to Daphnia magna," Final report No.
*. ^            -  \
6£:     .    	
           TTWET-9506.1, submitted to the Silver Council c/o the National Association of Photographic
0         Manufacturers by the Department of Environmental Toxicology, The Institute of Wildlife and
*                   *                                                                      *
           Environmental Toxicology (TTWET), Clemson University, Pendleton, SC.
f'
* :         Brown, David S. and Jerry D. Allison, February, 1987, "MINTEQA1 Equilibrium Metal Speciation
}          Model: A User's Manual," prepared for the USEPA Environmental Research Laboratory, Office of
<
*          Research and Development, Athens, Georgia.
i
I
                                                   3-83

-------
Bury, N.R., J.C. McGeer and C.M. Wood, 1999a.  "Effects'of Altering Freshwater Chemistry on
Physiological Responses of Rainbow Trout to Silver Exposure," Environmental Toxicology and
Chemistry, Vol. 18, No. 1, pp. 49-55.

Bury,N.R.,F. Galvez and CM; Wood, 1999b.  "Effects of Chloride, Calcium and Dissolved Organic
Carbon on Silver Toxicity: Comparison Between Rainbow Trout and Fathead Minnows," Environmental
Toxicology and Chemistry, Vol. 18, No. 1, pp. 56-62.

Campbell, P.G.C., 1995. "Interactions Between Trace Metals and Aquatic Organisms: A Critique of the
Free-ion Activity Model," Metal Speciation and Bioavailability in Aquatic Systems, A. Tessier and D.R.
Turner, eds., IUPAC, John Wiley and Sons.

Davies, P.H., J.P. Goettl, Jr., J.R. Sinley, 1978. "Toxicity of Silver to Rainbow Trout (Salmo gairdneri)."
Water Research. Vol. 12, pp. 113-117.

Di Toro, D.M., C.S. Zarba, DJ. Hansen, W.J. Berry, R.C. Schwartz, C.E. Cowan, S.P. Pavlou, H.E.
Allen, N.A. Thomas and P.R. Paquin, 1991.  "Technical Basis for Establishing Sediment Quality Criteria
for Nonionic Organic Chemicals by Using Equilibrium Partitioning," Environmental Toxicology and
Chemistry, Volume 10, No. 12, pp. 1541-1583.

Di Toro, D.M., H.A. Allen, H.L. Bergman, J.S. Meyer, P.R. Paquin and R.S. Santore, 1999. "A Biotic
Ligand Model of the Acute Toxicity of Metals. I. Technical Basis,"  manuscript in preparation.

Erickson, R.J., L. Brooke, M.D. Kah!, F.V. Venter, S.L. Harting, T.P. Markee and R.L. Spehar, April
1998. "Effects of Laboratory Test Conditions on the Toxicity of Silver to Aquatic Organisms,"
Environmental Toxicology and Chemistry, Vol.  17, No. 4, pp. 572-578.

Hogstrand, C., F. Galvez and C.M. Wood, 1996. "Toxicity, Silver Accumulation and Metallothionein
Induction in Freshwater Rainbow Trout During Exposure to Different Silver Salts," in Environmental
Toxicology and Chemistry, 15:1102-1108.
                                            3-84

-------
  I
           Hollis, L., L. Muench and R.C. Playle, 1997. "Influence of Dissolved Organic Matter on Copper
           Binding, and Calcium on Cadmium Binding, by Gills of Rainbow Trout," Journal of Fish Biology,
           50:703-720.

           Janes, N., and R.C. Playle, 1995. "Modeling Silver Binding to Gills of Rainbow Trout (Oncorhynchus
           mykiss), Environmental Toxicology and Chemistry, Vol. 14, No. 11, 1847-1858.

           Kramer, J.R, H.E. Allen, W. Davison, K.L. Godtfredsen, J.S. Meyer, E.M. Perdue, E. Tipping, D. van de
           Meent and J.C. Westall, 1997. "Chemical Speciation and Metal Toxicity in Surface Freshwaters," in
           Reassessment of Metals Criteria for Aquatic Life Protection: Priorities for Research and
           Implementation, Bergman, H.L., and E.J. Dorward-King, eds., Proceedings of the Pellston Workshop on
           Reassessment of Metals Criteria for Aquatic Life Protection, February 10-14,1996, Pensacola, Florida,
           SETAC Press.

           MacRae, R.K., December, 1994.. "The Copper Binding Affinity of Rainbow Trout (Oncorhynchus
           mykiss) and Brook Trout (Salvelinus fontinalis) Gills," a thesis submitted to the Department of Zoology
           and Physiology and The Graduate School of the University of Wyoming in partial fulfillment of the
           requirements for the degree of Master of Science in Zoology and Physiology.
fi
?••:
           McDonald, D.G., J.P. Reader and T.R.K. Dalziel, 1989.  "The Combined Effects of pH and Trace Metals
           on Fish lonoregulation," in: Acid Toxicity and Aquatic Animals, Edited by R. Morris, E.W. Taylor, D. J.A.
           Brown and J.A. Brown, Soc. Exp. Biol. Seminar Series, Vol. 34, Cambridge University Press,
:           Cambridge, U.K., pp. 221-242.
fe
'?•         McGeer, J> C. and C. M. Wood, August 4,1998.: "Protective Effects of Water Cl on Physiological
           Responses to Waterbome Silver in Rainbow Trout," final manuscript accepted by Canadian Journal of
{•          Fisheries and Aquatic Sciences,
*.
           Morel, F.M., 1983. "Complexation: Trace Metals and Microorganisms," in Chapter 6 of Principles of
           Aquatic Chemistry, Wiley Interscience, New York, pp. 301-308.
                                                       3-85

-------
Morgan, I.J., R.P. Henry, C.M. Wood, August 1996. "The Mechanism of Acute Silver Nitrate Toxicity
in Freshwater Rainbow Trout (Oncorhynchus mykiss) is Inhibition of Gill Na* and C!" Transport." (Draft
of manuscript published in Aquatic Toxicology)
Pagenkopf, G.K., 1983. "Gill Surface Interaction Model for Trace-metal Toxicity to Fishes: Role of
Complexation, pH, and Water Hardness, " Environmental Science and Technology, 17:342-347.

Playle, R.C., R.W. Gensemer and D.G. Dixon, 1992. "Copper Accumulation on Gills of Fathead
Minnows: Influence of Water Hardness, Complexation and pH on the Gill Micro-environment,"
Environmental Toxicology and Chemistry, 11:381-391.

Playle, R.C., D.G. Dixon and K. Bumison, 1 993a. "Copper and Cadmium Binding to Fish Gills:
Modification by Dissolved Organic Carbon and Synthetic Ligands," Can. J. Fish. Aquat. Sci., 50:  2667-
2677.
Playle, R.C., D.G. Dixon and K. Bumison, 1993b. "Copper and Cadmium Binding to Fish Gills:
Estimates of Metal-Gill Stability Constants and Modeling of Metal Accumulation," Can. J. Fish. Aquat.
Sci., 50: 2678-2687.
is?
Playle, R.C., 1998a. "Modelling Metal Interactions at Fish Gills," The Science of the Total Environment,
219, pp. 147-163.

Playle, R., 1998b.  Personal communication to Paul Paquin.

Prothro, M.G., 1993.  "Office of Water Policy and Technical Guidance on Interpretation and
Implementation of Aquatic Life Metals Criteria," Memorandum from USEPA Acting Assistant
Administrator for Water to Water Management Division Directors and Environmental Services Division
Directors, Regions I-X.

Renner, R., 1997.  "Rethinking Water Quality Standards for Metals Toxicity," Environmental Science
and Technology, 31:466-468.
                                            3-86

-------
I
Santore, R.C. and C.T. Driscoll, 1995. "The CHESS Model for Calculating Chemical Equilibria in Soils
and Solutions," Chemical Equilibrium and Reaction Models, SSSA Special Publication 42, The Soil
Society of America, American Society of Agronomy.

Santore, R.S., D.M. Di Toro, P.R. Paquin, 1999. "A Biotic Ligand Model of the Acute Toxicity of
Metals. II. Application to Fish and Daphnia Exposure to Copper," manuscript in preparation.

Sikora, F.J. and F.J. Stevenson, 1988. "Silver Complexation by Humic Substances: Complexation
Constants and Nature of Reactive Sites," Geoderma, 42:353-363.

Tipping, E., 1994.  "WHAM--A Chemical Equilibrium Model and Computer Code for Waters,
Sediments, and Soils Incorporating a Discrete Site/Electrostatic Model of Ion-Binding by Humic
Substances," Computers and Geosciences, Volume 20, No. 6, pp.  973-1023.

Tipping, E., 1996.   Modeling the Interactions of Metals with Dissolved Organic Matter in Surface
Waters," In Press.

USEPA, October 1980,. "Ambient Water Quality Criteria for Silver," Office of Water Regulations and
Standards, Criteria and Standards Division, Washington, DC. EPA 440/5-80-071; PB81-117822.

USEPA, January 1985.  "Ambient Water Quality Criteria for Copper -1984," Office of Water
Regulations and Standards, Criteria and Standards Division, Washington, DC.
    \
USEPA, September 24,1987. "Ambient Aquatic Life Water Quality Criteria for Silver." - Draft.

USEPA, August 1994. "Water Quality Standards Handbook: Second Edition, Appendix L, Interim
Guidance on Determination and Use of Water Effect Ratios for Metals," USEPA, Office of Water, EPA-
823-B-94-005a.

Varshal, G.M., T.K. Velyukhanova, N.N. Baranova, I.Y. Koshcheyeva, T.V. Shumskaya, Y.V. Kholin,
1995. "Geochemical Significance of Complexing of Silver(I) with Humus Acids," Geochemistry
                                                    3-87

-------
International, 32:1-9.

Westall, J.C., J.L. Zachary and F. Morel, 1976. MINEQL, A Computer Program for the Calculation of
Chemical Equilibrium Composition of Aqueous Systems. Technical Note 18. Department of Civil Engr.
MIT. Cambridge MA.
Wood, C.M., C. Hogstrand, F. Galvez, and R.S. Munger, 1996a. "The Physiology of Waterborne Silver
Toxicity in Freshwater Rainbow Trout (Oncorhynchus mykiss) 1. The Effects of Ionic AgY Aquatic
Toxicology, Vol. 35, pp. 93-109.

Wood, C.M., C. Hogstrand, F. Galvez, and R.S. Munger, 1996b. "The Physiology of Waterborne Silver
Toxicity in Freshwater Rainbow Trout (Oncorhynchus mykiss) 2. The Effects of Silver Thiosulfate,"
Aquatic Toxicology, Vol. 35, pp. 111 -125.

Wood, C.M., R.C. Playle and C. Hogstrand, 1999.  "Physiology and Modeling of the Mechanisms of
Silver Uptake and Toxicity in Fish," Environmental Toxicology and Chemistry, 16:1, 71-83.
                                            3-i

-------
 t
            Table 1.  Biotic Ligand Model Parameter Values for Silver
i
t
                                                     3-89

-------
                                       LIST OF FIGURES

Figure 1       Schematic of Biotic Ligand Model for Silver

Figure 2.       Probability distribution of Genus Mean Acute Values (GMAVs) used to establish the
               1987 draft water quality criteria for silver (left side) and probability distribution of
               measurements of freshwater silver concentrations downstream of POTWs (right side).

Figure 3.       Gill silver accumulation on the gills of 1 to 3 gram rainbow trout exposed for 2 to 3
               hours to 0.07 uM Ag and varying levels of thiosulfate (data from Janes and Playle,
               1995).

Figure 4       Calibration of WHAM to Ag-DOC complexation data of Sikora and Stevenson (1988).

Figure 5       Comparison of data to biotic ligand model results for gill silver accumulation varying
               with silver concentration, sodium, chloride and DOC (data from Janes and Playle, 1995).

Figure 6       Comparison of data "(Janes and Playle, 1995) to biotic ligand model results for gill silver
               accumulation resulting from exposure to silver in control waters (1 and 2) and 6 natural
               waters (3-8) (data from Janes and Playle, 1995).

Figure 7 .      Effect of calcium on 96-hour silver LC50 for fathead minnows and rainbow trout,
               calculated free silver concentration, and calculated gill silver concentration (data from
               Buryetal., 1999b).

Figure 8       Effect of DOC on 96-hour silver LC50 for fathead minnows and rainbow trout,
               calculated free silver concentration, and calculated gill silver concentration (data from
               Buryetal., 1999b).

Figure 9       Effect of chloride on 96-hour silver LC50 for fathead minnows and rainbow trout,
             ,  calculated free silver concentration, and calculated gill silver concentration (data from
               Buryetal., 1999b).
Figure 10 .      DOC levels in treatments at 4 chloride levels, variation in 96-hour silver LC50 for
               fathead minnows by treatment, and calculated gill silver concentration, (data from Bills
               et al, 1997).               '

Figure 11       Biotic ligand model predictions of silver LC50 for fathead minnows and rainbow trout
               versus measured silver LC50. Biotic ligand LA50 = 17 nmol/gw. AgCl binds to gills but
               is only toxic to fathead minnows.  Diagonal solid line is line of perfect agreement and
               dashed lines are within a factor of +/- 2 of solid line.

Figure 12      Biotic ligand model predictions.of silver LC50 for fathead minnows, rainbow trout and
               Daphnia magnet versus measured silver LC50. Biotic ligand LA50 = 17 nmol/g,, for fish
               and 2.3 nmol/gw for D. magna. AgCl binds to gills but is only toxic to fathead minnows.
               Diagonal solid line is line of perfect agreement and dashed lines are within a factor of
               +/-2 of solid line.
                                             3-90

-------
  I
i
    LL

    05,
    -100
    UJO)
    Qr-



    §1
         D)
I
      O
      U.LU

      0>




        55
      
                                                                       C3
                                      3-91

-------
                                       I
                                         i
                                     CM

                                     LU
                                     DC
                                       •
(i/Bri) AVWO
ainov
   3-92

-------
  t
'••„->
I
                   10.
                    8l
 o>

 05

I
 o
I    4,

I
<5    ;
                        Legend:
                         Q Control Un!t
                         H Means with 95% C.I.



                                              I

                                                         t
                          0.00  0.00  0.05  0.10  0.50  1.00   2.00
                                        Thiosulfate (uM)
   i
                                           FIGURE  3
 t
                                            3-93

-------















•















^


«• 1
o. p
C (0
8 S
O •g
1. 1
T- OQ
£i
* «
D CO
3 d




1
I
\
\
^ .
I !
*
\
\
t
\
1 \ \
\
• \ v
* 1
• \ \
\
\
m \ ^ -
\
. \ ^
\
\
\
* \ \
- f
w . •', V .

|1 i\ \
o — •*« V
1 | ^ \
LL. ^^ * A
S S * *

i 1 *• \
g g i> v *
i g *\« v*
S o \ B \«
£D £p **..« \*

KtO,tO ^f CO CM i- C
- d d ;• d d d . d d


,
- h-




- CD


~<




m
Cf . i
UJ
t~
X ^f
l_
*• lit
^1 CC ^
i <£ 1
w "- |
| !
UL |
- CO i


i
i
1
|
•
CM






I
'

(6/|ouJiu) BV punog
           3-94

-------
                                                                 •o
                                                                 0>

                                                                 to
                                                                 a
                                                                 O

                                                                 I

                                                                 oa
                                                                  %
                                                                 O>
                             o    in    o    in
                             CM    r-.    T-
                                                                 •o

                                                                 I

                                                                 o
                                                                 "3
                                                                 O
                             to
                              %
                             o
                             O)
                             2

                             a
                                   CA
                                   4^

                                   C
                                   0)
                                                                        a
                                                                        x
                                                                       Ul
      O

      2
      o
     -Q Ifl
      CO O)
     —I O)
      I  «-

     s^
      I  fr
6/6v joiuu) 6v-|l!9
B/BV |ouju)
     h-

      5
                                                                              111
                                                                              cc

                                                                 o
                                                                 in
                                                                 CD
                                                                  w

                                                                  i
                                                                 T3
                                                                  CD
                                                                  O
                                   xi  «>  a
                                    C  g  U.

                                   '5 jg
                                   cc "»
                                       * *
                                   *»-  (0
                                    o  15
                                    CO O

                                                                  s .    >
                                    O)
                                    c

                                   TJ


                                   5
                                    1-
                                    0)
 B/BV
                                                BV-IHD
                                                                  o

                                                                 a
0)
D)
a>
                                  3-95

-------
W
0)
J^
CO
•^
To
*••
^^
CO
c
— to
4-» 0>
3 0)
O r-
HCD
•2S
•* >*
5 .2
0 0-
-Q T3
^ §
£ 2
DC o
•sl
J2 «J
= «
0 Q
0 .
+*
^ •
O)
c
s
c •
S
0)
>
^^^ *
w .8
,
i
. i , , , . r 	 (
	 < ' ' ' ' i ' ' i i 1 i i i i

» — r
»Si
'
i — i
_ • t — a 	 i
• ^— i


i
i ^ 	 1

t-0H(
-—• • '
o
vP » 	 * 	 i
0X ^ '•
to
0)
•tk t__ . i
• 2 '" 1
> CO K*H '
C4 fl> "-W-«
c o>
co C'
fl) ID
55 ^ OC
K ^-»
••" V I 	 1
^ C *IP^ I I
•gal -•-
S IJ
CO £ 0
•fl% • ^"*
1,33. ...
^•uO^i — i
j— ' ' 	 i ••• i i .... i ....
* • 8 S g °
-i 1
i
• °o
• i
.'
j
rs
i
ii
i;
|i
Mf± El
Hj Jl
1
I
v. ii
0)
.» |
£ S
3 13 /
- I Si
iS
i
CO



CM !


f '


                                         I
                                         I
6/joiuu) 6v-|l!0
  3-96

-------
I
IS
8
                 30

             _  25

           •0*5)  20
           §3
              o>  15
              3
              co  1P

                   5

                   0


                 20
          Is  15
          (0 D)
          "53
          £ m  10
           ' ^—
            o
          O
                   - I—I Fathead Minnows
                   • E23 Rainbow Trout
   (D
52   5
*"-

       0

      50

-o I  40


If  30

OS  20

S5*
02  1°
   5
                           50
                         250
800
                                         EJ
                           50
                         250
800
                    r~l Fathead Minnows: GilI:AgCI + GilhAg
                    E3 Rainbow Trout: GilhAg only
                           50
                         250      800

                         Chloride (uM)
1500
1500
         1500
                      Effect of Chloride on Silver 96-hour LC50
                       Data:  Bury, Galvez and Wood,  1998
                                      FIGURE 7
                                    3-97

-------
       30

   .-.  25
   HT
 •o"o>  20


 IS  ^
 CD —
   |2
        5

        0

       20
  '—   1R
       15
          - O Fathead Minnows
           £32 Rainbow Trout
                    0.3
 «o en
"5.5
0


-------

I
              30


         '—   25

       •o^   20
        £3
        go)   15


       i"5   10

         H    5

               0


              20



       I?   15
       «"o>
       "33
       -2m   10
        O
        OQ
           2    5
           u.
       0


      50

  4^
T, I  40


ff  3°

3s  20

§5*
->*?  10
Mi
  S
                      Fathead Minnows
                      Rainbow Trout
                           50
500
2000
                           50
500
2000
                      Fathead Minnows: 6!l!:AgCI + Gfl!:Ag
                   VTA Rainbow Trout: GIllrAg  only
t

I
                           50          500       2000

                                  Calcium (uM)


                    Effect of Calcium on Silver 96-hour LC50
                      Data: Bury, Galvez and Wood,  1998

                                     FIGURE  9

                                     3-99

-------
^
^
             £8
in
                            I
                      I	I  I
                                    	I	•  I  i  |"-r
                                   o
                                 S B
                                 C9 O
                                  i i
                                                  1
                                     i  i  i  . t , i
                                                      "5>


                                                      o
                                                      CD
                                                      "8>
                                                      E
                                                     O)


                                                    O
                                                      O)
                                                    o
                                                    s
                                                              (0

                                                              o
                                                              c
                                                              c
                                                            •o
                                                             CO  _
                                                             0)  fN
                                                            J=  O>
                                                              >. if
                                                             •g  s
                                                              '  2
                                                             co  .a
                                                            s  I
                                                            O -2
                                                             CD
                                                                .g

                                                             C o
                                                             O  CD
                                                             o> o
                                                             "O  w
                                                           £  O)
                                                           O  E

                                                           18
                                                            CO
                                                           O
                                                           H-
                                                            o
                                                                      LLJ
                                                                      CC

000
           (1/Bn) 0901
               pajnsee|/\|
                                   B/JOUJU) Q901 Ii!D
                          3-100

-------
  I
for
           O
           LO

           O
            0)



           55

           •o
            0)
           .Q

           O
            0)


           •o
            0)

           ts
           -2
           a.
                 JIM  \. I—i	1	j i I  i  i i   i—i	r
a
   CO
      s
   1  1
   •n
   to  o
                    LL.  QC


                    •  D
                                                  o
                                                  o
                                                      0)

                                                      3

                                                      o
                                                      LO

                                                      O
Ul
QC
                                                                       i   c
                                                                       CO
                                                                       CO

                                                                       o
               o
               o
                                  0901 PW!P«»d iAHS
                                        3-101

-------
o
in
O
 o>
-o
 0>
 o
 V)
 CO

 I
 Q)
 O
 ff
£
         (1 M  I	1	 INI II  I  I   1     1111 I I  I  l	1
JC  C  Q.

^  « 5
£ cc Q
                                                   o
                                                   o
                                                                      t
                                                         _  o
                                                       o
                                                       ID
                                                       O


                                                       •o
                                                       0)

                                                       3
                                                       w
                                                       ro
li -l
1
<3
Uli
               0901
                                          IAI18
                               3^102

-------
I
      Surface Water Assessment Presentation Materials

-------
 -
I
I
j

-------
                                             DRAFT
  I
Site Specific Water Quality Criteria for Metals

                Herbert E. Allen
 Department of Civil and Environmental Engineering
              University of Delaware
               Newark, DE 19716
                Water Quality Criteria (WQC) are derived from bioassay tests in which the toxicant is
•-                added to the water. Because organisms have different sensitivities to toxicants, the LC50
'   .             values (lethal concentration to 50% of the organisms exposed) will vary greatly as shown
                in Slide 1 for the freshwater criteria. Saltwater criteria are similarly derived.  Criteria are
                developed for protection of 95% of the species which is achieved by statistical
                calculation.  This provides the Final Acute Value (FAV). The freshwater criteria are
                functions of the hardness, which is acknowledged to affect the toxicity. However, in
"I              addition to hardness a number of other water quality parameters, including solids, pH,
r               dissolved organic matter (DOM), and inorganic ligands (Slide 2) also affect the toxicity
                of metals.  Presently, WQC are implemented on a dissolved basis. Thus, partitioning of
!'                metals to suspended solids need not be considered.

                Site-specific variation of water quality parameters from those present in the waters in
|                which the bioassays used to develop the criteria often results in the criteria not being
&              predictive of the actual effects. This is due to the effects of water chemistry on the metal
•                speciation which affect the bioavailability of the metal.  The Water Effect Ratio (WER),
                in which the ratio of toxicity test results for a site water compared to that for a reference
                water are used to correct the criteria, has been used as a means to account for the effects
                of metal speciation on a site specific basis (Slide 3); Water Effect Ratios have been
|:J              determined for a number of metals (including Cd, Cr, Cu, Pb, Ni, Ag, and Zn) and for
                other constituents, such as ammonia.  As shown in Slide 4, at many sites WERs are very
                large, even exceeding a value of 10.  This means that the WQC are over protective
                relative to the intended level of protection. At some sites values less than one have been
"                reported for WERs indicating that the WQC are under protective. The value of the WER
                is not a constant for a given water, but depends on the organism tested.  More senitive
-;,              .organisms (lower LC50 values in Slide 1) give higher WERs (Slide 4).

.-,..              The principles involved in the WER are shown in Slide 5 for the acute toxicity of copper
                in a laboratory reference and a site water.  In the bioassay test various concentrations of
                copper are added to both the reference and the site waters (x-axis). Because these two
j^               waters have different concentrations of DOM present, the metal speciation (y-axis)
                differs between the two waters.  If the pH, hardness and other water quality parameters
                are kept constant, the values on the y-axis are a good predictor of bioavailability. The
                importance of this can be followed for the three organisms shown. The WQC for
                polychaetes is 120 ^,g/L. I have assumed that toxicity in the reference water is also 120
                Hg/L. Following the solid arrow from this value on the x-axis to the y-axis gives the
                equivalent bioavailable copper concentration in the reference water.  Clearly,  for the
                same biological effect to occur in the site water, the same concentration of bioavailable
                                                      4-2

-------
 metal must be present irrespective of the total metal concentration. This is shown by the
 dashed arrow which extends from the y-axis to the solid line for the site water and thence
 to the x-axis. The value intercepted on the x-axis is the toxicity for copper in the site
 water.  The ratio of the x-axis values for the site water to the reference water is the WER

 Inspection of Slide 5 indicates that the WER will not be the same for the 3 organisms
 considered. This is more clearly shown in Slide 6 which presents the same data in bar
 graph form for these three organisms plus two that are more sensitive. Lines for the more
 sensitive organisms cannot be easily plotted in Slide 5. The sensitivity of the organisms
 increases in Slide 6 from left (mysid) to right (mussel) as indicated in smaller
 concentrations of copper for the LC50 values in both the reference water and the site
 water. However, the non-linearity of the curves in Slide 5 result in inconsistent WERs
 between the five species considered in Slide 6.  The WER values dramatically increase
 for the more sensitive organisms.

From these considerations of chemical speciation, it is clear that development of WQC
based on bioavailable metal would obviate the need for determining a WER.  The biotic
iigand model shown in Slide 7, which will be discussed in the next presentation, couples
the environmental chemistry of metals to the response of the organism.
                                      4-3

-------
I
                         Copper - Fresh Water Species
I
             0)
I
cti
o>
             CO
             0
                100000
                 10000
                  1000
100
                    10
         Adjusted to Hardness = 50 mg/L

      Acute:   1 -hr C < 9.23 [H/50]0'9422
      Chronic: 4-d C < 6.54 [H/50]0'8545    O
      Exceedance Frequency
        < once in 3 years
      1. Northern Squawfish
      2. Daphnia
      3. Ceriodaphnia
- 4. Garnmarus
                 -  FAV=18.48^g/L
              	CMC = FAV/2 = 9.23 jig/L
                     0.1   1      10   30  50 70   90
                                      Probability
                                      99  99.9
           HEA Slide 1
                                          4-4

-------
          Water Quality Parameters
           Affecting Metal Toxicity
  Hardness (Ca and Mg)
• Solids
  pH

  Dissolved Organic Matter (DOM)

  Inorganic Ligands
HEA Slide 2
                      4-5

-------
 I
             Water Effect Ratio


       Site-Specific WQC = NWQC x WER

               = NWQC x   site-water LC5Q
                         reference-water LC50
m
I
~-;-

t".
       HEA Slide 3
 I
                       4-6

-------
                                Water Effects Ratios
          LOCATION
                               METAL
                                                SPECIES
 St. Louis R., MN
 Nemadji R., WI
 Little Pokegama R., WI
 Selser's Cr., LA
 Cuyahoga R., OH
 Lehigh River, PA (4 months)
                                 Cadmium
                                 Cadmium
                                 Cadmium
                                 Cadmium
                                 Cadmium
                                 Cadmium
                                                                            Es
                                                                      Most to Least
                                                                         Sensitive
                                                                   11,4.4, 5.0,4.3, 0.8
                                                                   3.6, 2.8
                                                                   1.8, 1.0
                                                                   3.0
                                                                   1.0, 1.7
    Boggy Cr., OK
    Leon Cr., TX
Naugatuck R., CT
St. Louis R., MN
Nemadji R.,WI
Little Pokegama R., WI
Quinnipiac River, CT •
Columbia R.WA
Lehigh River, PA (4 months)
Wissahickon Creek, PA
                                 Copper
                                 Copper
                                 Copper
                                 Copper
                                 Copper
                                 Copper
                                 Copper

                                !^E*S£?(iKa
                                          A
                                          A
                                          D-P,
                                          c.d.
1.0, -I.I
15, 9.2, 6.3, 3.3
4.3
3.1
2.5 - 36.6 (18 values)
1.2 - 3.8 (5 values)
3.5 -17.1
1.6 - 7.6 (8 values)
   St. Louis R., MN
   Norwalk River, CT
   Lehigh River, PA (4 months
                                          A, Bin., Q.n
                                          &m,j.Q
                                          c.d., E.D.
                                                                  .l', 3.7
                                                                 1.0-25.6
   Selser's Cr., LA
   LehihRier,PA (4 months
  orwalk R., CT
 Boggy Cr., OK
 St. Louis R., MN
 Naugatuck R., CT
 Cuyahoga R., OH
 Lehigh River, PA (4 months)
                                                                   2.2, 1.5
                                                                   1.2
                                                                   2.9, 1.1, 0.8, 0.7
                                                                   0.9, 0.7
                                                                   0.8
                                                                   1.0-7.6
A=Amphipod; C=Caddisfly; C.d.=Ceriodaphnia dubia: D.ra=Daphnia magsa; JD p =Daphnia pulex-
am=0nchorhynchus myjdss; P.E.=Pimephales promelas; P.sp.=Palaemonftrkfi sp.'; sTKH^^al
serrulatus; S.t.=SalmQ trutta                                           "^      	*—
                                                                                     us
HEASlide4
                                            4-7

-------
  I
Effect of Metal Complexation on Toxicity

  t
&>.-
Vi-
     Total Copper (mol/L)



CO    l">  ' O    ^"
CD    CM   O>    1C



      Total Copper (|u.g/L)
                                                    Is-    T-
                                                    >-    CO
                                                    co    co
                                               cq

                                               ^
                                               Tj-
-*

00
o
m
            bashed line is Reference Water and solid line is Site Water. Lines with arrows show

            relationship between total metal and the bioavailable metal for the LC50 values for three

            marine species.
            HEA Slide 5
                                           4-8

-------
      Effect of Species Sensitivity on WER
          500
                          D Reference Water
                          E Site Water
                            wm
          niiM»«i iNMnnnw JNT*»«*  NNimaa r-^yp»^ |  o

              Mysid  Polychaete Copepfod   Oyster   Mussel
HEA Slide 6
                            4-9

-------
I
                  AQUATIC BIOTIC LIGAND MODEL
                      CONCEPTUAL DIAGRAM
                        (after Pagenkopf, 1983)
                   Complexes
                              j     Inorganic
                              1   Complexes
                       e.g. Cu - Hydroxides
                          Cu - Carbonates
                                                    Gill Surface
                                                    (biotic ligand)
	| Metal Binding
  I     Site
         HEASlide?
                   Tipping, 1994 (WHAM)     Playle, 1993
                                 4-10

-------
               DRAFT

Biotic-ligand Model:  Future Directions
           Other freshwater species
           Other metals
           Marine species
           Chronic toxicity
           Explanatory power
               Joseph S. Meyer
       Department of Zoology and Physiology
             University of Wyoming
              Laramie,WY 82071
                 Presented to:

    Ecological Processes and Effects Committee
            Science Advisory Board
       U.S. Environmental Protection Agency
               Washington, DC

            *    7 April 1999
                    4-11

-------
M
                CO O
&
o
0)
0)
                      3 *>
                      w"S
                      .2>3
                      h-XI
        0)
        (/)
        0)


        o
        Q.

        Q)
        C
       CD
                                             0)
                                             c
                                             re
                                             I
  osxn Jo oson
s g
Is
I
    
-------
        FRESHWATER SPECI
•  Rainbow trout (Oncorhynchus mykiss)



•  Fathead minnow (Pimephales promotes)



•  Tubificid worm (Lumbriculus variegatus)



•  Amphipod  (Hyallela azteca)



•  Daphnid (Daphnia magna)



•  Algae
                                       i
                 4-13

-------
           ID
           Q.
                                  - co
                                  - ^r
- CM
0-
0



(ft

d>
c
"Q

(5
X
1
                            3 33
                            tu
                                      C
                                      •g

                                      (5
                  no |0iurl)  M-9OSV1
             JO (no
                     4-14

-------
      •s
      O
 0)
 Q,
 Q,
 O
O

 O
 0)
 O
 Q.
 X
W

 (Q
 O
 Q>
5
0
      D
     O
r
CM


O
      ^^


     O
                                             •
                                            - II
                                            pr
                                            o>
                                       - CM
                         U) ID IO
                          •  m  m
                         (0 h. 00

                         XXX
                         Q. Q. Q,
                         fit
                           I 7
                                       - CM
         (Mp kB no loiurl)  M"9OSV1
           JO (no |A|rl)  QSOl M-
                     4-15

-------
frf
yi
&~t


K:"l
&
V '
Ag*   Playle,
      McMaster U.

Cd   Campbell,
      Playle,
      McMaster U.

Co   Playle,

Cu*   Borgmann,
      Playle,
      McMaster U.,
      U. Delaware,
      U. Wyoming
          *
Hg   Borgmann
Ni*
                fish toxicity, gill binding &
                physiology

                fish toxicity & gill binding;
                algal toxicity and uptake
                gill binding

                fish toxicity, gill binding &
                physiology;
                invertebrate toxicity & uptake
                invertebrate toxicity & uptake
U. Wyoming    fish toxicity & gill binding
Pb    Borgmann

Tl    Borgmann

Zn*   Allen,
      Borgmann,
      Campbell,
      McMaster U.
                invertebrate toxicity & uptake

                invertebrate toxicity & uptake

                fish toxicity, gill binding &
                physiology;
                invertebrate toxicity & uptake
                algal toxicity & uptake
 = metals directly tested for biotic-ligand model
                        4-16

-------
Fathe
Minnows Exposed t

(Pimephales promelas)
      2-,

 10 -

 Oo
 —I U)
         pH7.5
               24       6


           Hardness (meq L~1)
               4-17

-------
m
           (0
          O
           O
          X
          LU
               03

•o
O
                     O
                                                     ID
                                                     T™
                                                     CO

                           8    8    8
                         8
                                                     8
                           8    8    83
g.S
                                                           tlt§
                                                               a-.fr
          o
          (O
          o
          T-

          10
          ID
          O
                                               00
                                                                     o
                                                                     CO
                                               to
                                               o
                                                •V

                                     *§«:*<   S
                                     ^a-o«   o>
                                                                     CTJ
                                               (0

                                               ff
                                               o
                                               CO

                                               o
                                       4-18

-------
  Rainbow Trout Exposed to Zinc
        (Oncorhynchus my kiss)
2200-1

2000-

1800-

1600-
                   96 h LC
                       50
                              I
        Control   2mMNa*  1mMMg2+  1mMCa2*
      [from Alsop and Wood 1999, manuscript]
                 4-19

-------
I             MARINE SPECIES
0
        Tidepool sculpin (Oligocottus maculosus)



        European eel (Anguilla anguilla)*



        Grass shrimp (Palaemonetes pugio)
        Algae
                      4-20

-------
                SI
               f  I
                        N Z N
                        K E 09
       • O9o
     SeSSSo S
        ||
      looj
                                «  c
       top/suoiswput (if) ffoj (jfrnojS syptds


                  .4-21
                                                        I
                                      0>
                                      CM
                                      1O
                                                iO
o
a:

CO
                                                CD
                                                K
                                                0>
                                             i
                                                O
                                                •d
                                                c
                                                10
                                                10
                                                •o
                                                c
                                                3
                                                (0


                                                I
                                             t

-------
i

.  ,

W
                                         IVAIAini 1N33M4

                                                    2£o
                                                    XXK
                            •o
                            O
                                                                                             t
                                                                                             o
                                                                                             * •

                                                                                            CM
                                                                                             O
 o
(0
  •


 I

 >

HI
  «t
oo
h*
O>
 CO
*»
 0)

 (0
T3
 C


CO



 p
                                     e
                                     S
                                                    4-22

-------
        O
               J
               .1
                t"
             JJ 4 • » * *
                                g

 CO

 t
 O>

 §
 • m
 CM
 O
 C

 o

12
"o
CO
                  1VAIA«nS
0)
       TO
                    n ^
                                       J
                                  i\
 >

III

ccT
h-
O)
T-
  •

"i

3
c
3
CO


o
             •s «•»••«

             5  lilii
              . 4 • » • •
             S
                         4-23

-------

 S2
il
 o>
mc

 <5
73
 C
 (0

 o>

u.

 c
 0)


 Q>


 O
i..
 (0
•o
 c
 CO
 c
 o
slfffl-

sipllu
Ililtfi
     cS =
     ••* JS

LU
UJ %




I

ill
                                           S


                                           I
                r
                c:
                <
                                    (fl
                                    (0

                                    £
                                    Q.
                                    0)
                                    o>
(0
o
                             1
                             o
                             i/t
                                    O
                                                    a.

                                                    "5
                                    o

                                    o>

                                    c
                                    4)
                                    (0
                                    "5

                                    z
                                    o
                                    (0


                                    O
                         4-24

-------
 in
 oo
 o>
 CL
 UJ
 CO

i i
9 0
•••*.
«
o *-
>
3
1
-3
o
 O)


 
-------
  t
                       Q.
                       .a
                       C
sg
                                                                          in
                                                                          oo
                                                   c
                                                   c
                                                                         1
                                                                         5
                                                                          
-------
  Why is the slope of In(LC50) vs.
  In(hardness) approximately 1.0?
   10000 -,
<—* 1000-
 3
o
 D)
ID
o
100^
     10-
       1
          slope
                       1 i i

                    1000  10000
                     -1
   Hardness (mg L"1 as CaCOJ
               4-27

-------
  A                  Winderinere Humic Aqueous Model (WHAM)




                                 Presentation Materials
T$§*~ ^^^




i w
i.
   I
                                         4-28

-------
 Humic
   Acid
 Stevenson, F. J. (1982).
 Humus Chemistry.
                              H,CO
                                         OOH
                                                 Aliphatic
                                                dicarboxyiic
                                                  acid
       Aromatic
      dicarboxyiic
         8Cid
                                             Aliphatic
                                               acid
                               Aromatic
                                acid
Thurman, E. M. (1985).
Organic geochemistry of natural waters.
                           4-29

-------
it' O
             WHAM REACTIONS
       Proton Binding

       (A) Carboxylic groups: COOK

       (B) Phenolic groups:  /"\-QH
             RA-H
             RB-H
RA
RB
H+
H+
        Mass Action for 1 site
                             exp{2wZ}
         Electrostatic Effect
          w =
                     4-30

-------
                                                      1 .
                                                    8 •"•>
            Distribution of Proton Binding Sites
A-types sites: fy to K4
     = Site Density
       = Midpoint pK
        = Spread in pK
                               B-types sites:  K4toK8
                               = Site Density
                            pKg = Midpoint pK


                            ApKg = Spread in pK
   1.5
^)


CD
c
0) _ _
Q 0.5 -
V)
    0-
             COOH
                        phenolic-OH
     0.00
K1 K2 K3 K4
. . . .nA/4
^ 	 - - fr
ApKA
K5
nA/8 .
P«A
	 -.., —A.
K6 K7 K8
ApKB
pKB
A. 	 	 -.„.
                5.00        10.00        15.00

                       pK
                        4-31

-------
  t
                    Proton Binding Calibration Results
m
                   nA     pKA    pKB   ApKA  ApKB log10(-P)
                                   4-32

-------
                                            tf* <•-
         Metal Binding Reactions
A and B Sites

       RA-H +  M2+

       RB-H.+  M2+
R-M+
       RB-M
                 H
 pK's for A and B Sites: KM1 to KM8
       KMi = KMHA :/ Ki   i = 1 ..... 4
       KMi = KMHB/Ki  i = 5 ..... 8
       PKMHB = 1 -38 PKMHA + 2-57
Bidentate
  -RA-H

    RB-H
             M
2+
       \
                               HP
                                 n
                                 %•*
M2+  + 2 H*
   KM12 = (KMA KMfi)
   Sites (1,2),(1,4),(1,6),(1,8),(2,3),(2,5)1
        (3,4),(3,6),(3,8),(4,5),(4,7)
                               i
                   4-33

-------
I
9
          pN
2.5

3.0

3.5


4.0

4.5

5.0
          pN
 2.5

 3.0

 3.5


 4.0

 4.5

 5.0
                11
          pN
 2.5


 3.0

 3.5


 4.0


 4.5


 5.0
            WHAM CALIBRATION
  10    8     6     4
         p[Cu]
        9753

            PlCu]
         654    3
            p[Cd]
pN
                                            2.5 r
                                            3.5
                                            415
                                            5.5
                                              10    8     6    4
                                                    PlCu]
 PN4
     10    8      6     4
             PlCu]
                                            2.5
                                         pN 3.5
   4.5
                                                6.0      5.0      4.0
                                                       p[Pb]
                                     4-34

-------
                 Binding Site Occupancy



Fraction
OJ
0)
2
•o
c
1




1.0
0.8

0.6
0.4
0.2
Q 0
pM = 9
•
K
I

m

0.5
0.4
0.3
Q.2
0.1
on
pM = 5



M^
i —
                                        pM =
                                    .EL
12345678  123456789101112 D
  MONO-       BIDENTATE
 DENTATE
12345678  123456789101112 D
  MONO-       BIDENTATE
 DENTATE
                         4-35

-------
                              Data Fitting  Results.
                                      pKMA
                          Ca
&r
t
               Table 7. Tipping, E. and M, Hurley (1992).
               Geochim. Cosmochlm. Acto 56: 3627-3641.
                                       4-36

-------
 ft
  i
I
 ii
II

-------
Integrated Approach to Assessing the Bioavailability
    and Toxicity of Metals in Surface Waters and
                     Sediments
                    Appendices
      Presented to the EPA Science Advisory Board
                     April 6-7,1999

            U.S. Environmental Protection Agency
                     Office of Water
            Office of Research and Development
                   Washington, D.C.

-------

-------
APPENDIX A

-------
i

-------
        Research Papers on the Bioavailability and Toxicity of Metals in Sediments'

Kev Papers (Roughly in order of importance)

Ankley, G.T., D.M. Di Toro, DJ. Hansen, J.D. Mahoney, WJ. Berry, R.C. Swartz, R.A. Hoke, N A
Thomas, A.W. Garrison, H.E. Allen and C.S. Zarba. 1994. Assessing potential bioavailability of
metals in sediments: A proposed approach. Environ. Management.  18:331-337.

Ankley, G.T., D.M. Di Toro, D J. Hansen and W J. Berry.  1996. Technical basis and proposal for
deriving sediment quality criteria for metals. Environ. Toxicol. Chem.  15:2056-2066.

Di Toro, D.M., J.D. Mahoney, DJ. Hansen, K.J. Scott, M.B. Hicks; S.M. Mayr and M.S. Redmond.
1990. Toxicity of cadmium in sediments: The role of acid volatile sulfide. Environ. Toxicol. Chem.
9:1487-1502.                                                                     '

Hansen, D.J., W.J. Berry, J.D. Mahoney, W. S. Boothman, D.M. Di Toro, D.L. Robson, G.T. Ankley,
D. Ma, Q. Yan and C.E. Pesch. 1996. Predicting the toxicity of metal-contaminated field sediments
using interstitial concentration of metals and acid-volatile sulfide normalizations. Environ. Toxicol
Chem.  15:2080-2094.

Berry, WJ., DJ. Hansen, J.D. Mahoney, D.L. Robson, D.M. Di Toro, B.P. Shipley, B. Rogers, J.M.
Corbin and W.S. Boothman. 1996. Predicting the toxicity of metal-spiked laboratory sediments
                                                            \
      \                                      .                               •        '
using acid-volatile sulfide and interstitial water normalizations. Environ. Toxicol. Chem.  15:2067-
2079.

Call, D J., C.N. Polkinghorne, T.P. Markee, L.T. Brooke, D.L. Geiger, J.W. Gorsuch and K.A.
Robillard.  1999. Silver toxicity to Chironomus tentans  in two freshwater sediments. Environ.
Toxicol. Chem. 18:30-39.

Berry, W.J., M.G. Cantwell, P. A. Edwards, J.R. Serbst and D J. Hansen. 1999. Predicting toxicity
of sediments spiked with silver. Environ.  Toxicol. Chem.  18:40-48.

Di Toro, D.M., J.D. Mahoney, D J. Hansen, KJ. Scott, A.R. Carlson and G.T. Ankley. 1992. Acid
Volatile Sulfide predicts the acute toxicity of cadmium and nickel in sediments.  Envrion. Sci.
Technol 26:96-101.
Other Relevant Papers

Ankley, G.T. 1996. Evaluation ofmetal/acid-volatile sulfide relationships in the prediction of metal
bioaccumulation by benthic macroinvertebrates. Environ. Toxicol. Chem.  15:2138-2146.
       'Articles listed as "Key Papers" are reprinted in this volume.  Additional relevant papers
have been listed, but not reprinted.

-------
 DeWitt, T.H., R.C Swartz, DJ. Hansen, D. McGovern and WJ. Berry. 1996.  Bioavailability and
. chronic toxicity of cadmium in sediment to the estuarine amphipod Leptocheirus plumulosus.
 Environ. Toxicbl. Chem.  15:2095-2101.

 Di Toro, D.M., J.D. Mahoney and A.M. Gonzalez. 1996. Particle oxidation model of synthetic FeS
 and sediment acid-volatile sulfide. Environ.  Toxicol. Chem.  15:2156-2167.

 Di Toro, D.M., J.D. Mahoney, DJ. Hansen and WJ. Berry. 1996. A model of the oxidation of iron
 and cadmium sulfide in sediments. Environ. Toxicol. Chem. 15:2168-2186.

 Gonzalez, A.M.  1996., A laboratory-formulated sediment incorporating synthetic acid-volatile
 sulfide. Environ. Toxicol. Chem. 15:2209-2220.

 Hansen, D J., J:D. Mahoney, W J. Berry, S J. Benyi, J.M. Corbin, S.D. Pratt, D.M. Di Toro and M.B.
 Abel. 1996.  Chronic effect of cadmium in sediments on colonization by benthic marine organisms:
 An evaluation of the role of interstitial cadmium and acid-volatile sulfide in biological availability.
 Environ. Toxicol. Chem.  15:2126-2137.

 Hassan, S.M.rA.W. Garrison, H.E. Allen, D.M. Di Toro and G.T. Ankley.  1996. Estimation of
 partition coefficients for five trace metals in sandy sediments and application to sediment quality
 criteria. Environ. Toxicol. Chem. 15:2198-2208.

 Leonard, E.N., G.T. Ankley and R.H. Hoke. Evaluation of metals in marine and freshwater surficial
 sediments from the environmental monitoring and assessment program relative to proposed sediment
 quality criteria for metals. Environ. Toxicol Chem. 15:2221-2232.         '

 Liber, K., DJ. Call, T.P. Markee, K.L.  Schmude, M.D. Balcer, F.W. Whiteman and G.T. Ankley.
 1996.   Effects  of acid-volatile sulfide  on zinc  bioavailability and  toxicity to benthic
 macroinvertebrates: A spiked-sediment field experiment Environ. Toxicol. Chem. 15:2113-2125.
                          • /                                  '

 Mahoney, JJX, D.M. DiToro, A.M. Gonzalez, M. Curto, M. Dilg, L.D. De Rosa and L.A. Sparrow.
 1996. Partitioning of metals to sediment organic carbon. Environ. Toxicol. Chem. 15:2187-2197.

 Peterson, G.S., G.T. Ankley and E.N.  Leonard.  1996.  Effect of bioturbation on metal-sulfide
 oxidation in surficial freshwater sediments. Environ. Toxicol. Chem. 15:2147-2155.

.Sibley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard.  1996. Predicting chronic toxicity of
 sediments spiked with zinc: An evaluation of the acid-volatile sulfide model using a life-cycle test
 with the midge Chironomus tentans.  Environ. Toxicol. Chem.  15:2102-2112.

-------
Assessing  Potential  Bioavailability of Metals in
Sediments:  A Proposed Approach
GERALD T.ANKLEY*
NELSON A. THOMAS
US Environmental Protection Agency
6201 Congdon Blvd., Duluth, Minnesota 55804. USA


DOMINIC M.-DITORO
Hydroqual, Inc.
Mahwah, New Jersey 07430, USA


DAVID J.HANSEN
US Environmental Protection Agency
27 Tarzwell Dr., Narragansett. Rhode Island 02882, USA
                       i

JOHND.MAHONY
Chemistry Department, Manhattan College
Bronx, New York 10471, USA


WALTER J. BERRY
SAIC Corporation
27 Tarzwell Dr., Narragansett, Rhode Island 02882, USA


RICHARD C. SWARTZ
US Environmental Protection Agency
Hatfield Marine Science Center
Marine Science Drive, Newport, Oregon 97365, USA


ROBERT A. HOKE
SAIC Corporation
411 Hackensack Ave.,
Hackensack,  New Jersey 07601. USA
 A. WAYNE GARRISON
 US Environmental Protection Agency
^College Station Rd., Athens, Georgia 30605. USA

 HERBERT E. ALLEN
 Department of Civil Engineering, University of Delaware
 Newark, Delaware 19716, USA

 CHRISTOPHER S. ZARBA
 US Environmental Protection Agency
 401 M Street S.W., Washington, DC 20460, USA

 ABSTRACT / Due to anthropogenic inputs, elevated
 concentrations of metals frequently occur in aquatic
. sediments. In order to make defensible estimates of the
 potential risk of metals in sediments and/or develop sediment
 quality criteria for metals, it is essential to identify that fraction
 of the total metal in the sediments that is boavailabte.
 Studies with a variety of benthic invertebrates indicate that
 interstitial (pore) water concentrations of metals correspond
 very well with the bioavailability of metals in test sediments.
 Many factors may  influence pore water concentrations of
 metals; however, in anaerobic sediments a key phase
 controlling partitioning of several catkxiic metals (cadmium,
 nickel, lead, zinc, copper) into pore water is acid volatile
 sulfide (AVS). In this paper, we present an overview of the
 technical basis for predicting bioavailabiiity of cationic metals
 to benthic organisms based on pore water metal
 concentrations and metal-AVS relationships. Included are
 discussions of the advantages and limitations of metal
 bioavailability predictions based on these parameters,   s
 relative both to site-specific assessments and the .
 development of sediment quality criteria.
Introduction and Technical Background
   The occurrence of elevated concentrations of met-
als in  aquatic sediments is common, and resource
managers often must decide whether these concentra-
tions may result in adverse ecosystem impacts, and
KEY WORDS: Sediment; Metal; Bioavailability; Toxicity; Sediment
           quality criteria
•To whom correspondence should be addressed.
 whether remedial action is required. Unfortunately,
 this can be difficult because it is impossible to utilize
 total concentrations of metals in sediments to predict
 the occurrence of environmental impacts. Differences
 in physicochemical properties among sediment types
 result in variations in biologically available metal at
 any given total metal concentration {for review, see
 Luoma 1989). Some have attempted to address this by
 predicting metal bioavailability in sediments through
 the use of differential extraction and fractionation
 schemes (e.g., Tessier and Campbell 1987). The ma-
Environmental Management Vol. 18. No. 3. pp. 331-337
                      © 1994 Springer-Verlag New York Inc.

-------
332
G. T. Ankley and others
jor shortcoming with this type of approach, however,
is that often it has been difficult to relate observed
chemistry results to the end point of concern, that is,
biological exposure and response.
   Early studies by Swartz and others (1985) demon-
strated a correlation between sediment  interstitial
(pore) water concentrations  of cadmium and  the
acute toxicity  of cadmium-spiked sediments to  the
marine amphipod Rhepaxymus abronius. This observa-
tion suggested that, based on acute toxicity, cadmium
in pore water represented the  bioavailable fraction of
the total cadmium in sediment. This observation also
was consistent  with the results of studies demonstrat-
ing that the toxicity of nonionic organic chemicals to
epibenthic and benthic species could be predicted by
comparison of chemical concentrations in  interstitial
water to toxicity data generated in water-only expo-
sures (e.g., Adams and others 1985, Swartz and others
1990). These observations served as the  basis for the
development of the equilibrium partitioning approach
to deriving sediment quality criteria (SQC) for non-
ionic organic compounds (Di Toro and others 1991).
In the case of nonionic organic chemicals, organic
carbon is the  major partitioning  phase controlling
pore water concentrations and bioavailability in sedi-
ments; however, until recently, comparable partition-
ing phases had not been as well defined for metals.
   In spiking experiments  by Di  Toro and others
(1990), it was demonstrated that the acute  toxicity of
cadmium to amphipods (Ampelisca abdita, R. kudsoni,
R. abronius) in  marine sediments could be predicted
based upon acid volatile sulfide (AVS) content of the
sediments. AVS is defined as that fraction of sulfide in
sediments extracted by cold HC1, and it exists in natu-
ral sediments  primarily as iron sulfide complexes
commonly referred to as mackinawite and greigite
(Berner 1967, Goldhaber and  Kaplan 1974). A num-
ber of cationic metals of environmental concern (zinc,
lead, copper, nickel, cadmium) will displace iron from
the monosulfide and thereby may be rendered biolog-
ically unavailable. Thus, in the experiments by Di
Toro and others (1990), when die molar ratio of cad-
mium  to AVS  exceeded one  (i.e., the AVS binding
pool was exhausted), interstitial water concentrations
of free cadmium increased dramatically,  and there
was a corresponding increase in amphipod  mortality.
   Subsequent   spiking experiments conducted  by
Carlson and others (1991) used AVS concentrations
to accurately predict cadmium toxicity in freshwater
sediments. In those experiments, acute toxicity to oli-
gochaetes (Lumbricuhu variegatus) and snails (Helisoma
sp.) was observed only when the molar cadmium/AVS
ratio exceeded one.
                                         Based on these studies, Di Toro and others (1992)
                                      proposed that AVS should be a suitable normalization
                                      phase for predicting the lack of acute toxicity of cad-
                                      mium, nickel, lead, zinc, and copper in both marine
                                      and freshwater sediments. Di Toro and others (1992)
                                      also described the utilization of simultaneously  ex-
                                      tracted metal (SEM) for normalization to AVS. That
                                      is, it would be inappropriate to use total sediment
                                      metals for normalization to AVS; metals should be
                                      measured in die same acid extraction fraction as that
                                      used for liberation of the AVS from the test sedi-
                                      ments. With the harsher extractions commonly used
                                      to measure "total" metals in sediments, a significant
                                      portion of the metals liberated are from the mineral-
                                      ogical matrix of the sediment, and hence are of little
                                      significance from die standpoint of biological avail-
                                      ability and effects..
                                         Ankley and others (1991) further investigated  the
                                      use of AVS normalization for prediction of the bio-
                                      availability of metals in field-collected sediments con-
                                      taminated by multiple metals. Sediments from die up-
                                      per end of a marine tidal estuary contaminated widi
                                      cadmiunvand nickel were tested for toxicity to fresh-
                                      water amphipods (Hyalella azteca) and oligochaetes (L.
                                      variegatus). Toxicity of the sediments to H. azteca, a
                                      relatively sensitive species, could be predicted based
                                      on SEM (nickel plus cadmiumXAVS ratios in die sedi-
                                      ments. Elevated  amphipod mortality was observed
                                      when die molar SEM/AVS ratio exceeded one, while
                                      toxicity was not seen at ratios of less than one. Oli-
                                      gochaetes, which  are relatively tolerant of metals,
                                      were less sensitive than the amphipods to die test sed-
                                      iments; however,  bioaccumulation of metals by  die
                                      worms was correlated with the sediment SEM/AVS
                                      ratios. Ankley and others (1991) also noted a correla-
                                      tion between sediment toxicity and pore water cad-
                                      mium and nickel concentrations.
                                         The same sediments evaluated by Ankley and oth-
                                      ers (1991) also were tested in salt water using marine
                                      amphipods (A. abdita). In these experiments, toxicity
                                      was not observed at SEM/AVS ratios less than one;
                                      however, toxicity also was not observed in several sam-
                                      ples with ratios greater than one (unpublished data).
                                      Evaluation of pore water concentrations of the nickel
                                      and cadmium, however, did provide an accurate pre-
                                      diction of the occurrence of toxicity. Regardless of the
                                      SEM/AVS ratio, toxicity was observed only in those
                                      sediment samples with pore water concentrations of
                                      cadmium and nickel that exceeded the joint water-
                                      only toxicity of the two metals to A. abdita. These data
                                      suggested the presence of binding phases in addition
                                      to AVS for cadmium and nickel in the test sediments.
                                         Further  laboratory spiking of marine sediments

-------
                                                        Metal Bioavailability in Sediments
                                          333
 with cadmium, zinc, nickel, and copper, either singly
 or in combination, demonstrated that acute toxicity to
 A. abdita could be predicted based upon SEM/AVS
 ratios in the test sediments (Hansen and others 1990;
 Berry and others 1991). Toxicity was observed when
 SEM/AVS ratios exceeded one, while at ratios less
 than one, survival of the amphipods was comparable
 to that in the control sediments. Pore water concentra-
 tions of the cadmium, zinc, nickel, and copper also
 could be used to accurately predict the occurrence
 and extent of toxicity to the amphipods, that is, mor-
 tality was observed only at pore water metal concen-
 trations exceeding those  causing mortality  to the
 amphipods in water-only exposures. In the spiking
 experiments with multiple metals, it also was noted
 that, in the sediments with SEM/AVS ratios greater
 than one, relative' pore water concentrations of the
 test metals were inversely proportional to their bind-
 ing constants for the sulfide.
   Ankley and others (1993a) evaluated AVS as a nor-
 malization phase for determining copper bioavailabil-
 ity in sediments from two freshwater sites contami-
 nated with the metal: Steilacoom Lake, Washington,
 and  the  Keweenaw   Waterway,  Michigan.  SEM
 (copperXAVS ratios in the test sediments were found
 to overpredict copper bioavailability in these studies.
 Acute toxicity to H. axteca did not occur in several test
 sediments with extremely high copper/AVS ratios,
 nor were pore water copper concentrations elevated
 in these sediments. This suggested the presence of an
 additional binding phase(s) for copper. Sediment ti-
. tration  experiments, similar to those described by Di
 Toro and others (1990) for cadmium, subsequently
 confirmed the presence of, strong binding phases,
 other than AVS, for copper in anaerobic sediments
 (Mahony and others  1991). This additional binding
 capacity was correlated with the organic carbon con-
 tent of the test sediments.  Significantly, although
 copper/A VS ratios could not be used to predict the
 occurrence of acute toxicity of copper to amphipods,
 measurement of pore  water copper concentrations
 and comparison of these data to water-only copper
 toxicity data'for H. asJteca enabled accurate predic-
 tion of the presence and extent of sediment toxicity
 (Ankley and others 1993a).
   The majority of studies examining SEM/AVS ra-
 tios and/or pore water metal concentrations as expo-
 sure models has been limited to short-term tests with
 lethality as an end point. One exception .was a recent
 study by Ankley and others (1993b) which evaluated
 the bioaccumulation of copper, lead, zinc, cadmium,
 nickel,  and  chromium by L. variegatus held  for an
 extended time in three different  sediment samples
from the lower Fox River, Wisconsin. These sedi-
ments had elevated concentrations of all six metals;
however, based upon SEM/AVS ratios or pore water
metal concentrations, metals in the sediments were
predicted to be of minimal biological availability to the
oligochaetes. After 30 days of exposure, metal con-
centrations in L. variegatus held in the sediments were
similar to concentrations in worms held in dean wa-
ter. This indicates that it may be possible to use metal/
AVS relationships  to predict metal exposure in long-
term  as well as  short-term studies  with  benthic
invertebrates. -     '
  AVS concentrations in aquatic sediments  are inti-
mately related to a number of biogeochemical cycles
and thus may  vary  greatly. In addition to varying
among sites, AVS  concentrations at a particular site
change with depth and with season. Therefore, in the
absence of alternative binding phases, it is  possible
that metals in some systems may be bioavailable dur-
ing some portions  of the year but not during others.
To address this, studies were initiated to better define
the seasonal variability in AVS concentrations among
sites and at sediment depths. For example, one study
has monitored sediment AVS concentrations in three
northern Minnesota lakes for, approximately two
years (Leonard and others 1993). Although the three
lakes were relatively similar, both geographically and
hydrologically, there were marked variations among
them in sediment AVS concentrations. Furthermore,
AVS concentrations varied with depth, with decreas-
ing amounts of AVS at greater depths. Most impor-
tantly there were marked seasonal variations in AVS,
in some instances  almost two orders of magnitude,
with the minimum concentrations generally occurring
in late winter and  the maximum concentrations ob-
served in late spring and early summer. These sea-
sonal variations were much more pronounced in surf-
icial sediments than in sediments at greater depths.
Limitations to Prediction of Metal Toxicity
in Sediments
   Based on the above studies it appears that evalua-
tion of pore water metal concentrations and/or SEM/
AVS ratios can  lend  insights concerning metal bio-
availability  in  sediments.  We feel that the two
techniques are complementary and should be used in
conjunction with one another as one approach to pro-
viding assessments of the potential ecological impacts
of metals in sediments. Both techniques have specific
problems and common limitations. These problems
and limitations are discussed below.

-------
334
G. T. Ankley and others
   It appears that comparison of metal concentrations
in pore water to water-only toxicity data can be used to
predict not only the presence, but also the extent, of
metal toxicity in sediments. The ability to actually
quantify bioavailable metals in sediments is attractive
for a number of reasons. For example, quantification
of bioavailable metal facilitates the evaluation of dif-
ferences in relative species sensitivity and thus enables
the identification of species at risk. This is not possible
with SEM/AVS ratios because when SEM/AVS ratios
exceed one, due to other possible binding phases, it is
impossible to predict actual pore water concentrations
of metals (e.g., Ankley and others 1991, 1993a). An-
other advantage to monitoring pore water metal con-
centrations is that they should be useful for predicting
the toxicity of metals, such as chromium, not reactive
with A VS. Finally, because AVS is readily oxidized, it
is not an important binding phase for metals in aero-
bic sediments; however, the bioavailable fraction of
metals should still be predictable based upon pore
water concentrations.
   There also are disadvantages to  using pore water
concentrations to predict metal bioavailability. First,
because pore water is operationally defined (i.e., there
is no standard method for isolation), there is a valid
concern that among-Iaboratory variations in prepara-
tion may result in significant differences in metal con-
centrations found in pore water. A* second disadvan-
tage to using pore water  metal concentrations  to
predict toxicity is that, if one accepts the paradigm
that pore water is indeed a major route of contami-
nant exposure  for epibenthic and benthic inverte-
brates, it may be difficult to account for the effects of
the pore water matrix (e.g., dissolved organic carbon,
hardness,  salinity) on metal complexation and bio-
availability. This is, of course, also an issue of ongoing
concern in the area of water quality criteria issued by
the US Environmental Protection Agency. A final po-
tential complication is that it is-necessary to have a
water-only effects data base for comparative purposes
for the metal and species of concern.
   Our studies have dearly demonstrated that AVS
can be a key factor influencing interstitial water con-
centrations and bioavailability of metals in sediments.
In no instance have we seen metal toxicity at SEM/
AVS ratios less than one, and ratios greater, than one
often have been predictive of the presence (but not
extent) of metal toxicity. The use of SEM/AVS con-
centrations alleviates the need  for water-only effects
data.in an assessment since no bioavailabitity is ex-
pected at  SEM/AVS ratios  less than one.  A further
advantage to measuring sediment SEM and AVS in
sediments is that it gives an indication as to the relative
                                       size of the pool of both components. This is not possi-
                                       ble through monitoring pore water metal concentra-
                                       tions; pore water metal concentrations should be low
                                       in sediments with SEM/AVS ratios ranging from ex-
                                       tremely low values to, theoretically, a ratio of 0.99.
                                       Yet, sediments with relatively high ratios would be of
                                       more potential concern than those with low ratios; in
                                       the absence of other metal binding phases, slight in-
                                      , creases in SEM or decreases in AVS could cause the
                                       SEM/AVS ratio to exceed one and, thereby, result in
                                       toxicity.
                                         There are a number of limitations inherent in us-
                                       ing either pore water or SEM/AVS ratios to predict
                                       metal bioavailability. First, because AVS varies season-
                                       ally in a system-specific manner, it is desirable that
                                       SEM/AVS ratios and pore water metals be measured
                                       over time, or  at least when AVS is expected to be
                                       minimal (e.g., late winter in our studies).  A single
                                       sampling is only a snap-shot of what occurs through
                                       the course of the year. Furthermore, even under re-
                                       ducing conditions, biological activity in surficial sedi-
                                       ments may serve to effectively oxidize AVS. At
                                       present, there is little understanding of the role of
                                       AVS in deeper sediments relative to metal partition-
                                       ing at the sediment surface, where most biological
                                       activity and exposure occurs. This is significant be-
                                       cause the AVS pool in deeper sediments appears to
                                       remain relatively constant, as opposed to AVS in surf-
                                       icial sediments.. It may be, for example, that as surfi-
                                       cial  sediments are depleted of AVS, metals will
                                       subsequently bind to AVS in deeper sediments. Alter-
                                       natively, as AVS concentrations are depleted in sur-
                                       face sediments, other binding phases for metals may
                                      'become important in determining bioavailability.
                                         Neither pore water metal concentrations nor SEM/
                                       AVS ratios can be used to assess potential metal bio-
                                       availability in situations where sediments are expected
                                       to be altered and  become aerobic through physical
                                       disturbance (e.g., storms, boat traffic, dredging). In
                                       fact, in these cases, it may be appropriate to "exhaust"
                                       sediment AVS (e.g., by aerating the sample) before
                                       attempting to evaluate the presence of bioavailable
                                       metals, possibly through evaluation of pore, water
                                       metal concentrations.
                                         An important limitation in using either pore water
                                       metal concentrations or SEM/AVS ratios  to evaluate
                                       metal bioavailability is that these approaches have not
                                       been thoroughly validated for predicting chronic tox-
                                       icity. With the exception of the 30-day bioaccumula-
                                       tion study described by  Ankley and others (1993b)
                                       and a recently completed marine colonization experi-
                                       ment, all the laboratory exposures conducted thus far
                                       .have been relatively short and  generally used only

-------
                                                        Metal Bioavailability in Sediments
                                           335
lethality as an end point. Thus, because of a current
lack of understanding of the dynamics of metals and
AVS during long-term exposures, we do not sanction
the use of either pore water metal concentrations or
SEM/AVS ratios for predicting the absence or pres-
ence of chronic toxicity.
   A final limitation to using either approach for pre-
dicting even acute toxicity  to epibenthic or benthic
species is that neither model has been field validated.
To date, the most exhaustive studies have been con-
ducted in a laboratory setting. Preliminary field stud-
ies focused on changes in benthic community struc-
ture or btoaccumulation of metals by benthos have
been consistent with  predictions based upon a pore
water exposure model (Campbell and Tessier 1991,
L. Hare, University of Quebec, personal communica-
tion); however,'further work  in this area is needed.
Therefore, care should be taken in interpretation of
SEM/AVS ratios and pore water metal concentrations
in the field relative to predicting the presence or ab-
sence of shifts in benthic community structure.

Recommendations for Predicting Metal
Bioavailability in Sediments
   Based both on the technical considerations and
limitations described above, we present the following
recommendations/caveats for assessing the potential
bioavailability of metals in sediments.
1.  Both SEM/AVS ratios and pore water metal con-
    centrations should be measured in  sediment as*
    sessments focused on defining metal bioavailabil-
    ity. A standard method for the extraction and
    measurement of SEM  and AVS has been de-
    scribed by Alien and others (1993). For the stud:
    ies described above, pore water was isolated using
    either of two different techniques: dialysis cham-
    bers (peepers) or centrifugation (e.g., see Di Toro
    and others 1990, Ankley and others 1991). Other
    pore water isolation techniques also may be use-
    ful; however, we have had little experience with
    •them.
2.  If AVS is used as a normalization phase, it should
    be used only for cadmium, nickel, lead, zinc, and
    copper, and only for these metals when simulta-
    neously extracted with the AVS. Molar concentra-
    tions of the metals then can be summed to gener-
    ate SEM/AVS ratios (Di Toro and others 1992).
    Theoretically, however, it is  possible to utilize
    pore water measurements of  metals, other than
    the five listed above, to evaluate their potential
    bioavailability (e.g., chromium; Ankley and  oth-
    ers 1993b).
3.  It is strongly recommended that cadmium, nickel,.
    copper, lead, and zinc all be measured when eval-
    uating SEM/AVS ratios and pore water metal con-
    centrations, at least in  initial test samples. This is
    because although all five of these metals have a
    higher affinity than iron for sulfide in monosul-
    fide complexes, individually they also have vary-
    ing affinities (solubility products) for the sulfide.
    Thus, for example, cadmium will displace nickel
    from sulfide, and if excess sulfide is not available,
    nickel will be released to the pore water. If only
    pore water metals were measured, or only nickel
    was measured  in the solid phase,  the analyst
    would erroneously conclude that nickel was die
    only problem in the sediments, when in fact, ele-
    vated' concentrations  of  cadmium also  were
    present. In order to have a complete understand- ;
    ing as to why a particular metal is present at ele-
    vated concentrations in pore water, it is necessary
    to know the molar concentrations of all the SEM.
    This is particularly true when considering the fact
    that metal concentrations often covary in contam-
    inated aquatic sediments, that is, rarely is only one
    metal of concern.
4.  In fully aerobic sediments (e.g., sand), AVS con-
    centrations should not be used to attempt to pre-
    dict the  bioavailability of metals in sediments.
    Theoretically, however,  it should be possible to
    infer bioavailability based on pore water metal
    concentrations. Moreover, significant progress is
    being made in identifying alternative normaliza-
    tion phases for metals in aerobic sediments (Camp-
    bell and Tessier 1991, Mahony and others'1991,
    Tessier and others 1993).
' 5.  Only a limited amount of research has been con-
    ducted to assess the utility of SEM/AVS  ratios or
    pore water concentrations for predicting metal
    bioavailability in long-term exposures. Given un-
    certainties in kinetics of metal and AVS interac- •
    tions in temporal cycles, and a lack of information
    of the importance of other metal binding phases
    relative to these cycles, extrapolations of the ex-
    posure model to long-term situations should be
    made  with  care.  Further information also  is
    needed concerning the nature of the microhabitat,
    of invertebrates relative to long-term changes in
    metal bioavailability in sediments.
6.  As  with any chemical-specific monitoring meth-
    od, the analyst should be aware  that: (a)  not all
    chemicals of possible toxicological concern can be
    measured in environmental samples; and (b)  in
    most instances, it is difficult to account for possi-
    ble toxicological interactions among measured

-------
336
G. T. Ankley and others
    toxicants. For both these reasons, we strongly rec-
    ommend that toxicity tests be an integral part of
    any assessment concerned with the effects of sed-
    iment contaminants.


Summary                        .

   Significant progress has been made in understand-
ing factors influencing the bioavailability of metals in
sediments, particularly from the standpoint of acute
toxicity. Evaluation of pore water metal concentra-
tions, as well as A VS-metal interactions, has enabled a
much better understanding of the dynamics of metal
bioavailability in sediments. However, researchers uti-
lizing SEM/AVS ratios and/or pore water metal con-
centrations to evaluate metal bioavailability in sedi-
ments should be aware of possible drawbacks and
limitations to these approaches. For example, the con-
sequences of spatial (depth) and/or seasonal variations
in AVS relative to metal bioavailability need to be
better defined, and further research  is required to
determine if metal bioavailability predictions made
using  these approaches are  valid for  predicting
chronic toxicity and/or impacts in situ.  Research in all
these areas is ongoing, and a summary of our results,
including a theoretical basis for developing SQC for
metals, will receive a formal critical review by the Sci-
ence Advisory Board of the US Environmental Pro-
tection Agency.
             i,
Acknowledgments

   We thank all of our colleagues at the Duluth and
Narragansett EPA laboratories and Manhattan Col-
lege for input to various technical aspects of this work.


Literature Cited

Adams, W. J., R. A. Kimerle, and R.  G. Masher. 1985.
  Aquatic safety assessment of chemicals sorbed  to sedi-
  ments. Pages 429-453 in R. D. Cardwell, R. Purdy and
  R. C. Banner (eds.), Aquatic toxicology and hazard assess-
  ment: Seventh symposium. STP 854. American Society
  for Testing and Materials, Philadelphia, Pennsylvania.
Alien, H. £.,  G.  Fu, and B. Deng. 1993.  Analysis of'acid
  volatile sulfide (AVS) and simultaneously extracted metals
  (SEM) for the estimation of potential toxicity in aquatic
  sediments.  Environmental Toxicology  and Chemistry  12:
  1441-1453.
Ankley, G. T., G. L. Phipps, E. N. Leonard, D. A. Benoit,
  V. R. Mattson, P. A. Kosian, A. M. Cotter, J. R. Dierkes, D.
  J. Hansen, and J. D. Mahony. 1991. Acid volatile sulfide as
  a factor mediating cadmium and nickel  bioavailability in
  contaminated  sediments. Environmental Toxicology  and
  Chemistry 10:1299-1307.
                                         Ankley, G. T., V. R. Mattson, E. N. Leonard, C. W. West,
                                           and J. L. Bennett 1993a. Predicting, the acute toxicity of
                                           copper in freshwater sediments: Evaluation of the role of
                                           acid volatile sulfide. Environmental Toxicology and Chemistry
                                           12:315-320.
                                         Ankley, G. T., E. N. Leonard, and V. R. Mattson. 1993b.
                                           Prediction of die bioaccumulation of metals from contam-
                                           inated sediments by the oligochaete, Lumbricuhu variega-
                                           tus. Water Research (in press).             '
                                         Berner, R. A.  1967. Thertnodynamic stability of sedimen-
                                           tary iron sulfides. American Journal of Science 265:773-785.
                                         Berry, W. J., D. J. Hansen, J. D. Mahony, D. L. Robson, and
                                           J. M. Corbin.  1991. The role of acid volatile sulfide in
                                           controlling the toxicity of a, metals mixture in sediment.
                                           Page 91 in Abstracts of the 12th Annual Meeting of the
                                           Society  of Environmental Toxicology  and Chemistry.
                                           Novermber 1991. Seatde, Washington.
                                         Campbell, P. G. C.. and A. Tessier. 1991. Biological avail-
                                           ability of  metals in sediments: analytical approaches.
                                           Pages 161-174 m J.-P. Vernet (ed.), Heavy metals in the
                                           environment. Elsevier, New York.
                                         Carlson, A. R., G. L. Phipps, V. R. Mattson, P. A. Kosian,
                                           and A. M. Cotter. 1991. The role of acid volatile sulfide in
                                           determining cadmium bioavailability and toxicity in fresh-
                                           water sediments. Environmental Toxicology and Chemistry
                                           10:1309-1319.

                                         Di Toro, D. M., J. D. Mahony, D. J. Hansen, K. J. Scott,
                                           M. B. Hicks, S. M. Mayr, and M. S. Redmond. 1990. Tox-
                                           icity of cadmium in sediments: The role of acid volatile
                                           sulfide.  Environmental Toxicology and Chemistry 9:1487-
                                           1502.
                                         Di Toro, D. M., C. Zarba, D. J. Hansen, R. C. Swartz, C. E.
                                           Cowan, H. E.  Allen, N. A. Thomas,  P. R. Paquin, and
                                           W. J. Berry. 1991. Technical basis for establishing sedi-
                                           ment quality criteria for non-ionic organic chemicals us-
                                           ing equilibrium partitioning. Environmental Toxicology and
                                           Chemistry 10:1541-1583.
                                         Di Toro, D. M., J. D. Mahony, D. J. Hansen, K. J. Scott, A. R.
                                           Carlson, and G.  T. Ankley.  1992. Acid volatile sulfide
                                           predicts the acute toxkity of cadmium and nickel in sedi-
                                           ments. Environmental Science and Technology 26:96-101.
                                         Goldhaber, M. B. and I. R. Kaplan. 1974. The sulfur cycle.
                                           Pages 569-655 in E. D. Goldberg (ed.), Sea, Vol. 5—Ma-
                                           rine chemistry. John Wiley & Sons, New York.
                                         Hansen, D. J., W. J. Berry, J. D. Mahony, A. R. Carlson,
                                           K. M. McKenna, and D. L. Robson.  1990. Acid volatile
                                           sulfide controls divalent metal toxicity in sediments. Page
                                           88 in Abstracts of the 11th Annual meeting of the Society
                                           of Environmental Toxicology and Chemistry,  Washing-
                                           ton, DC.
                                         Leonard, E.  N., V. R. Mattson, D. A. Benoit, R. A. Hoke,
                                           and G. T. Ankley. 1993. Seasonal variation of acid volatile
                                           sulfide in sediments from three northeastern Minnesota
                                           lakes. Hydrobiologia (in press).
                                         ' Luoma, S. N.  1989. Can we determine the biological avail-
                                           ability of sediment-bound trace elements? Hydrobiologia
                                            176/177:379-396.
                                         Mahony, J. D.. D. M. DiToro, A. M. Gonzalez, D. J. Hansen,

-------
                                                              Metal Bioavailability in Sediments
                                                337
  W. J. Berry, and G. T. Ankley. 1991. A sediment compo-
  nent in addition to acid volatile sulfide that may further
  control toxicity of metals. Page 91 in Abstracts of the 12th
  Annual Meeting of the Society of Environmental Toxicol-
  ogy and Chemistry, Seatde, Washington.
Swartz, R. C., G. R. Ditsworth, D.  W. Schults, and J. O.
  Lamberson. 1985. Sediment toxicity to a marine infaunal
  amphipod: Cadmium and  its interaction with sewage
  sludge. Marine Environmental Research 18:133-153.
Swartz, R. C., D. W. Schults, T. H. DeVVitt, G. R. Ditsworth,
  and J. O. Lamberson. 1990. Toxicily of fluoranthene in
  sediment to marine amphipods: A test of the equilibrium
  partitioning approach to sediment quality criteria. Envi-
  ronmental Toxicology and Chemistry 9:1071-1080.
Tessier, A., and P. G. C. Campbell. 1987. Partitioning of
  trace metals in sediments: Relationships with bioavailabii-
  ity. Hydrobiologia 149:43-52.
Tessier, A., Y. Couillard, P. G. C. Campbell, and J. C. Au-
  dair.  1993. Modeling cadmium partitioning in oxic lake
  sediments and Cd concentrations in  the freshwater  bi-
  valve Anodonta grandis (Mollusca, Pelecypoda). Limnology
  and Oceanography 38:1-17.

-------

-------
                                                             Environmental Toxicology and Chemistry, Vol. 15, No. 12, pp. 2056-2066, 1996
                                                                                                             O 1996 SETAC
                                                                                                           Primed in the USA
                                                                                                     0730-7268/96 $6.00 4  .00
               TECHNICAL BASIS AND PROPOSAL FOR DERIVING SEDIMENT

                                    QUALITY CRITERIA FOR METALS

          1                                    ''.'.'•                ~     ~

            GERALD T. ANKLEY,*T DOMINIC M. Di TORO,$ DAVID J. HANSEN§ and WALTER J. BERRY§
       flf.S. Environmental Protection Agency, Mid-continent Ecology Division, 6201 Congdon Boulevard, Duluth, Minnesota 55804
                             JHydroqual, Inc., One Lethbridge Plaza, Mahwah, New Jersey 07030, USA
                      Environmental Engineering Department, Manhattan College, Bronx, New York 10471, USA
         §U.S. Environmental Protection Agency, Atlantic Ecology Division. 27 Tarzwell Drive, Narragansett, Rhode Island 02882

                                        (Received 3 March \996; Accepted 18 June 1996)


      Abstract—In developing sediment quality criteria (SQC)  for metals, it is essential that bioavailability be a prime consideration.
      Different studies have shown that while dry weight metal  concentrations in sediments are not predictive of bioavailability, metal
      concentrations in interstitial (pore) water .are correlated with observed biological  effects. A key partitioning phase controlling
      cationic metal activity and toxicity in the sediment-interstitial water system is acid-volatile sutfide (AVS). Acid-volatile sulfide
      binds, on a mole-to-mote basis, a number of cationic metals of environmental concern (cadmium, copper, nickel, lead, zinc) forming
      insoluble sulfide complexes with minimal biological availability. Short-term (10-d) laboratory studies with a variety of marine and
      freshwater benthic organisms have demonstrated that when AVS concentrations in spiked or field-collected sediments exceed those
      of metals simultaneously extracted with the AVS, interstitial water metal concentrations remain below those predicted to cause
      effects, and toxicity does not •occur. Simitar observations  have, been made in life-cycle laboratory toxicity tests .with amphipods
      and chironomids in marine and freshwater sediments spiked with cadmium and zinc, respectively. In  addition, field colonization
      experiments, varying in length from several months to more than 1 year,  with cadmium- or zinc-spiked freshwater and marine
      sediments, have demonstrated a lack of biological effects when there is sufficient AVS to limit interstitial water metal concentrations.
      These studies on metal bioavailability  and toxicity in  sediments serve as the basis for proposed SQC for the metals cadmium,
      copper, nickel,  lead, and zinc. Specifically, four approaches for deriving criteria are described: (a) comparison of molar AVS
      concentrations to the summed molar concentration of the five metals simultaneously extracted with the AVS; (b) measurement of
      interstitial water metal concentrations and calculation of summed interstitial water criteria toxic units (IWCTU) for the five metals,
      based  upon final chronic values from water quality criteria documents; (c) calculation of summed IWCTU based upon sediment
      AVS concentrations and metal-specific partitioning of the metals to organic carbon; and (d) calculation of summed IWCTU based
      upon partitioning of the metals to a minimum binding phase sorbent (chromatographic sand). For a number of reasons. SQC derived
      using these approaches generally should be considered "no effect" values, i.e., with these techniques it is possible to predict when
      sediment metals will not be toxic, but not necessarily when metal toxicity will be manifested. Currently, approaches (a) and (b)
      are the most useful in terms of predicting metal bioavailability and deriving SQC. Further research is required, however, to fully
      implement approaches  (c) and (d). Additional research also is required to thoroughly understand processes controlling bioaccu-
      mulation of metals from sediments by benthic organisms, as well as accumulation of metals by pelagic species that ingest metal-
      contaminated benthos.

      Keywords—Sediment     Metal     Criteria    Programmic input   . Interstitial water     Acid-volatile sulfide
                     BACKGROUND

  Due to their widespread release and persistent nature, metals
such as silver, cadmium, copper, mercury, nickel, lead, and
zinc are commonly elevated in aquatic sediments. Thus, there
have been various proposals for deriving sediment criteria or
standards for protecting benthic communities from metal tox-
icity. Many such attempts have featured measurement of total
sediment metals followed by comparison to background metal
concentrations, or in some cases an effects-based endpoint [1-
5]. An important limitation to these types of approaches is that
causality is difficult to establish because values are based on
total rather than bioavailable metal concentrations, i.e., for any
given total metal concentration, adverse lexicological effects
may or may not occur, depending upon physicochemical char-
acteristics of the sediment of concern [6-8]. The phenomenon
of differential  bioavailability across sediment type also  has

   * To .whom correspondence may be addressed.
   This document has received an official EPA technical review; how-
ever, the content does not reflect official EPA policy. Mention of trade
names does not indicate endorsement by EPA or the federal govern-
ment.
been noted for other classes of contaminants, including non
ionic organic chemicals such as pesticides [9]. The sediment •
specific nature of contaminant bioavailability clearly repre-
sents  a challenge  for regulatory agencies, such as the U.S.
Environmental Protection Agency (EPA), attempting to derive
technically defensible sediment quality  criteria (SQC) with
broad applicability.
  As  a prelude to the development of national SQC by EPA,
a workshop was held in 1984, at which  experts in the disci-
plines  of environmental toxicology and chemistry met to re-
view possible approaches for dealing with the issue of differ-
ential contaminant bioavailability across sediments [10]. The
results  of various studies concerning contaminant bioavail-
ability in sediments were presented, including one evaluating
the organochlorine pesticide Kepone*. An intriguing obser-
vation from this study was that, while Kepone toxicity was
not predictable based upon total sediment dry weight concen-
trations, effects of the pesticide were strongly correlated with
its interstitial (pore-)water concentrations. That is, irrespective
of sediment type,  toxicity-response relationships based upon
pore-water concentrations were similar  [11]. In the  case'of
                                                          .2036

-------
 Sediment quality criteria for metals
                    Environ. Toxicol. Chan. 15, 1996   2057
. Kepone, it appeared that pore-water concentrations were con-
' trolled by partitioning of the organochlorine pesticide to or-
 ganic carbon in the test sediments. This was an important
 determinant in die recommendation that equilibrium partition-
 ing (EqP) be pursued as an approach for deriving SQC [10].
 -  Initially, EPA focused upon using the EqP approach to de-
 velop and validate SQC for nonionic organic chemicals; the
 technical basis of  the paradigm currently used for deriving
 criteria for these types of contaminants is'described in detail
 by Di Toro et al. [9J. The two basic elements of the model are
 (1)  prediction of bioavailability based upon organic carbon
 content of the sediment and the octanol-water partition  co-
 efficient (£ow) of the nonionic organic contaminant of concern,
 and (2) comparison of the resultant organic carbon "normal-
 ized" concentration to effects data from existing water quality
 criteria (WQC) for the contaminant. The EqP approach for
 predicting bioavailability of nonionic organics has received
 extensive peer review, including two assessments by the EPA
 Science Advisory Board, an independent panel of internation-
 ally recognized scientists, who recommended the method as
 the best currently available for deriving SQC [12,13]. To date,
 EqP has been used as the basis for developing draft national
 SQC for five nonionic pesticides and polycyclic aromatic hy-
 drocarbons [14-18] and also has  served as  a basis for  the
 derivation of "Tier 2" sediment quality guidelines for a  na-
 tional sediment inventory [19].
   More recently, EPA has evaluated the potential utility of the
 EqP approach for deriving SQC for metals. Initial studies sup-
 porting the concept were described by Swartz et al. [20] and
 Kemp  and Swartz [21]  who found that toxicity of cadmium
 to amphipods in marine sediments could be accurately pre-
 dicted  based upon pore-water concentrations of the metal.
 However, as opposed "to  the situation for nonionic organic
 chemicals and  organic  carbon, those  sediment partitioning
 phases controlling interstitial water concentrations of metals
 initially were not readily apparent.
   The purpose of this paper is to present a proposed approach
 for deriving SQC for the metals cadmium, copper, nickel, lead,
 and zinc based  upon a series of relatively recent studies ex-
 amining the bioavailability and toxicity of metals in sediment.
 As a prelude to describing the derivation of these SQC,  we
 first present a brief overview of the technical basis of  the
 approach. ,

 PREDICTING METAL TOXICITY: SHORT-TERM STUDIES
   Considerable research has focused upon elucidating  sedi-
 ment partitioning phases controlling metal bioavailability, of-
 ten through the use of elaborate sequential extraction proce-
 dures to identify physicochemical fractions with which metals
 are associated [22,23]. Through these types of techniques, it
 has  been established that key binding phases for metals in
 sediments include iron and manganese oxides and organic car-
 bon. However, an important shortcoming with these approach-
 es was that much of the work was done with sediments that
 had intentionally, or unintentionally, been oxidized through
 procedures such as drying. Thus, the techniques were appro-
 priate only for examining metal bioavailability in oxidized
 sediments. This resulted in an underestimate of the importance
 of metal-sulfide binding in the anaerobic horizons character-
 istic of most natural in-place sediments. A number of cationic
 metals form stable complexes with sulfide generated in sedi-
 ments by sulfate-reducing bacteria  [24-28].  From a lexico-
 logical standpoint this could be a particularly critical reaction
because predictions based upon chemical equilibria [27] sug-
gest that a miinbersof metals of environmental concern (e.g.,
cadmium, copper, mercury, nickel, lead, silver, zinc) form rel-
atively insoluble sulfides that should not be present in pore
water, and hence, biologically unavailable.
  Di Toro et al.  [8] investigated  the significance of sulfide
partitioning in controlling metal bioavailability in marine sed-
iments spiked with cadmium. In those experiments, the op-
erational definition of Cornwell and Morse [29] was used to
identify that fraction of amorphous sulfide, commonly termed
acid-volatile sulfide (AVS), available to interact with cadmium
in the  sediments. Specifically, the AVS was defined as the
sulfide liberated from wet sediment by treatment with 1 N HC1
acid. Di Toro et al. [8]'found that when the molar concentration
of AVS in the test sediments was larger than that of the cad-
mium (i.e., when the cadmium-to-AVS ratio was less than 1,
or the cadmium-to-AVS difference was less than  0), interstitial
water concentrations of the metal were small and no toxicity
was observed in 10-d tests with the amphipods Rhepoxynius
hudsoni or Ampelisca abdita.. Studies by Carlson et al. [30]
with cadmium-spiked freshwater sediments yielded similar re-
sults; when there was more AVS than metal, significant toxicity
was not observed in 10-d tests with oligochaetes (Lumbriculus
variegatus) or snails (Helisoma sp.). Based upon these initial
studies, Di Toro et al. [31] suggested that it may  be feasible
to derive metal SQC by direct comparison of molar AVS con-
centrations to die molar sum of the concentrations of cationic
metals  (specifically, cadmium, copper, nickel, lead, zinc) ex-
tracted with the AVS (summed simultaneously extracted met-
als; 2SEM).
  Casas and Crecelius [32] further explored this possibility
by conducting 10-d toxicity tests with the marine polychaete
Capitella capitata exposed to sediments spiked with zinc, lead,
or copper. As was true in earlier studies, elevated pore-water
metal concentrations and toxicity  were observed  only when
SEM concentrations exceeded those of AVS. Green et al. [33]
reported results  of another spiking experiment supporting the
general EqP approach to deriving SQC for metals. In their
study metal-sulfide partitioning was not directly evaluated, but
it was found that toxicity of cadmium-spiked marine sediments
to the meiobenthic copepodAmphiascus tenuiremis was pre-
dictable based upon interstitial water (but not sediment dry
weight) cadmium concentrations. Further spiking experiments
by Pesch et  al.  [34] demonstrated that 10-d  survival of the
marine polychaete Neanthes arenceodentata was comparable
to controls in cadmium- or nickel-spiked sediments with more
AVS than SEM. An important observation in their study was
that significant mortality did not always occur  in sediments
with more SEM than AVS. This appeared to be related, in
part, to the ability of the polychaete to sense elevated metal
concentrations and avoid burrowing into the test  sediments,
thereby limiting their exposure to metals in the interstitial
water.
  Berry et al. [35] described experiments in which A. abdita
was exposed for 10 d to  sediments spiked either singly, or in
combination, with cadmium, copper, nickel,  lead and zinc. As
in previous studies, significant toxicity  to the amphipod did
not occur when  AVS concentrations exceeded those of SEM.
Berry et al.  [35] also analyzed their data by comparing ob-
served  mortality to interstitial water metal concentrations ex-
pressed as toxic units (IWTU):
                  IWTU = [MJ/LC50
(D

-------
 2058    Environ. Toxicol. Chem. 15, 1996
                                         G.T. Aokley et al.
 where [Md] is the dissolved metal concentration in the inter-
 stitial water, and the LC50 is the concentration of the metal
 causing 50% mortality of the test species in a water-only test.
 If pore-water exposure in a sediment test is indeed equivalent
 to that in a water-only test, then 1IWTU should result in 50%
 mortality of the test animals. Berry et al. [35] reported that
 significant (>24%) mortality occurred in only 2.6% of sedi-
 ments with less than 0.5 IWTU, while samples with greater
 than 0.5 IWTU were toxic 91.9% of the time. Berry et al. [35]
 also made an important observation relative to pore-water met-
 al chemistry in their mixed-metals test. Chemical equilibrium
 calculations suggest that the relative affinity of metals for AVS
. should be copper > lead > cadmium > zinc > nickel [27,31].
 Hence, the appearance of the metals in pore water as AVS is
 "exhausted" should occur in an inverse order, e.g., zinc would
 replace nickel in a monosulfide complex and nickel would be
 liberated to the pore water, etc.  This trend was, in fact,  ob-
 served by Berry et al. [35].
  In addition to short-term laboratory experiments with spiked
 sediments, there have been a number of laboratory toxicity
 tests with metal-contaminated sediments from the field. Ankley
 et al. [36] exposed  L. variegatus and the ampbipod Hyalella
 azteca to 17 sediment samples from a freshwater/estuarine site
 along a gradient of cadmium and nickel contamination. In 10-d
 toxicity tests, H.  azteca  mortality was greater than controls
 only in sediments with more SEM (cadmium plus nickel) than
 AVS. Lumbriculus variegatus was  far. less sensitive to  the
 sediments man H. azteca, which correlates with the differential
 sensitivity of the two species in water-only tests with cadmium
 and nickel. Ankley et al. [37] examined the significance of
 AVS as a binding phase for copper in freshwater sediments
 from two copper-impacted sites. Based upon pore-water copper
 concentrations in the test sediments, the 10-d LC50 for H.
 azteca was 31 n-g/L; this corresponded very well with a mea-
 sured LC50 of 28 |tg/L  for amphipod in a 10-d water-only
 test. However, Ankley et al. [37] also found that survival of
 H.  azteca was comparable to control values in several sedi-
 ments with markedly more SEM than AVS suggesting that, in
 these samples, SEM in excess of AVS was not biologically
 available. Measurement of pore-water concentrations' in  the
 samples corroborated- this lack of bioavailability;  when sur-
 vival was comparable to controls, pore-water copper was non-
 detectable. This observation suggested the presence of binding
phases in addition to AVS for copper in the test sediments.
 Recent studies suggest that an important source of the extra
 binding capacity in these sediments was organic carbon [38].
  Hansen et al. [39] summarized the results of 10-d laboratory
 toxicity tests with amphipods, oligochaetes, and polychaetes
 using metal-contaminated field sediments from five different
 marine sites and four freshwater sites, including the three de-
 scribed above. They compared the toxicity data both to SEM
 (summed for mixed metal sites):  AVS relationships and pore-
 water metal concentrations (expressed as IWTU). In the 49
 sediments evaluated where metals were the likely cause of
 toxicity, 100% with less SEM than AVS and less than  0.5
 IWTU did not exhibit significant toxicity to the test species.
 Prediction of the occurrence of metal toxicity was less certain
 than prediction of an absence of toxicity; 66.7% of the 45
 samples with more SEM than AVS and more than 0.5 IWTU
 were toxic.
 PREDICTING METAL TOXICITY: LONG-TERM  STUDIES
  Taken as a whole, the short-term laboratory experiments
 with  metal-spiked  and .field-collected  sediments  present  a
 strong argument for the ability to predict an absence of metal
 toxicity based upon sediment SEM-to-AVS relationships and/
 or interstitial water metal concentrations. However, for this
 approach to serve as a valid basis for SQC derivation, com-
 parable predictive success must be demonstrated in long-term
 laboratory and field experiments where chronic effects could
 be manifested [40,41]. This demonstration was  the goal of
 experiments described by Hare et al. [42], DeWitt et al. [43],
 Hansen et al. [44], Liber et al. [45], and Sibley et al. [46]. An
 important experimental modification to these long-term stud-'
 ies, as opposed to the short-term tests described, above, was
 the collection of horizon-specific chemistry data.' This is re-
 quired because AVS concentrations often vary inversely with
 depth [47,48]; hence, chemistry performed on homogenized
 samples might not reflect the true exposure of benthic organ-
 isms dwelling in surficial sediments [40,42,49].
   DeWitt et al. [43] conducted a life-cycle test with the marine
 amphipod Leptocheirus plumutosus  exposed to cadmium-
 spiked sediments for 28  d. There were no significant effects
 on survival, growth, or reproduction in sediments containing
 more AVS than cadmium, in spite of the fact that these samples
 contained up to 363 mg  cadmium/kg  on a dry weight basis.
 Sibley et al.  [46] reported similar results from  a life-cycle
 (56-d) test conducted with the freshwater midge Chironomus
 tentans exposed to zinc-spiked sediments. In that experiment,
 significant effects on survival, growth, emergence, and repro-
 duction were only observed in sediments with SEM in excess
 of AVS. Performance of the midge was comparable to controls
 in sediments with more AVS  than SEM, at dry weight zinc
 concentrations as high as 270  mg/kg.
   Hansen et al. [44] conducted a 118-d colonization experi-
 ment-in which cadmium-spiked sediments  were  held  in the
 laboratory in a constant flow of raw seawater. Qualitative and
 quantitative analysis of colonization of the test sediments re-
 vealed no discernable difference from  controls when surficial
 AVS concentrations were greater man  those of SEM. In those
 instances where SEM was greater than AVS, pore-water cad-
 mium concentrations were elevated, and significant alterations
 in the types of benthic species present, and their richness and
 abundance were observed. Among-treatment differences in
 benthic community structure based on interstitial water con-
 centrations of cadmium were consistent with known sensitiv-
 ities of the species in water-only tests with cadmium.
   For approx. 1 year Hare et al. [42] conducted field colo-
 nization experiment in which uncontaminated freshwater sed-
 iments were spiked with cadmium and replaced in the  oligo-
 trophic lake from which they originally had been collected.
 They reported little or no effects on abundance or biomass of
 major benthic taxa in  any of  the treatments.  However, they
 did report a significant accumulation of cadmium by organisms
 from sediments with surficial SEM concentrations greater than
 those of AVS. These sediments also contained elevated con-
' centrations of cadmium in interstitial  water. Liber et at. [45]
 reported the results of colonization experiment conducted us-
 ing a design similar to that of Hare et al. [42], Sediments from
 a freshwater mesotrophic pond were spiked with zinc, replaced
 in the  field, and sampled over 16 months for collection of
 biological and analytical data. With the exception of the high-
 est spiking  concentration (ca.  700 mg/kg, dry weight), AVS
 concentrations remained larger than those of SEM, and pore-
 water zinc concentrations were generally nondetectable. The
 only observed  difference  in  benthic community structure

-------
 Sediment quality criteria for metals
                     Environ. Toxicol. Chem. 15, 19%    2059
.across the treatments was a slight decrease in the abundance
 of Naididae oligochaetes at the highest spiking concentration.
               *

  CONSIDERATIONS IN PREDICTING METAL TOXICITY

   In summary, results of the long-term laboratory and field
 experiments conducted to date represent convincing support
 of the conclusions reached in the initial short-term laboratory
 experiments, i.e., an absence (but not necessarily a presence)
 of metal toxicity can be reliably predicted based upon metal-
 sulfide relationships and/or  pore-water metal concentrations.
 However, to use these observations as a basis for predicting
 metal bioavailability or deriving SQC, a number of conceptual
 and practical issues, related to analytical measurements, sam-
 pling, additional binding phases, and metal "activity"  must
 be considered. Many of these were addressed by Ankley et al.
 [50]; the most salient to our proposed derivation of SQC are
 described below.
              i
 Analytical measurements

   An important aspect to deriving "global" SQC values is that
 the methods necessary to implement the approach be reasonably
 standard. From the standpoint of the metal SQC proposed below,
 a significant amount of research has gone into defining the
 extraction of SEM and AVS, the measurement of AVS, and the
 extraction of pore water The basic SEM/AVS extraction method
 recommended by EPA is that of Allen et al. [51]. In terms of
 AVS measurement, a number of techniques cave been success-
 fully utilized including gravimetric approaches [8,52], colon-
 metric assays [29], gas chromatography-photoionization detec-
 tion [32,53], and.specific ion electrodes [54-57]. Bufflap and
 Allen [58] recently reviewed approaches for isolating interstitial
 water for metal analysis. They concluded that centrifugation
 under a nitrogen atmosphere was the simplest technique likely
 to result in an unbiased estimate of metal concentrations in pore
 water. At present, die EPA recommends filtration of the isolated
 pore water through 0.4 to 0.45-ji polycarbonate filters to better
 define that fraction of aqueous metal  associated with toxicity
 [59]. An alternative approach to sampling pore water for de-
 termination of dissolved metals that has proven to be  quite
. successful is in situ dialysis  [35,39,42,45].

 Sampling

   As alluded to above, accurate prediction of exposure of ben-
 thic organisms to metals is critically dependent upon sampling
 appropriate sediment horizons at appropriate times. This is be-
 cause of the relatively labile nature of AVS due to oxidation.
 In fact it is this seemingly labile nature that has led some to
 question the practical utility of using AVS as a basis for EqP-
 derived metal SQC [40,41]. For example, there have been many
 observations of spatial (depth) variations in AVS concentrations,
 most of which indicate that surficial concentrations are smaller
 than those in deeper sediments [42,44,45,47,48,54.57,60 ]. This
 likely is due to oxidation of AVS at the sediment surface, a
 process that is enhanced by bioturbation [49]. In addition to
 varying with depth, AVS can vary seasonally. For example, in
 systems where overlying water contains appreciable oxygen dur-
 ing cold weather months, AVS tends to decrease, presumably
 due  to a constant rate  of oxidation of the AVS linked to a
 decrease  in  its generation  by sulfate-reducing  bacteria
 [47,52,61]. Based on these studies, it appears the way to  avoid
 possible underestimation of metal bioavailability is to sample
 the biologically "active" zone of sediments (e.g., 0-2 cm) at
 times when AVS might be expected to be present at small con-
 centrations (e.g., during the winter in aerobic systems).
   The  somewhat  subjective  aspects of these sampling rec-
 ommendations have been of  concern to us. However, recent
 research suggests that the transient nature of AVS may be
 overstated relative to predicting the fate of all  metal-sulfide
 complexes in aquatic sediments. Observations from the Duluth
 EPA laboratory made in the  early 1990s indicated that AVS
 concentrations in sediments contaminated by metals  such as
 cadmium and zinc tended to  be elevated over concentrations
 typically  expected in freshwater systems (G.T. Ankley, un-
 published data). The probable underlying basis  for these ob-
 servations did not become apparent, however, until a recent
 series of spiking and metal-sulfide stability experiments. The
• field colonization study of Liber et al. [45] demonstrated a
 strong positive correlation between the amount of zinc added
 to test  sediments and the resultant concentration of AVS in
 the samples. In fact, the initial design of their study attempted
 to produce test sediments with as much as five times more
 SEM (zinc) than AVS: however, the highest SEM-to-AVS ratio
 achieved  was only slightly larger than 1. Moreover, the ex-
 pected surficial depletion and seasonal variations in AVS were
 unexpectedly low in the  zinc-spiked sediments [45]. These
 observations suggested that zinc sulfide, which comprised the
 bulk of AVS in the spiked sediments,  was more stable than
 the iron sulfide that presumably was the source of most of the
 AVS in the control sediments. The apparent stability of both
 zinc sulfide and cadmium sulfide versus iron sulfide also has
 been noted in laboratory spiking experiments with freshwater
 sediments [46,49,62].
   In support of these observations, recent metal-sulfide oxi-
 dation  experiments conducted by Di. Toro et al. [63]  have
 confirmed mat  cadmium  and zinc form more stable sulfide
 solid phases than iron. If this is also true for sulfide complexes
 of copper, nickel, and lead, the issue of seasonal/spatial vari-
 ations in  AVS becomes of less concern because most of the
 studies evaluating variations in AVS have focused on iron
 sulfide (i.e., uncontaminated sediments). Thus, further research
 concerning the differential stability of metal-sulfides,  both
 from a temporal and spatial perspective, is definitely war-
 ranted.

 Additional binding phases
   Although AVS  is an important binding phase for metals,
 clearly other physicochemical factors influence metal parti-
 tioning in sediments. In aerobic  systems, or those with low
 productivity (i.e., where the absence of organic  carbon limits
 sulfate reduction) AVS plays little or no role in determining
 interstitial water concentrations of metals. For example, Leon-
 ard et al. [48] found that a relatively large percentage of sur-
 ficial sediments from open areas in Lake Michigan did not
 contain detectable AVS. In fact the great majority (42 of 46)
 of samples analyzed by Leonard et al. [48] contained less AVS ,
 than SEM, yet pore-water metal concentrations of cadmium,
 copper, nickel, lead, and zinc were consistently small or non-
 detectable. Even in  sediments where concentrations of AVS
 are significant, other partitioning phases may provide addi-
 tional binding capacity for SEM [37,53,64]. In anaerobic sed-
 iments, organic carbon appears to be  an additional binding
 phase controlling metal partitioning, in particular for cadmium,
 copper, and lead [38], while in aerobic sediments both organic ,
 carbon and iron and manganese oxides undoubtedly play a role
 in controlling  interstitial water concentrations of metals

-------
 2060   Environ. Toxicol. Chem. 15, 1996
                                          G.T. Ankley et,.»!.
 [22,23,64,65]. Even in substrates with very little metal binding
 capacity (e.g., chromatographic sand), surface adsorption as-
 sociated with cation exchange capacity will control pore-water
 metal concentrations to some degree [66]. Although an ideal
 SQC model for metals  would incorporate all possible .metal
 binding phases, current knowledge concerning partitioning/
 capacity of phases other than AVS is insufficient for practical
 application of this model. However, in the proposed metal SQC
 presented below, we describe two approaches through which
 phases other than AVS could be incorporated into criteria der-
 ivation.

 Meted activity                   •
  A substantial number of water-only experiments suggest that
 biological  effects often are correlated to the divalent metal
 activity,  {M2*}. Metal  activity is the molar.divalent metal
 concentration [M2*] corrected for the shielding effect of anions
 that are electrostatically attracted to the atoms in high ionic
 strength (e.g., concentrated) solutions. The law of mass action
 for dissolved chemical species is expressed in terms of metal
 activities rather than metal concentrations. .This procedure cor-
 rects for ionic strength effects. The claim is not that the only
 bioavailable form is M2+ (for example MOH+ may also be
 bioavailable) but that dissolved organic carbon and certain
 other ligand-complexed fractions generally are of less impor-
 tance in terms of toxicological significance.
  There are a number of examples of correlation of biological
 responses with metal activity. The acute toxicity of cadmium .
 to grass shrimp was determined at various concentrations of
 chloride and nitriloacetic acid (NTA), both of which form
 cadmium complexes [67]. In that study, 'the response was quite
 different at different concentrations of chloride, indexed by
 salinity, and NTA. However, when the concentration/response
 relationships were evaluated with respect to  Cd2* activity in
 the solution, the response curves collapsed into a single re-
 lationship. Comparable results have  been reported for exper-
iments with Gonyaulax tamarensis  exposed  to varying con-
centrations of copper and the inorganic ligand EDTA; in these
 studies, concentration/response relationships correlated with
Cu2* activity [68]. Chronic toxicity of zinc, with phytoplank-
ton growth as  the endpoint, also has been examined [69]; in
this case as NTA was added, the toxicity of zinc to Microcystis
decreased. However, the toxicity data were correlated with free
zinc activity. Similar  results for diatoms exposed to copper
 and the completing ligand Tris also have been found [70].
 Variations in Tris concentrations arid pH produced markedly
different growth rates, which all could be related to the Cu2*
 activity. A similar set of results with copper and algae were
described by Sunda and Lewis [71] with dissolved organic
carbon from river water as the completing ligand. Bioavail-
 ability, as a function of free metal activity, also has been ex-
 amined from the standpoint of metal accumulation. For ex-
 ample, uptake of copper by oysters was correlated, not to total
 copper concentration,  but to Cu2* activity [72].
  There are, however, examples of where metal toxicity seem-
 ingly is not well correlated with metal activity, suggesting that
 other metal species or variables also can be important [73].
 Nevertheless, even in these instances, factors that reduce metal
 activity limit toxicity. What can be drawn from this is that any
 approach to deriving no effect SQC for metals should focus
 on  minimization of metal activity.  In doing so, there is no
 implication that M2+ is  the only form of metal bioavailable,
 only that if metal activity is low,  no toxicity should occur.
                 PROPOSED METAL. SQC
   Based upon the technical information summarized above,
 we propose the following approach to deriving metal SQC.
 These values are intended to be the best .estimate of sediment
 metal concentrations that should not cause toxicity to benthic
 organisms. The SQC  are for five metals:  cadmium, copper,
 nickel, lead, and zinc. They are derived from EqP-based es-
 timates of metal concentrations associated with a lack of. ad-
 verse biological effects. This approach has been presented to
 and reviewed by the Science Advisory Board of EPA [38,74].'
   The SQC for all five metals collectively can be derived using
 four procedures: (a) comparing the sum of their molar con*
 centrations to the molar concentration of AVS in sediments
 (AVS criteria); (b) comparing the measured interstitial water
 concentrations of the metals to  WQC final chronic values
' (FCVs) (interstitial water criteria); (c) using organic carbon-
 based partition coefficients,, in additioiuo the AVS and SEM
 relationships, to compute interstitial water concentrations, fol-
 lowed by comparison  to FCVs (AVS and organic carbon cri-
 teria); or (d) using minimum partition coefficients (e.g., gen-"
 crated from chromatographic sand [66]) to  compute sediment
 concentrations that would not result in interstitial water ex-
 ceeding  metal FCVs (minimum partitioning criteria). These
 approaches are described in more detail below. When an SQC
 is exceeded, based upon any one of the four procedures, metal
 toxicity should not occur. Failing all of the approaches is in-
 dicative of a potential  problem that would entail further eval-
 uation. At present, we believe that the technical basis for im-
 plementing approaches (a) and (b) is supportable. As discussed
 below, however, additional research is required to implement
 procedures (c) and (d).                   '
   The following nomenclature is used in subsequent discus-
 sion of SQC derivation for metals. The SQC for the metals
 are expressed in molar units because of the molar stoichi-
 ometry of metal binding to AVS. Thus, solid-phase constitu-
 ents (AVS, SEM) are  in jjimol/g dry weight. The interstitial
 water metal concentrations are expressed in u.mol/L, either as
 dissolved concentrations [MJ or activities {M2*} [75]. The
 partition coefficients are in L pore water/g dry weight, con-
 sistent with the above solid- and aqueous-phase concentration
'units. The subscripted  notation, M,,, is used to distinguish dis-
 solved aqueous-phase  molar concentrations from solid-phase
 molar concentrations with no subscript. For the combined con -
 centration,  [SEMT], the units are moles of metal per volume
 of solid plus liquid phase (i.e., bulk).
   One final point should be made with respect to nomencla-
 ture. When we use the terms nontoxic or having no effect, ws
 mean only with respect to the five metals considered in this
 paper. The toxicity of field-collected sediments can be caused
 by other chemicals. Therefore, avoiding exceeding  the SQC
 for metals does not mean that the sediments are nontoxic. It
 only ensures that the five metals being considered should not
 have an undesirable biological effect. Moreover, as discussed"
 in detail below, exceeding the criteria for the five metals does
 not necessarily indicate that metals will cause toxicity. For
 these reasons, we  strongly recommend toxicity tests as an
 integral part of any assessment concerned with the effects of
 sediment-associated contaminants [50].

     SINGLE METAL  SEDIMENT QUALITY CRITERIA
  . Single metal criteria are not usually applicable  to field sit-
 uations because there  is almost always more than one metal
 to  be considered. In any case, as will become subsequently

-------
Sediment quality criteria for metals
                                                  Environ. Toxicol Chem. 15, 1996    2061
clear, it would be technically indefensible to derive criteria for
one metal at a time because of the competitive nature of AVS
binding. Nevertheless,, it is illustrative to present the logic for
single  metals as a prelude to the derivation of the multiple
metal criteria.

AVS criteria             '    .
  It has been demonstrated that if the SEM of a sediment is
less than or equal to the AVS .

                    [SEM] £  [AVS]                 (2)

then no toxic effects are seen. This is consistent with the results
of a chemical equilibrium model for the sediment-interstitial
water system [31]. The resulting metal activity {M2*} can be
related to die total SEM of the sediment and water, and to the
solubility products of the metal sulfide (£MS) and iron sulfide
    ). In particular it is true that at [SEM] £ [AVS) then
                     [SEMT]
Because  the ratio of metal-sulfide to iron-sulfide solubility
products (^Ms/^fts) is very small (<10-I) even for the most
soluble of the sulfides, the metal activity of the sediment is at
least five orders of magnitude smaller than the SEM (see Di
Toro et al. [31] for data sources and references). This indicates
that no biological effects would be seen if this sediment were
tested. Therefore the condition [SEM] £ [AVS] is a no effect
SQC.
  The reason we use the term "no effect" is that for the
condition [SEM] £ [AVS] no biological impacts are expected.
However, for .[SEM] > [AVS], which might  seemingly be
considered an SQC violation, there are many documented in-
stances where no biological  impacts occur, e.g., because or-
ganic carbon partitioning controls metal bioavailability, or spe-
cies of concern are insensitive to metals.
  Again, because of potential temporal and spatial variability
of AVS, we recommend that AVS and SEM measurements be
made using surficial sediments during periods when AVS con-
centrations are  smallest.  Note, that this  will not always be
during cool-weather seasons; for example, systems that be-
come anaerobic during the winter can maintain relatively large
sediment AVS concentrations [45].

Interstitial water criteria
  The condition [SEM] & [AVS] indicates that the metal ac-
tivity of the sediment-interstitial  water  system is low and, '
therefore, below lexicologically significant concentrations.
Another way of ensuring this is to place a condition on the
interstitial water activity directly. Suppose that we knew the
metal activity, denoted by {FCV}, that corresponded to the
[PCV]. Then the SQC corresponding to this  effect level is
{FCV}
                                                    (4)
It is quite difficult, however, to measure and/or calculate metal
activity in a solution phase at the low concentrations required
because it depends on the identities, concentrations, and ther-
modynamic affinities of other chemically reactive species that
are present. Also the WQC are not expressed on an activity
basis. -An approximation to this condition is

                     [Md] as [FCVJ                  (5)
where [FCVJ is the FCV applied to total dissolved concen-
trations. That is, we require that the total dissolved metal con-
                               centration in the pore water [Md] be less than the FCV applied
                               as a dissolved criterion. Although this requirement ignores the
                               effect of chemical speciation on both sides of the equation—
                               compare Equations 4 and 5—it is the approximation that is
                               currently being suggested by EPA for die  WQC for metals
                               [59]. That is, the WQC should be applied to the total dis-
                               solved—rather than the total acid recoverable—metal concen-
                               tration. Hence, if this second condition is satisfied it is con-
                               sistent with the level of protection afforded by the WQC.
                                 In- situations where the SEM exceeds the AVS ([SEM] >
                               [AVS]), but the interstitial water total dissolved metal is less
                               than the final chronic value ([MJ £ [FCVJ), this sediment
                               would not violate the  criteria. These cases-occur when sig-
                               nificant binding to other phases occurs. It should be noted that
                               using the FCV for metals in freshwater samples requires that
                               the hardness of the interstitial water be measured because the
                               WQC vary with hardness;
                                                    ,y.,     AVS and organic carbon criteria
                                 For sediments with an appreciable AVS concentration rel-
                               ative to SEM, the SQC requirement that [SEM] £ [AVS] is
                               a useful comparison. However, if,the AVS concentration is
                               small, then this comparison  is of little value. Similarly, as
                               described above, even in situations where significant AVS oc-
                               curs in sediments, other sorption phases may limit the metal
                               activity when SEM exceeds the AVS.
                                 One well-established example of an important metal-binding
                               phase in sediment is organic carbon. It has been demonstrated
                               that a relationship exists between the SEM that is in excess
                               of the AVS and the interstitial water metal activity {M2*}
                                           [SEM] - [AVS] = JT..OC/OC
                                                    (6)
                               where JE..OC is the partition coefficient between organic carbon
                               and interstitial water of the sediment on a metal activity basis,
                               and/oc is the weight fraction of organic carbon of the sediment
                               [76]. If it is required that the pore-water metal activity be at
                               the FCV, then the SQC for SEM in the single metal case would
                               be
                                                     [AVS]  + i^oc/oc{FCV}
                                                    (7)
                               If the activity is replaced with the total dissolved FCV, the
                               resultant SQC is
                                                     [AVS] +
                                                    (8)
where K^ is the partition coefficient between organic carbon
and interstitial water on the basis of dissolved metal. Note that
the metal-specific organic carbon-based partition coefficients
vary with respect to pH [76], so pH of the interstitial water
must be determined. In addition, because the FCV for the five
metals is hardness dependent in freshwater, an estimate  or
measurement of pore-water hardness is required.
  This is the third approach from which an SQC can be de-
rived. For sediments where organic carbon provides all binding
capacity in addition to AVS, it is interpreted the same as SQC
for nonionic organics [9]. That is, exceeding  the criterion
would imply that unacceptable biological impacts would occur.
Because the analysis  of sediment binding data and the esti-
mation of the K^. attributes all the binding to organic carbon,
using these constants would  imply  that this criterion is the
boundary between no effect and effect. It is  likely, however,
that the assumption that organic carbon is the only important
phase in addition to AVS often will not be correct.  In these
cases, it becomes a no effect criterion. Of course, using this

-------
 2062    Environ. Toxicol. Cheat. 15, 1996

 as an effect criterion assumes that applying the FCV on a total
 dissolved basis is appropriate. If, in fact, a significant fraction
 of the interstitial water metal is not bioavailable, then again,
 this criterion would be a no effect value.

 Minimum partitioning criteria
   It would be useful to have solid-phase  criteria that would
 effectively screen sediments for which metal concentrations
 are low enough so that no problem is anticipated. In developing
 this approach we examined substrates for which the partition
 coefficients were likely to be quite low, such as clean chro-
 matographic sand [66]. From these experiments it was possible
 to establish minimum partition coefficients (Ktj&S which could
 be applied to any sediment. Based on this, no effect SQC would
 be:
    = J^JFCVJ
                                                     (9)
Unlike Equation 8, no AVS term is included because it is to
be applied only to sediments having no detectable AVS. If
AVS is quantifiable, then Equation 2'.would apply. In deriving
JL.J. values appropriate to a given sediment, measurement or
estimation of interstitial water pH is required, as metal binding
is highly pH dependent [66]. Also required is pore-water hard-
ness for appropriate adjustment of FCVs for freshwater sam-
ples.   .                         .•'•••

Multiple metals criteria
  As described in the previous section, from a practical stand-
point it is insufficient and inappropriate to consider each metal
separately. Because of the interactive nature of metal-sulflde
binding, this is of particular concern for the AVS criteria.

AVS criteria
  Hie results of calculations using chemical equilibrium mod-
els  indicate mat metals act in a competitive manner when
binding to AVS. The five metals copper, lead, cadmium, zinc,
and nickel will bind to AVS and be convened to their re-
spective sutfides in this sequence, i.e., in the order of increasing
solubility. Therefore, they must be considered together. There
cannot be a criterion for just nickel, for example, because all
the other metals may be present as metal sulfides and, there-
fore, to some  extent as AVS. If these other metals are not
measured as SEM, then the SSEM will be misleadingly small,
and it  may appear that [2SEM] <  [AVS] when  in fact this
would not be true if all the metals are considered together. It
should be noted that we currently restrict this discussion to
the five metals listed above; however, in situations where other
sulfide-fonning metals (e.g., mercury, silver) are present at
high concentrations, they also must be considered.
  The equilibrium model prediction of the metal activity is
similar to the  single metal example when a mixture of the
metals is present. If the molar sum of SEM for the five metals
is less than or equal to the AVS, i.e.,
men
[SEM,] £ [AVS]
                       (M.)
                                                    (10)
where [SEM^] is the total SEM (}unol/L (bulk)) for the ith
metal. Thus the activity of each metal, {Mil. is unaffected by
the presence of the other sulfides. This can  be understood as
follows. Suppose that the chemical system  starts initially as
                                         G.T. Ankley et al.

 iron and metal sulfide solids and that the system proceeds to
 equilibrium by each solid dissolving to some extent. The iron
 sulfide dissolves until the solubility product of iron sulfide is
 satisfied. This sets the sulfide activity. Then each metal sulfide
 dissolves until reaching its solubility. Because so little of each
 dissolve relative to the iron sulfide, the interstitial water chem-
 istry is not appreciably changed. Hence, the sulfide activity
 remains the same and the metal activity  adjusts'to meet each
 solubility requirement. Therefore, each metal sulfide behaves
 independently of one another. The fact that they are only slight-
 ly soluble relative to iron sulfide is the cause of this behavior.
 Thus, the AVS criteria is easily extended to the case of multiple
 metals.

 Interstitial water criteria
  The application of the interstitial water criteria to multiple
 metals is complicated, not by the chemical interactions of the
 metals in the sediment-interstitial water system (as in the case
 with the AVS criteria) but rather because of their possible toxic
 interactions. Even if the individual concentrations do not ex-
 ceed the FCV of each metal (FCV,), the metals could exert
 additive effects that might result in toxichy [77-80]. Therefore,
 to address this potential additivity, the interstitial water metal
 concentrations are converted to toxic units (TUs)  and these
 are summed. Because FCVs are used as the effects concen-
 trations, these TUs are referred to as interstitial water criteria
 toxic units (IWCTUs). For freshwater sediments, the FCVs are
 hardness dependent for all of the metals under consideration
 and, thus, need to be adjusted to the hardness of the pore water'
 of the sediment being considered. For the  ith metal with a total
dissolved concentration [M^, the IWCTU is
                                                                                           (12)
                                       Failure to exceed  this SQC requires that the sum of the
                                       IWCTUs be less than or equal to one
                                                                                          (13)
                                                              [FCVJ
                                       Hence, the multiple metals criterion is quite similar to the
                                       single metal case (Eqn. 5) except that it is expressed as summed
                                       TUs.

                                       AVS and organic carbon criteria
                                         The case in which the sediment organic carbon is considered
                                       in addition to AVS is more complicated. Consider, first, a single
                                       metal. As discussed above, the relationship between the in-
                                       terstitial water concentration and sediment concentration for
                                       the ith metal is given by the equation
            [SEM,] = [AVS] +
                                                                                          (14)
where #d/xy is me metal-specific partition coefficient between
sediment organic carbon and interstitial water, and fMJ is the
total dissolved interstitial water metal concentration. For this
case, where the  interstitial water concentration is predicted
using the SEM in excess of the AVS, and the partition coef-
ficient between the excess SEM and the interstitial water
                                                                             [SEMJ - [AVS]
                                                                                               [Mw]i
                                                                                           (15)
                                       In order to apply this equation to the case.of multiple metals,
                                       it is first necessary to identify and quantify the metals that are
                                       not entirely present as metal sulfides. The best way to do this

-------
 Sediment quality criteria for metals
                                                                    Environ. Toxicol. Chem. IS, 1996   2063
 is to establish which metals are present as the metal sulfides
 and in what quantity. The procedure is to assign the AVS to
 the metals in the sequence of their solubility products from
 the lowest to the highest: SEM^, SEM^, SEMcd, SEM^, and
 SEMm. That is, the AVS-complexed metals would be copper,
 followed by lead, followed by cadmium, etc., until the AVS
 is exhausted. The remaining SEM is that amount present in
 excess of the AVS.                           ' .
   To be specific, let A[SEMJ be the excess SEM for each of
 the ith metals. The least soluble metal sulfide (of the five metals
 being considered in this analysis) is copper sulfide. Thus if
 the  simultaneously extracted copper  is  less than the AVS
 ([SEMcJ < [AVS]), then essentially all of it must be present
 as copper sulfide with no additional SEM^, present, such that
 A[SEMcJ = 0. .The remaining AVS is A[AVS] = [AVS] -
 [SEMcJ.
  This  computation is repeated next for lead because lead
 sulfide is the next least soluble sulfide. Suppose, unlike copper,
 the simultaneously extracted lead is not less than the remaining
 AVS ([SEMpJ > A[AVS]). Hence, only  a portion of the si-
 multaneously extracted lead is present as lead sulfide and the
 remaining SEM, which is denoted as A[SEMn], is the differ-
 ence between the remaining AVS, A[AVS] and the simulta-
 neously extracted lead: A[SEMpJ = [SEM.J - A[AVS]. Thus,
 a portion of the lead is present as lead  sulfide, and the re-
 mainder is excess SEM. Because the AVS has been exhausted
 by the lead in this .example, the remaining three metals would
 all be present as excess SEM such that: AfSEM]^ = [SEMcJ;
 A^EMk. = [SEMzJ; and A[SEM]Ni = [SEMNJ.
  For each of these metals, interstitial water concentrations
 can be determined from  the appropriate partition coefficients
                           A[SEMJ
                                                   (16)
This equation is analogous to Equation 14 for the single metal
example. Note that if A[SEMJ = 0, then so is the interstitial
water metal concentration. The IWCTUs are computed using
this equation for the interstitial water concentrations
y  DMUI  _ ,y
           ~
                               A[SEMJ
                                                   (17)
where Equation 15 is used'to compute the interstitial water
concentrations. Note  that the organic carbon-based partition
coefficients vary with, respect to pH, so pH of the interstitial
water must be determined, together with the hardness if nec-
essary.
  If mis SQC is not exceeded, the computed total IWCTU
concentration must be less or equal to one
                       A[SEMJ
                                                   (18)
Thus, this criterion is the IWCTU value, Equation 13, with
the interstitial water concentrations calculated from the excess
SEM for each metal and the appropriate organic carbon par-
tition coefficients.

Minimum partitioning criteria       .   •           .    •
  The no-effect criterion using the minimum partition coef-
ficients (^d^Binj) is analogous to that using the organic carbon-
based coefficients    '
      S
                        [SEMJ
                                                Again, minimum binding phase coefficients vary with pH, so
                                                pH of the pore water must be determined, along with hardness
                                                for adjustment of freshwater PCVs. Because the minimum
                                                partition coefficients are being used, mis criterion corresponds
                                                to the upper bound estimate of the IWCTUs.
                                                   To summarize, the proposed SQC ate as follows. The sed-
                                                iment passes the SQC if any one of these conditions is satisfied:
                                                (a) AVS criteria
                                                                     [SEMJ s [AVS]
                                                (b) Interstitial water criteria
                                                (c)  AVS and organic carbon criteria
                                                                Y     A[SEMJ
                                                (d) .Minimum partitioning criteria
                                                                 ^    [SEMJ
                                                                                    1
                                                   (10)
                                                                                                   (13)
                                                                                                   (18)
                                                   (19)
                                                   (19)
If any one of these conditions is violated, this does not mean
that the sediment is toxic. For example, if the AVS in a sed-
iment is nondetectable, then'condition (a) will be violated.
However, if there is sufficient organic carbon sorption so that
either condition (b) or (c) is satisfied, then the sediment would
be deemed acceptable.
  If all of these conditions are violated, then.there is reason
to believe that the sediment may be unacceptably contaminated
by these metals. Further testing and evaluations would there-
fore be useful in order to assess actual toxicity and its causal
relationship to the five metals. These may include acute and
chronic tests with species that are sensitive to the metals sus-
pected to be  causing the toxicity. Also, in situ community
assessments, sediment toxicity identification evaluations, and
seasonal characterizations of the SEM, AVS, and  interstitial
water concentrations would be appropriate [50]'.

         SUMMARY AND RECOMMENDATIONS
  This paper summarizes the technical basis for predicting the
bioavailability of metals in sediment and a proposal for es-
tablishing SQC for copper, cadmium, nickel, lead, and zinc.
The basis of the overall approach is the use of EqP theory
linked to the concept of maintaining metal activity in the sed-
iment-interstitial water system below effects levels. Extensive
toxicological data from short-term and long-term laboratory
and field experiments, with both marine and freshwater sed-
iments, and a variety of species indicates that it is possible to
predict reliably an absence of metal toxicity based  upon EqP
theory. At present, two of the four proposed components for
deriving metal SQC, the AVS and interstitial water approaches
(a and b), are technically defensible; work remains to establish.
the applicability of the AVS and organic carbon and minimum
partitioning methods (c and d) [74], Research issues for these
latter two approaches include the development of robust par-
titioning data sets for the five metals, as well as investigation
of factors such as metal competition for binding sites. With
the possible exception of the organic carbon approach, all these
criteria are intended as "no  effects"  rather than  "effects"
values. Even so, they should be useful for dealing with the
majority of sediments to which they are applied [4&66J.

-------
2064    Environ. Toxicol. Chem. 15, 1996
                                           G.T. Ankley et al.
   Additional research required to implement folly the pro-
posed SQC includes the development of uncertainty estimates
associated with any of the four approaches; part of this would
include their application to a variety of field settings and sed-
iment types. Research also is needed to establish the technical
basis for SQC for metals other than the five described herein,
such as  mercury, silver, arsenic, and chromium. Finally, the
SQC approaches described in this paper are intended to protect
benthic organisms from direct toxicity associated with expo-
sure to metal-contaminated sediments. They are not designed
to protect aquatic systems from metal release associated, for
example, with sediment suspension, or the transport of metals
into the  food web either from sediment ingestion or  the in-
gestion of contaminated benthos. This latter issue, in particular.
should be die  focus of future research given existing uncer-
tainty in the prediction of bioaccumulation of metals by ben-
thos [81].
Acknowledgement—This work was supported by the EPA Office of
Research and Development and Office of Water. The support of Mary
Reiley and Chris Zarba is gratefully acknowledged.  Technical con-
tributions and assistance were provided by numerous scientists at the
Duluth and Narragansett EPA laboratories, and cooperating univer-
sities, including Manhattan College, University of Wisconsin-Supe-
rior, University of Rhode  Island and the University of Minnesota.
David Mount, Russell Erickson, Chris Ingersoll, and Nelson Thomas
provided valuable comments on an earlier version of this manuscript.
Marlene Johnson provided considerable assistance in manuscript prep-
aration.
                       REFERENCES
 1. Sullivan, J-, J. Ball, E. Brick, S. Hausmann, G. Puarskl and
    D. Sopcich. 1985. Report of the technical subcommittee on de-
    termination of dredge material suitability for in-water disposal.
    Wisconsin Department of Natural Resources, Madison, WI, USA.
 2. Pereaud, D,, R. Jacqumagi and A. Hayton. 1989. Development
    of provincial sediment quality guidelines. Ontario Ministry of the
    Environment, Toronto, ON, Canada.
 3. Long, E.R. and L.G. Morgan. 1991. The potential for biological
    effects of sediment-sorbed contaminants tested in the national
    status and trends program. NOAA Technical Memorandum NOS
    OMA 52. National Oceanic and Atmospheric Administration, Se-
    attle, WA, USA.
 4. MacDonald, DJ>., US. Carr, FD. Calder, E.R. Long and C.G.
    Ingersoll. 1996. Development and evaluation of sediment quality
    guidelines for Florida coastal waters. Ecotoxicology 5:253-278.
 5. Ingersoll, C.G., PS. Havcrland, E.L. Branson, TJS. Canfield,
    F.S. Dwyer, C.E. Henke and NJJ.  Kemble. 1996. Calculation
    of sediment effect concentrations. J. Great Lakes Res. 22:602-
    623.
 6. Tessier, A. and P.G.C. Campbell.  1987. Partitioning of trace
    metals in sediments: Relationships with bioavailability. Hydro-
    biologia 149:43-52.
 7. Luoma, S.N. 1989. Can we determine the biological availability
    of sediment-bound trace elements? Hydrobiologia 176/177:379-
    396.           .     '  •
 8. Di Toro, D.M., JJ>. Mahouy, DJ. Hansen, KJ. Scott,  MJ.
    Htcks, S.M. Mayr and MS. Redmond. 1990. Toxicity of cad-
    mium in sediments: The role of acid volatile sulfide. Environ.
    Toxicol. Chem. 9:1487-1502,
 9. Di Toro, D.M., C. Zarba, D J. Hansen, R.C. Swartz, C.E.
    Cowan, HE. Allen, N A. Thomas, P.R. Paquin and W J. Berry.
    1991. Technical basis for establishing sediment quality criteria
    for non-ionic organic chemicals using equilibrium partitioning.
    Environ, Toxicol. Chem. 10:1541-1583.
10. Battelle. 1984. Sediment quality criteria development workshop.
    Office of Water Regulations and Standards, Criteria and Standards
    Division, Washington, DC.
11. Adams, WJ., RA. Kimerle and R.G. Mosher. 1985. Aquatic
 .   safety assessment of chemicals sorbed to sediments. In R.D. Card-
    well, R. Purdy  and R.C. Bahner, eds.. Aquatic Toxicology and
    Hazard  Assessment: Seventh Symposium.  STP 854.  American
    Society for Testing and Materials, Philadelphia, PA, USA, pp.
    429-453.
 12. U.S. Environmental Protection Agency. 1990. An SAB Report.
    Evaluation of the equilibrium partitioning (EqP) approach for
    assessing sediment quality. EPA-SAB-EPEC-90-006. Office of
    Water, Washington, DC.
 13. VS. Environmental Protection Agency. 1992. An SAB Report.
    Review of sediment criteria development methodology for non-
 .   ionic organic contaminants. EPA-SAB-EPEC-93-002. Office of
    Water, Washington, DC.                          '
 14. U.S. Environmental Protection Agency. 1993. Sediment quality
   .criteria for the protection of benthic organisms: Acenaphthene.
    EPA 822-R-93-013. Office of Water, Washington, DC.
 15, VS. Environmental Protection Agencyi 1993. Sediment quality
    criteria for the protection of benthic organisms: Endrin. EPA 822-
    R-93-016. Office of Water, Washington, DC.
 16. U.S. Environmental Protection Agency. 1993. Sediment quality
    criteria for the protection of benthic organisms: Dieldrin. EPA
    822-R-93-015. Office of Water, Washington, DC.
 17. VS. Environmental Protection Agency. 1993. Sediment quality
    criteria for the protection of benthic organisms: Phenanthrene.
    EPA 822-R-93-014. Office of Water, Washington, DC.
 18. ILS. Environmental Protection Agency. 1993. Sediment quality
    criteria for the protection of benthic organisms: Fluoranthene.
    EPA 822-R-93-012. Office of Water. Washington, DC.
 19. VS. Environmental Protection Agency. 1996. The national sed-
    iment quality  survey. A report to Congress on the extent and
    severity of sediment contamination in surface waters of the United
    States: EPA 823-D-96-002. Office of Water, Washington, DC.
20. Swartz, R.C., G.R. Ditsworth, D.W. Sennits and J.O. Lam-
    bersou. 1985. Sediment toxicity to a marine infaunal amphipod:
    Cadmium and its interaction with sewage sludge. Mar. Environ.
    Res. 18:133-153.
21. Kemp, PJ. and R.C. Swartz. 1986. Acute toxicity of interstitial
    and particle-bound cadmium to a marine infaunal amphipod. Mar.
    Environ. Rex. 26:135-153.
22. Tessier, A., P.G.C. Campbell and M. Bisson.  1979. Sequential
    extraction procedure for the speciation of paniculate trace metals.
    Anal. Chem. 51:844-851.
23. Luoma, S.N. and W. Bryan. 1981. A statistical assessment of
    the form of trace metals in oxidized sediments employing chem-
    ical extractants. Sci. Total Environ. 17:165-196.
24. Goldhaber, M.B. and I.R. Kaplan. 1974.  The sulfur cycle. In
    E.D. Goldberg, ed.. The Sea, Vol. 5—Marine Chemistry John
    Wiley & Sons. NY, USA, pp. 569-655.
25. Berner, RA. 1980. Early Diagentsis: A Theoretical Approach.
    Princeton University Press, Princeton, NJ, USA.
26. Davies-Colley, RJ., P.O. Nelson and KJ. Williamson. 1985.
    Sulfide control of cadmium and copper concentrations in anaer-
    obic estuarine sediments. Mar. Chem. 16:173-186.
27. Emerson, Sn L.^ Jacobs and B. Tebo. 1983.  The behavior of'
    trace metals in marine anoxic waters: Solubilities at the oxygen-
    hydrogen sulfide interface. In C.S. Wong, E. Boyle, K.W. Bruland
    and J.D. Burton, eds., Trace Metals in Sea Water. Plenum Press,
    NY. USA, pp. 579^608.
28. Morse, J.W.,F4.Millero,J.C.ConiwellandD.Rlckard. 1987.
    The chemistry of the hydrogen sulfide and  iron sulfide systems
    in natural waters. Earth Sci. Rev. 24:1-42.
29. Cornwell, J.C. and J.W. Morse.  1987. The characterization of
    iron sulfide minerals in anoxic marine sediments. Mar. Cheat.
    22:193-206.
30. Carlson, A.R., G.L. Phipps, V.R. Mattson, PA. Kosian and
    A.M. Cotter. 1991. The role of acid volatile  sulfide in deter-
    mining cadmium bioavailability in sediments. Environ. Toxicol.
    Chem. 10:1309-1319.
31. Di Toro, DM., J.D. Mahony, DJ. Hansen, KJ. Scott, A.R.
    Carlson and G.T. Ankley. 1992. Acid volatile sulfide predicts
    the acute toxicity of cadmium and nickel in sediments. Environ.
    Sci. Technol. 26:96-101.          "                       •
32: Casas, A.M. and.EA.  Crectlius. 1994. Relationship between
    acid volatile sulfide and the toxicity of zinc, lead and copper in
    marine sediments. Environ. Toxicol Chem, 13:529-436.
33. Green, A.S., G.T. Chandler and  E.R. Blood.  1993. Aqueous-,
    pore-water-, and sediment-phase cadmium: Toxicity relationships
    for a meiobenthic copepod. Environ. Toxicol.  Chem, 12:1497-
    1506.
34. Pesch, C.E., DJ. Hansen, W.S. Boothman,  W J. Berry and

-------
 Sediment quality criteria for metals
                      Environ. Toxical. Chem. 15, 1996    2065
    J.D. Mahony. 1995. The role of acid volatile sulfide and inter-
    stitial water metal concentrations  in determining bioavailability
    of cadmium and nickel from contaminated sediments to the ma-
    rine polychaete Neanthes  arenaceodentata. Environ. Toxicol.
    Chem. 14:129-141.
 35. Berry, W J., DJ. Hansen, JJ>. Mahony, D.L. Robson, D.M.
    Di Toro, B.P. Shipley, B. Rogers, JM. Corbin and W.S. Booth-
    man. 1996. Predicting the toxicity of  metal-spiked laboratory
    sediments using acid-volatile sulfide and interstitial water nor-
    malizations. Environ. Toxicol. Chem. 15:2067-2079.
 36. Ankley, G.T., GX. Phipps, E.N. Leonard, DA. Benoit, V.R.
    Mattson, PA. Koslan, AM. Cotter, J.R. Dierkes, D J. Hansen
    and JJ>. Mahony. 1991. Acid volatile sulfide as a factor medi-
    ating cadmium and nickel  bioavailability in contaminated sedi-
    ments. Environ. Toxicol. Chem. 10:1299-1307.
 37. Ankley, G.T., V.R. Mattson, E.N. Leonard, C.W. West and
    JX. Bennett.  1993. Predicting the acute toxicity of copper in
    freshwater sediments: Evaluation of the  role of acid-volatile sul-
    fide. Environ. Toxicol Chem. 12:315-320.
 38. UJS. Environmental Protection Agency. 1994. Equilibrium par-
    titioning approach to predicting metal bioavailability in sediments
    and the derivation of sediment quality  criteria for metals, Vol.
    1—Briefing Report to the Science  Advisory Board. EPA 822-D-
    94-002. Office of Water, Office of Research and Development,
    Washington, DC.
 39. Hansen, D J, W J. Berry, JJX Mahony, W.S. Boothman, D.M.
    Di Toro, D.L. Robson, G.T. Ankley, D. Ma, Q. Van and C.E.
    Pesch. 1996. Predicting the toxicity of metal-contaminated field
    sediments using interstitial concentration of metals and acid-vol-
    atile sulfide normalizations. Environ. Toxicol. Chem. 15:2080-
    2094.
40. Lnoma, S.N. and J.L. Carter. 1993. Understanding the toxicity
    of contaminants in sediments: Beyond the bioassay-based para-
    digm. Environ. Toxicol. Chem. 12:793-796.
41. Meyer, J.S., W. Davlson, B.  Snndby, XT. Orls, DJ. Lauren,
    U. Fdrstner, J. Hong and  D.G. Crosby. 1994. The effects  of
    variable redox potentials, pH and light on bioavailability in dy-
    namic water-sediment environments. In  JX. Hamelink, P.F Lan-
    dmm, H.L. Bergman and W.H. Benson^ eds., Bioavailability;
    Physical, Chemical and Biological Interactions. CRC, Boca Ra-
    ton, FL, USA, pp. 155-170.
42. Hare, L., R. Carignan and MA. Huerta-Diaz. 1994. An ex-
    perimental study of metal toxicity and accumulation by benthic
    invertebrates: Implications  for the acid volatile sulfide  (AVS)
    model. Umnol. Oceanogr. 39:1653-1668.
43. DeWltt, T.H., R.C. Swartz, DJ. Hansen,  D. McGovern and
    W.  J. Berry. 1996. Bioavailability and  chronic toxicity of cad-
    mium in sediment to the  estuarine amphipod. Leptocheirus plu-
    muiosus. Environ. Toxicol.  Cheat.  15:2095-2101.
44. Hansen, D J., J.D. Mahony, W J.  Berry, S J. Benyi, JM. Cor-
    bin, S.D. Pratt, D.M. DI Toro and M.B. Able. 1996. Chronic
    effect of cadmium in sediments on colonization by benthic marine
    organisms: An evaluation of the role of interstitial cadmium and
    acid-volatile .sulfide in biological availability. Environ. Toxicol.
    Chem. 15:2126-2137.
45. Liber, K., D J. Call, T.P. Market, K.L. Schmude, M J>. Balcer,
    F.W. Whiteman and G.T. Ankley. 1996. Effects of acid-volatile
    sulfide on zinc bioavailability and toxicity to benthic macroin- '
    vertebrates:  A spiked-sediment field experiment. Environ. Toxi-
    col. Chem. 15:2113-2125.
46. Slbley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard.
    1996. Predicting chronic  toxicity of sediments spiked with zinc:
    An  evaluation of the acid-volatile  sulfide (AVS) model using a
    life-cycle test with the midge Ckironomus tentans. Environ. Tox-
    icol. Chem.  15:2102-2112.
47. Howard, DM. and R.D. Evans. 1993. Acid-volatile sulfide (AVS)
   . in a seasonally anoxic mesotrophic lake: Seasonal  and spatial
  , changes in  sediment AVS.  Environ. Toxicol. Chem. 12:1051-
    1058.
48. Leonard, E.N., G.T. Ankley and RA. Hoke. 1996. Evaluation
    of metals in marine and freshwater surficial sediments from the
    Environmental  Monitoring and Assessment Program relative  to
    proposed sediment quality criteria  for metals. Environ. Toxicol.
    Chem. 15:2221-2232.
49. Peterson, G.S., G.T. Ankley and EJi.  Leonard. 1996.  Effect
    of bioturbation on metal-sulfide oxidation in surficial freshwater
    sediments. Environ. Toxicol. Chem. 15:2147-2155.
 50. Ankley, G.T., D.M. Di Toro, D J. Hansen, J.D. Mahony, W J.
    Berry, R.C. Swartz, RA. Hoke, NA. Thomas, A.W. Garrison,
    H.E. Allen and CS. Zarba. 1994. Assessing the potential bio-
    availability of metals in sediments: A proposed approach. Envi-
    ron. Manage. 18:331-337.                       •
 51. Allen, H.E., G. Fu and B. Deng. 1993. Analysts of acid-volatile
    sulfide (AVS) and simultaneously extracted metals (SEM) for the •
    estimation of potential toxicity in aquatic sediments. Environ.
    Toxicol.Chem.,12:1-13.
 52. Leonard, E.N.,V.R. Mattson, DA. Benoit, R A. Hoke and G.T.
    Ankley.  1993. Seasonal variation of acid volatile sulfide con-
    centration in sediment cores from three northeastern Minnesota
    lakes. Hydrobiologia 271:87-95.
 53. Slotton, D.G. and J.E. Renter. 1995. Heavy metals in intact and
    resuspended sediments of a California reservoir, with emphasis
    on potential bioavailability of copper arid zinc. Mar. Freshwater
    Res. 46:257-265.
 54. Boothman, W.S. and A. Helmstetter. 1992. Vertical and sea-
    sonal variability of acid volatile sulfides in marine,sediments.
    Research Report. U.S. Environmental Protection Agency, Nar-
    ragansett, RI.                                      '
 55. Brouwer, H. and TJ>. Murphy. 1994. Diffusion method for the
    determination of acid-volatile sulfides (AVS) in sediment En-
    viron. Toxicol. Chem. 13:1273-1275.
 56. Leonard, EJN., AM. Cotter and G.T. Ankley. 1996. Modified
    diffusion method for analysis of  acid volatile sulfides and si-
  .  multaneously extracted metals in freshwater sediment Environ.
    ToxicoL'Chem. 15:1479-1481.
 57. Brumbaugh, W.G., C.G. Ingenoll, N.E. Kemble, T.W. May
    and JX. Zajcek. 1994. Chemical  characterization of sediments
    and pore water from the upper Clark Fork River  and Milltown
    Reservoir, Montana. Environ. Toxicol. Chem.  13:1971-1983.
 58. Bufflap, W.E, and H.E. Allen. 1995. Sediment pore water col-
    lection methods: A  review. Water Res. 24:165-177.
 59. Prothro, M.G. 1993. Office of Water Policy and Technical Guid-
    ance on Interpretation and Implementation of Aquatic Life Metals
    Criteria. Letter from M.G. Prothro, Acting Assistant Administra-
    tor for the Office of Water, to Water Management Division Di-
    rectors, U.S. EPA Regions I-X. Office of Water, Washington, DC.
 60. Besser, J.M., C.G. IngersoU and J.P. Giesy. 1996.  Effects of
    spatial and temporal variability of acid volatile sulfide on the
    bioavailability of copper  and zinc  in freshwater sediments. En- .
    viron. Toxicol. Chem. 15:286^-293.
 61. Herlihy, A.T. and AX. Mills. 1985. Sulfate reduction in fresh-
    water  sediments receiving acid mine drainage. Appl. Environ.
    Micnbiol 49:179-186.
 62. Leonard, EJJ., V.R. Mattson and G.T. Ankley. 1995; Horizon-
    specific oxidation of acid volatile sulfide in relation to the toxicity
    of cadmium spiked into a freshwater sediment. Arch. Environ.
    Contain. Toxicol. 28:78-84.
63. Di Toro, D.M., J.D. Mahony, DJ. Hansen and WJ. Berry.
    1996. A model of the oxidation of iron and cadmium sulfide in
    sediments. Environ. Toxicol. Chem. 15:2168-2186.
64. Calamono, W., W. Ahlf and U. Fdrstner. 1990. Exchange of
    heavy metals between sediment components and water. In J.A.C.
    Broekart, S. Queer and F. Adams,  eds. Metal Speciation in the
    Environment. Springer-Verlag, Berlin, Germany, pp. 503-522.
65. Jenne, EA.  1968.  Controls on Mn, Fe, Co, Ni, Cu, and Zn
    concentrations in soils and water—The significant role of hydrous
    Mn and Fe oxides. ID Advances in Chemistry. American Chemical
    Society, Washington, DC, USA, pp. 337-387.
66. Hassan, S.M., A.W. Garrison, H.E. Allen, D.M. Di Toro and
    G.T. Ankley.  1996. Estimation of partition coefficients for five
    trace metals in sandy sediments and application to sediment qual-
    ity criteria. Environ. Toxicol. Chem. 15:2198-2208.
67. Sunda, W.G., D.W. Engel and RJW. Thuotte.  1978. Effect of
    chemical speciation  of toxicity of cadmium to  grass shrimp, Pa-
    laemonetes pugio: Importance to free cadmium ion. Environ. Set.
    Technol. 12:409-413.
68. Anderson, DM. and F.M.M. Morel. 1978. Copper sensitivitj
    of Gonyaulax tamarensis. Limnol.  Oceanogr.  23:283-295.
69. Allen, H.E., R.H. Hall and T.D. Brisbin. 1980. Metal speciatioi
    effects on aquatic toxicity. Environ. Sci. Technol. 14:441-443.
70. Sunda, W.G. and R.R.L. Gutllard. 1976. The relationship tx
    tween cupric ion activity and the • toxicity of copper to phyt«
    plankton. J. Mar. Res. 34:511-529.

-------
2066     Environ. Toxicol. Chem. 15, 1996
                                            G.T. Ankley et al.
71. Sunda, W.G. and J.M. Lewis. 1978. Effect of compilation by
    natural organic ligands on the toxicity of copper to a unicellular
    alga, Monochrysis lutheri. Lunnol. Oceanogr. 23:870-876.
72. Zamuda, CD. and W.G. Sunda. 1982. Bioavailability of dis-
    solved copper to the American oyster Crassostrea virginica: Im-
    portance of chemical speciation. Mar. Biol. 66:77-82.
73. Erickson, RJ., DA.. Benoit, V.R. Mattson, HJP. Nelson and
    E.N. Leonard. 1996. The effects of water chemistry on the tox-
    icity of copper lo fathead minnows. Environ. Toxicol. Chem: 15:
    181-193.
74. U.S. Environmental Protection Agency. 1995. An SAB report:
    Review of die agencies approach for developing sediment criteria
    for five metals. EPA-SAB-EPEC-95-020. Office of Water, Wash-
    ington, DC.
75. Stumm, W.  and JS. Morgan. 1981. Aquatic Chemistry—An
    Introduction  Emphasizing Chemical Equilibria in Natural Wa-
    ters. John Wiley & Sons, New York, NY, USA.
76. Mahony, J.D., DM. Dl Tore, AM. Gonzalez, M. Curto, M.
    Dtlg, L.D. De Rosa and LA. Sparrow. 1996. Partitioning of
    metals to sediment organic carbon. Environ. Toxicol. Chem. IS:
    2187-2197.          '
77. Bieslnger, K.E., GM. Christensen and J.T. Fiandt 1986. Ef-
    fects of metal salt mixtures on Daphnia magna reproduction.
    Ecotoxicol. Environ. Saf. 11:9-14.
78. Spehar, R.L. and J.T. Fiandt. 1986. Acute and chronic effects
    of water quality criteria-based metal mixtures on three aquatic
    species. Environ. Toxicol. Chem. 5:917-931.    .
79. Enserink,  EX., J.L. Maas-DIepeveen and CS. van Leeuwen.
    1991. Combined toxicity of metals: An ecotoxicological evalu-
    ation. Water Res. 25:679-687.
80. Kraak, MILS., D. Lavy, H. Schoon, M. Toussaint, W.H.M.
    Peelers and NJM. van Straalen. 1994. Ecotoxicity of mixtures
    of metals to the zebra mussel Dreissena polymorpha. Environ.
    Toxicol. Chem. 13:109-114.
81. Ankley, G.T. 1996. Evaluation of metal/acid-volatile sulflde re-
    lationships in the prediction of metal bioaccumulation by benthic
    macroinvertebrates. Environ. Toxicol. Chem. 15:2138-2146.

-------

-------
Cnviroamtnul Ttakolotymil Chemistry. Vol. 9. pp. 1447-1502,1990
Primed in Ihe USA Pcrgamon Press pic
                             0730-7268/90 J3.00-f .00
                             Copyright® 1990 SETAC
                , TOXICITY OF CADMIUM IN SEDIMENTS:
                  THE ROLE OF ACID VOLATILE SULFIDE

                DOMINIC M. Di TORO,* JOHN D. MAHONY, DAVID J. HANSEN,
                  K. JOHN SCOTT, MICHAEL B. HICKS, SUZANNE M. MAYR,
                                 and MICHELE S. REDMOND
       •Environmental Engineering and Science Program, Manhattan College, Bronx, New York 10471

                        (Received 2Q October 19S9; Accepted 21 February 199Q)

                                               *
       Abstract-The toxicity of chemicals in sediments is influenced by the extent that chemicals bind to
       the sediment. It is shown that acid volatile sulfide (AVS) is the sediment phase that determines the LC50
       for cadmium in the marine sediments tested. Although it is well known that metals can form insolu-
       ble sulfides, it apparently has not been recognized that AVS is a reactive pool of solid phase sulfide
       that is available to bind with metals. Amphipod sediment toxicity tests were conducted in the labo-
       ratory and the observed amphipod LCSOs on a normalized cadmium concentration baas, [Cd]/[AVS],
       is the same for sediments with over an order of magnitude difference in dry weight normalized cad-
       mium LCSOs.
         Because other toxic metals also form insoluble sulfides, it is likely that AYS is important in de-
       termining their toxicity in sediments as well. Most freshwater and marine sediments contain sufficient
       acid volatile sulfide for this phase to be the predominant determinant of toxicity. The other sorptkm
       phases are expected to be important only for low AVS sediments, for example, fully oxidized sedi-
       ments. From the point of view of sediment quality criteria the other sorption phases would be impor-
       tant for metals with large partition coefficients and large chronic water quality criteria.

       Keywords-Sediment quality criteria    Metal bioavailabiliry    Iron sulfide
             ,  INTRODUCTION
   The toxicity of chemicals in sediments is influ-
enced by the extent that chemicals bind to the sed-
iment. This modifies the chemical potential to
which the organisms are subjected. As a  conse-
quence, different sediments will exhibit different
degrees of toxicity for the same total quantity of
chemical. These differences have been reconciled
by relating organism response to the chemical con-
centration in the interstitial water of the sediments
[1-7]; the relevant sediment properties, therefore,
are those that influence the distribution of chemi-
cal between the solid and aqueous phases.
   Hie varying toxicity of nonionic organic chem-
icals in different sediments is related to the organic
carbon content of the  sediments 11,6-8], This is

   *To whom correspondence may be addressed.
   The address of J.D. Mahohy, M.B. Hicks and S.M.
Mayr is Chemistry Department. Manhattan College,
Bronx, NY 10471.
   The address of D.J. Hansen is EPA Environmental
Research Laboratory, Narragansett, RI02592.
   The address of K. J. Scott and M.S. Redmond is Sci-
ence Applications International Corp., Narragansett, RI
02592.
due to the importance of sediment organic carbon
in determining the extent of sorption of nonionic
organic chemicals to sediments. The analogous
sediment properties for metals would be the phases
that influence partitioning behavior. It has been
suggested  that oxides of iron and manganese as
well as organic carbon may be relevant in deter-
mining the toxicity of metals in sediments [9].
   The purpose of this paper is to establish the pri-
macy of the acid volatile sulfide (AVS) phase-the
solid  phase sediment sulfides that are soluble in
cold acid—in determining the toxicity of cadmium
in sediments. The results of toxicity tests using
three  sediments with differing AVS concentrations
indicate that the LCSOs are different on a dry
weight basis. However, the interstitial water and
water only LC50s are similar. The problem is to
determine what sediment parameter is controlling
the cadmium activity.            /•
   Experimental cadmium titralions of iron sulfide
and natural sediments indicate that cadmium can
react  with the solid phase'AVS to form cadmium
sulfide precipitate. If the quantity of AVS in a sed-
iment exceeds the quantity of added cadmium, the
concentration of cadmium in the interstitial water
                                             1487

-------
 1488
                                     D. M. Dl TORO FT AL.
 is nondetectable and no mortality is observed. Be-
 cause the AVS is sufficiently reactive, the added
 cadmium precipitates as cadmium sulfide which is
 insoluble so that the  interstitial water cadmium
 concentration is low. In addition, CdS itself appar-
 ently is not bioavailable. As long as excess AVS is
 present no free cadmium exists. However,  if the
 added cadmium exceeds the AVS, free cadmium is
 measured in the interstitial water and amphipod
 mortality occurs. The presentation that follows
 gives the experimental evidence that leads to these
 conclusions.
          MATERIALS AND METHODS
•Organism collection, holding and
 exposure system design
   Ampeliscaabditav/erc collected from tidal flat
 sediments in the Pettaquamscutt (Narrow) River,
 a small estuary flowing into Narragansett Bay,
 Rhode Island; transferred to the laboratory within
 one-half hour and separated using a 0.5 mm mesh'
 screen. Rhepoxynius Hudson!—a new test species- .
 were collected in shallow water at Ninigret Pond,
 Rhode Island. Ampelisca is a tube-dwelling amphi-
 pod living in the aerobic surface sediments or in an
 oxic tube which penetrates anoxic sediments. Rhe-
poxyniusis a free-burrower in the oxic zone. Sub-
 adult annuals were separated from the sediment
 using a 1 mm mesh screen in the field, transported
 to the laboratory within 1 h, screened again and
 transferred to holding containers. The amphipods
 were acclimated in presieved uncontaminated col-
 lection site sediment and flowing filtered 20°C sea-
 water. During acclimation, the Ampelisca were
 fed, ad libitum, the laboratory cultured diatom
 Phaeodactylum tricomutum. R. hudsoni could ob-
 tain food from sediments but  received no sup-
 plemental diet.
   A 96-h static renewal toxicity test was con- •
ducted with the two marine amphipod test species
to determine their response to cadmium in a seawa-
ter-only exposure. The water-only tests were con-
ducted in 900 ml glass chambers at  20°C with no
 aeration.
   The sediment toxicity tests generally followed
 American  Society for  Testing and  Materials
 (ASTM) recommendations [10] with changes to ac-
 commodate experimental design requirements of
 this experiment. For  all sediment experiments,
 flowing  filtered seawater (10  volume replace-
 ments/d) and aeration ensured acceptable dis-
 solved oxygen concentration and cadmium-free
 overlying water, which was confirmed by measure-
ments. In all experiments-lighting was continuous
so that the amphipods would not leave their tubes.
   The interstitial and overlying water was sampled
using a diffusion sampler ("peeper") [11,12], de-
signed to fit within the gallon jar exposure cham-
ber. It  was constructed of Plexiglas  G grade
unshrunk cast acrylic sheet: 1S.2 x 7.6 x  5.1 cm
deep with six rows of three 1.9 cm diameter, 3.8
cm deep holes, each with a volume of about 5 ml.
The samples from the three horizontal cavities are
combined to yield  a sample volume of  IS  ml,
which is required for the electrode measurement.
The open side of the peeper is covered by  a sheet
of 1 micron polycarbonate membrane  (Nucleo-
pore, Pleasanton, CA),  followed by a 0.076 cm
low density polyethylene gasket and a 1.3 cm Plexi-
glas cover plate, both of which have the same hole
pattern as the body and  secured with PVC-1 cap
screws and nuts. Equilibration time for cadmium
in the peeper cavities was determined to  be less
than 1 d in water-only experiments in which seawa-
ter was allowed to diffuse bto the cavities and time
course measurements were made.
   Three 10-d tests were conducted exposing the
amphipod Ampelisca abdita to control and cad-
mium-spiked Long Island Sound sediment in gal-
lon jars, using one liter of sediment and two liters
of overlying water. In two of these tests the jars
also contained the diffusion samplers.
   The final toxicity test was conducted using three
sediments in 900 ml exposure vessels with 200 ml
of sediment (3.5 cm depth) and 600 ml of overly-
ing seawater. Amphipods were exposed for  10 d to
control and cadmium-spiked sediments. Interstitial
water samples were taken by centrifugrag the sed-
iment from the chemical control vessels at the end
of the experiment.

Sediment acid volatile sutfide
   The principal sediment property of concern in
these experiments was the  acid volatile'sulfide
(AVS) concentration. It is the solid phase sulfide in
the sediment that is soluble .in cold acid (Hd). The
measurement technique is to convert the sulfldes to
H2S(aq), purge it with nitrogen gas and  trap it
[13]. A 500  ml Erlenmeyer flask reaction vessel
fitted with a three-hole stopper is followed by three
sequentially  connected 250 'ml Erlenmeyer flask
trapping vessels, the first is a chloride trap with
200 ml of pH 4 buffer (0.05 M potassium hydrogen
phthlate) to prevent chloride carry over. The sec-
ond and third traps contain 200 ml of a 0.1 M sil-
ver nitrate solution for trapping H2S  as Ag2S

-------
                             Acid volatile sulfide and cadmium toxicity
                                          1489
precipitate. The four flasks are connected with air-
tight appropriately shaped glass and Tygon tubing.
   To prevent oxidation of HjS the nitrogen gas
flows through an oxygen-scrubbing system consist-
ing of a vanadous  chloride solution  in the first
scrubbing tower and seawater in the second tower.
Vanadous chloride is prepared using 4 g of ammo-
nium metavanadate boiled with 50 ml of concen-
trated hydrochloric acid  and diluted  to 500 ml.
Amalgamated one, prepared by taking about IS g
of zinc, covering it with deionized water and add-
ing three drops of concentrated hydrochloric acid
before adding a small amount of mercury to com-
plete the amalgamation, is then added to the vana-
dous chloride solution.
   The sediment sample (10-15 g of wet sediment)
or standard to be analyzed is placed in the reaction
vessel after the entire system has been purged with
nitrogen for about 1 h. The system is again purged
for 5  to tO mm, and deaerated 6 u hydrochloric
acid is added  from a thistle tube to achieve a 0.5 u
final concentration in the vessel. The system is run
at room temperature for 1 h which has been found
by experiment to be sufficient to complete the ex-
traction. The  nitrogen gas flows at a bubble rate of
about four per second. The sample vessel is swirled
every 5 or 10 min. At completion, all hydrogen sul-
fide produced has been converted to silver sulfide
in the first silver nitrate trap and no precipitate is
found in the second trap. The suspension in the
first silver nitrate trap is passed through a 1.2 mi-
cron GF fiber filter, dried at 102°C and weighed.
   Standards  prepared from appropriate quantities
of iron(II) sulfate and sodium sulfide (the latter be-
ing added from a solution standardized against
lead perchlorate), typically gave yields of 95 to
103%. Silver sulfide precipitates were usually in the
range 20 to 30 mg. When a blank was run (sample
without acid), about 0.9 mg silver sulfide was ob-
tained. When the actd was run without a sample,
about 0.6 mg silver chloride was obtained. This
corresponds to a detection limit of ~0.5 pmol/g.

Sediment characterization and
spiking procedure
   Sediments of three different acid-volatile sulfide
concentrations were used  in the toxicity tests. The
Long Island  Sound sediment, with a high AVS
concentration, was collected from an uncontami-
nated site in central Long Island Sound (40°7.95'N
and 72°52.7'W) with a  Smith-Mclntyre grab sam-
pler, returned to  the laboratory,  press sieved wet
through a 2  mm mesh stainless steel screen, ho-
 mogenized and stored at 4°C. A. abdita has been
 tested many times in this sediment and both its sur-
 vival and reproduction are satisfactory [14]. The
 Ninigret Pond sediment was a low AVS sand col-
 lected from the Rhepoxynius collection site. The
 upper few inches of sediment were collected with
 a shovel, returned to the laboratory, sieved wet
 through a 2 min stainless steel screen, rinsed sev-
 eral times to remove high-organic fine particles,
 homogenized  and stored at 4°C. The third sedi-
 ment was a 50/50 (v/v) mixture of Long Island
 Sound and Ninigret  Pond sediments. Sediment
 samples were also obtained from the saline region
 of the Hudson River and from Black Rock Harbor
 on the north shore of Long Island Sound. These
 were used only in the sediment titratipn procedure.
    Sediments were spiked by adding 1.0 liter of wet
 sediment to 2.0 liter of 20°C filtered seawater
 where a weighted amount of cadmium chloride
 had been dissolved. The mixture was stirred with.
 a nylon spatula, capped and placed on a Ro-Tap
 sieve shaker for 5 min to ensure complete mixing
 and held at ambient temperature (-15°Q in a wa-
 ter bath for 7 d (three-sediment experiment) or 3 to
 5 d (one-sediment experiment) with gentle aeration
 of the water column. The overlying seawater and
 a thin layer of cadmium sulfide precipitate that had
 formed on the surface of the sediment were re-
 moved. In the three-sediment toxicity test, the test
 sediments were then homogenized, and 200 ml
 were transferred to each of three replicate exposure
 containers. For the one-sediment experiments, sed-
 iments were not homogenized. Clean filtered sea-
 water was added to replace the water siphoned out,
 and the peepers, if used, were inserted. In all cases,
 containers were then placed in a water bath with
 air and seawater delivery.

 Toxicity tests
    The tests were initiated by adding the sediments
 to the exposure containers, inserting the peepers
 and waiting 1 d (except third one-sediment exper-
 iment: 6 d). The amphipods were separated from
 holding containers using a 0.5 mm stainless steel
 screen and distributed sequentially in 100 ml plas-
 tic beakers until 30 (three-sediment experiment), 50
 (one-sediment experiments) or 25  (water-only ex-
 periment) Ampeliscaoi  20 Rhepoxynius were ac-
 cumulated. After sorting and eliminating dead
. animals, the delivery in the exposure system was
 halted, and one beaker of amphipods was added
 randomly to each duplicate exposure container in
 each treatment. Rhepoxynius were tested in Nini-

-------
1490
                                    D. M. Dl ToRO ET At.
gret Pond sediments, and Ampelisca in Long Is-
land Sound and the sediment mixture. Additional
containers for each treatment 'were chemical con-
trols and received no amphipods. Temperature and
salinity of delivered seawater averaged approxi-
mately 20°C and 31 g/L during the exposure pe-
riod for the  three-sediment test and  the three
one-sediment tests, respectively.
   After termination, the contents of each expo-
sure container, but not the chemical controls, were
sieved through a 0.5 mm screen. For Ampelisca,
material retained on the screen was preserved in
5% buffered formaiin with Rose Bengal stain for
later sorting. For Rhepoxynius, material retained
on the screen was examined immediately. In both
cases, recovered amphipods were counted,  and
missing individuals were counted as mortalities.

Cadmium determinations and titrations
   The AVS and solid phase cadmium were mea-
sured in the chemical control  vessels at both the
start and end of the experiments. The average
of these concentrations is used in the subsequent
analysis. At the termination of the experiments the
cadmium ion concentrations from the peepers
(one-sediment experiments), or from the centrifu-
gate (three-sediment experiment) and during the ti-
trations described below, were measured as Cdz+
activity using an Orion 94-48 cadmium ion selective
electrode and a double junction reference electrode
(Orion 90-02, Cambridge. MA). The electrode was
standardized with a serial dilution of a 1 g/L cad-
mium solution that was also used as the titrant.
Cadmium concentrations are reported as cadmium
activity, mg Cd2+/L. Independent experiments in
seawater indicate that the cadmium activity is 5.0%
of the total dissolved cadmium concentration. Sed-
iment cadmium was determined using a cold con-
centrated nitric acid (16 M, 5 ml) .digestion of 10 ml
wet sediment followed by a peroxide  oxidation
(10 ml 30%) and evaporation to dryness. The res-
idue was reconstituted to 20 ml using 0.1  M nitric
add and the cadmium measured using flame atomic
absorption.                    '  *
   Cadmium titrations of FeS suspensions (pre-
pared in the same manner as the AVS standards)
and sediments were performed using sample sizes
of 5 to  10 g dry wt. added to 50 ml seawater which
was constantly stirred. Cadmium chloride was
added and dissolved cadmium was monitored using
the electrode. Oxygen-free conditions were main-
tained using a nitrogen atmosphere provided by a
glove  box or by constantly bubbling nitrogen
through the covered titration vessel. In the sedi-
ment titrations where electrode response was slow,
the readings were taken after the response had
stabilized to less than 0.3 mV/min as recommended
by the manufacturer (Orion 94-48 Instruction
Manual p. 22).

   CADMIUM TOXICTTY AND INTERSTITIAL
           WATER CORRELATIONS
Dry weight normalization
   The toxicity of cadmium to Ampelisca in Long
Island Sound sediment for the one-sediment exper-
iments are shown in Figure 1 A. The results of the
three-sediment experiment using Rhepoxynius hud-
soni in Ninigret Pond sediment; and Ampelisca in
both Long  Island Sound sediment and an equal
parts mixture of the two sediments, is shown in
Figure IB. Mean control mortalities were S.0,1.7
and 16.7%, respectively. The Spearman-Karber
median LC50 estimates and 95% confidence lim-
its [15] are  listed in Table 1. The curves are log-
logistic concentration response functions fit to the
data simultaneously using the same slope param-
eter. They are included as an aid in visualizing the
data. The LCSOs range from 290 pg/g to 2,850
jjg/g on a sediment dry weight basis.. As shown be-
low these two organisms have nearly the same 96-
h cadmium activity LCSOs in water-only exposures:
17.0 fig CdI+/L for Rhepoxynius and 32.0 ng
CdJVL for Ampelisca. Therefore the differences
in the cadmium toxicity are likely to be attributable
to sediment properties affecting bioavailability. In
addition, Swartz et al. |2] reported the LC50 for
cadmium to the amphipod Rhepoxynius abronius
in a water-only exposure to be 1.6 mg Cd/L, which
would be a cadmium activity of approximately
81.0 fig Cd2VL and a sediment toxicity  of 6.9
pg/g in a Yaquina Bay sediment. Thus, the sensi-
tivity of these three test species in water differs by
less than a  factor of 5 whereas the LC50s in  four
sediments differ by greater than a factor of 400
(Table 1).  An explanation for the variation in
LCSOs in sediments would be useful.

Correlation to interstitial water concentration
   Sediments with differing toxicities on a per unit
sediment dry weight basis have been shown to have
similar toxicity based on the interstitial water con-
centrations [1-7]. In addition, the evidence sug-
gests  [16] that biological response correlates to
chemical activity, in particular to the divalent metal
activity, [Me2*) [17-19]. Figure 2 presents a com-
parison of the observed mortality to the observed

-------
                            Acid volatile sulfide and cadmium toxicity
                                                                               1491
                      MORTALITY vs SEDIMENT CADMIUM

                             DRY  WEIGHT NORMALIZATION
     too

*y    80

>    6O


1    -

S    2O

        O

     too


H    "
>    60

H    40
K
O
S    20
                   W INITIAL EXPERIMENTS

                      • U SOUND
                   IB} JOINT EXPEIUMEKT

                       • U BOUND

                       • MIXTURE

                       O NINKMET POND
                10
                        100
1000
10000
100000
                        SEDIMENT CADMIUM (ug Cd/gm dry wt)
Fig. 1. (A) Toxicity test results for Long Island Sound sediments (Ampelisca). (B) Toxicity test results for Ninigret
Pond (Rhepoxynius hudsoni), Long Island Sound and the 50/50 (v/v) mixture of the two sediments (Ampelisca).
Cadmium concentrations on a sediment dry weight basis.
interstitial water cadmium activity, measured with
the specific ion electrode, for the three sediments
examined in the paper. The water-only response
data for Ampelisca and R, hudsoni are included
for comparison although they represent a shorter
duration exposure. The curve represents the pooled
water-only data. The interstitial water concentra-
tion data from the sediment exposures are some-
what scattered. However, if the data are grouped
by decades, then the medians (50th percentile) and
interquartile ranges (25th to 75th percentiles) par-
allel the water-only exposure results as shown in
Figure 2. These results conform to previous obser-
                                        vations that the concentration response curves for
                                        sediment exposures, which are different on a sed-
                                        iment cadmium dry weight basis (Fig. 1), are com-
                                        parable on an interstitial water basis (Table 1).

                                        Sediment cadmium vs. interstitial water
                                           The prediction of the toxicity of cadmium in
                                        sediments requires that the relationship between
                                        sediment  cadmium concentration and interstitial
                                        water concentration be established. A plot of solid
                                        phase vs. aqueous phase cadmium concentra-
                                        tions—which is referred to  as an isotherm plot
                                       .when used for the analysis of sorption data—is

-------
1492
D. M. Dl TORO ET AL.
                             Table 1. LC50s for cadmium toxicity tests
Experiment
                            LC50
95% Confidence
    limits
Water only exposure (4-d exposure) (pg Cd2*/L)
  Ampelisca abdita
  Rhepoxynius hudsoni
  Rhepoxynius abronius
Interstitial water (I0-d exposure) (jig Cd2+/L)  ,
 . A. abdita and R, hudsoni*
                             17.0
                             32.0
                             81.0


                             22.0
  15.0
  27.0
  71.0
19.0
37.0
91.0
   3.0 ,  130.0
Sediments— dry weight normalization (10-d exposure) (jig Cd/g) '
Long Island Sound (A. abdita)*
Long Island Sound (A. abditaY
Mixture (A. abditaY
Ninigret Pond (R. hudsoniY
Yaquina Bay (R, abronhu)* ,.
Sedimenu-AVS normalization (10-d exposure) (janol Cd/uraol
AVS)
Long Island Sound (A. abditaY
Long Island Sound (A. abditaY
Mixture (A. abditaY
Ninigret Pond (R. hudsoniY

2,580.0
2,850.0
1,070.0
290.0
6.9


1.54
1.70
2.19
1.97

2,310.0 , 2,880.0
2.400.0,3,390.0
870.0 , UIO.O
240.0, 360.0
5.6 , 7.9


1.38, 1.72
1.44, 2.02
1.79. 2.68
1.60, 2.44
•Computed for the medians of the grouped data in Figure 2.                      .       '
•"One-sediment Long Island Sound sediment experiments (Fig. 1A).
Three-sediment experiment (Fig. IB).
"SSf2 ?«A P1' C-*> concentrations estimated from total Cd concentrations using measured ratio of [Cd**]/
 [CdJ * 0.050.                                           -                               *    f
shown in Figure 3. The data can be envisioned as
a titration in which cadmium is added incremen-
tally to the sediment and the resulting aqueous and
solid phase cadmium concentrations are measured.
          Initially, the solid phase concentration increases
          but the aqueous phase concentration remains be-
          low the detection limit of the cadmium electrode.
          A critical sediment concentration is reached at
                   MORTALITY  vs  INTERSTITIAL WATER CADMIUM
               100
          £
          >


          I
                               0.0010O      O.10000     10.0OOOO    1000.OOOOO

                               CADMIUM ACTIVITY (mg Cd2+/L)

Fig. 2. Mortality versus interstitial water cadmium activity. Medians and interquartile ranges for each decade of in-
terstitial water activity. Water only exposure data for Ampelisca and Rhepoxynius hudsoni. The line is a joint.fit to
both water only data sets.

-------
                            Acid volatile sulfide and cadmium toricity
                                                          1493
                    SEDIMENT vs INTERSTITIAL WATER CADMIUM

                100000
              I
              o» 10000
              ^
              s

              i   1000
              s
              8
                    100
                     10
• U SOUND
< LESS THAN DETECTION
                          PRCCWfTATION f    TRANSITION
                      O.OOO1        0.0100        1.0000        100.0000

                              CADMIUM ACTIVITY 
-------
 1494
                                      D. M. DlTOROETAl.
fides [13]. Perhaps the source of the sulfide is this
solid phase sulfide initially present. As cadmium is
added to the sediment it causes the solid phase iron
sulfide to dissolve releasing sulfide which is avail-
able for the formation of cadmium sulfide.' Hie
plausibility of this mechanism is examined below.

Solubility relationships and
displacement reactions

   The majority of sulfide in sediments is in the
form of iron  monosulfides (mackinawite and
greigite) and iron bisulfide (pyrite) of which the
former are the most reactive. These sulfides can be
partitioned into three broad classes that reflect the
techniques used for quantification [13,23,24]. The
most labile fraction, acid volatile sulfide (AVS), is
associated with the more soluble iron and manga-
nese monosulfides. The more resistant sulfide min-
eral phase, iron pyrite, is not soluble in the cold
acid extraction used to measure AVS. Neither is the
third compartment, organic sulfide associated with
the organic matter in sediments [25].  /
   Iron monosulfide, FcS(s), is in equilibrium with
aqueous phase sulfide'by the reaction:
                                           (2)
If cadmium is added to the aqueous phase, the re-
sult is:       '
                                           (3)
             As the cadmium concentration increases, [Cd2+] x
             [S2~] will exceed the solubility product of cad-
             mium sulfide and CdS(s) will start to form. Be-
             cause cadmium sulfide is more insoluble than iron
             monosulfide, FeS(s) should start to dissolve in re-
             sponse to the lowered sulfide concentration in the
             interstitial water. The overall reaction is:
                   Cd2+ + FeS(s) -»CdS(s) + Fe2+
                                           (4)
             The iron in FeS(s) is displaced by cadmium to
             form  soluble iron and solid cadmium sulfide,
             CdS(s). A theoretical analysis of the Cd(U)-Fe(Ii>
             S(II) system, presented in Appendix I, supports
             this conclusion. The relevant parameter, which can
             be termed the metal sulfide solubility parameter for
             any metal, Me, is ow+#Mts- It is the product of
             OMC** = IS Me(aq)]/[Me2+], the ratio of total dis-
             solved Me to the divalent species concentration;
             and KM* = (Me2+][S21. the metal sulfide solubil-
             ity product (Table 2). The sulfide solubility param-
             eters determine the behavior of FeS(s) and any
             MeS(s) as the metal is added to the sediment.
             For example because the cadmium sulfide solubil-
             ity parameter is less than the iron sulfide solubility
             parameter, cadmium will form a sulfide at the ex-
             pense of the iron sulfide which will dissolve. Note.
             that all the metals listed in Table 2 below the
             dashed line are predicted to dissolve FeS and MnS.

             Titration results—amorphous FeS
               The calculations presented in Table 2 reflect the
             chemical composition expected at thermodynamic
                      Table 2. Metal sulfide solubility and ratio of total dissolved
                                to free cation metal concentration
sulfide
                     !<>«*„.>
                                                         loga
tog*,.
pH = 7.6
Average
MnS
FeS(am)
FeS
NiS
ZnS
CdS
PbS
CuS
HgS
-0.40
-3.05
-3.64
-9.23
-9.64
-14.10
-14.67
-22.19
-38.50
-19.15
-21.80
-22.39
-27.98
-28.39
-32.85
-33.42
-40.94
-57.25
0.13
0.10
0.10
0.11
0.12
1.50
1.12
0.50
15.10
0.13
• 0.12
0.12
0.17
0.14
1.50
1.32
0.92
15.10
-19.02
-21.69
-22.28
-27.84
-28.26
-31.35
-32.20
-40.23
-42.15
       Solubility products, K^, for the reaction: Mez+ HS~ *» MeS(s) + H+ for CdS (Greenockite),
       FeS(amorphous) and Mackinawite, MnS (Alabandite) and NiS (Millerite), from [21). Solubility
       products for CuS (CoveUite), HgS (Metatinnabar), PbS (Galena) and ZnS (Wurtzitej and pK2 -
       18.57 for the reaction HS~ « H+ + S2~, from [34]. K~ b for the reaction: Me1* + S4" « MeS(s)
       is computed from log Kv 2 and pK2. Ratios of total tofree metal concentrations: a = (£Me(aq)]/
        [Me2+ ], from [35] at T = 5°C. log(aJfv) = logo + \ogKv. All logs are log,,,.

-------
                              Acid volatile sulfide and cadmium toricity
                                         1495
 equilibrium. However many solid phase reactions
 are not at equilibrium with respect to either the
 aqueous phase of other solid phases because of the
 slow kinetics involved in the necessary transforma-
 tions. Therefore, a direct test of the extent to which
 this reaction takes place was performed.
    A quantity of freshly precipitated iron sulfide
 was titrated by adding dissolved cadmium. The re-'
 suiting aqueous cadmium activity, measured with
 the cadmium electrode vs. the ratio of cadmium
 added, {Cd]A, to the amount of FeS initially pres-
 ent, [FeS(s)Ii, is shown in Figure 4. The lines con-
 necting the data points are an aid to visualizing the
 data. The electrode potentials (left) correspond to
 a low cadmium concentration during the initial
 portion of the titration. Then a sharp upward in-
 flection occurs near [Cd]A *» [FeS(s)lj indicating
 that all the iron sulfide has dissolved to form CdS.
 Any additional cadmium added appears as free
 cadmium. The plot of dissolved cadmium vs. cad-
 mium added'(right) illustrates the rapid increase
 in dissolved cadmium that occurs near [Cd]A/
 [FeS(s)]i = 1. A similar experiment has been per-
 formed for amorphous MnS with comparable re-
 sults. These displacement reactions among metal
; sulfides have been observed by other investigators
 [26]. The reaction was also postulated (27] to ex-
 plain an experimental result involving copper and
 FeS.
    These experiments demonstrate that solid phase
 amorphous iron and manganese sulfide can be dis-
 solved by adding cadmium. As a consequence it is
 a source of available sulfide which must be taken
into account in evaluating the relationship be-
tween solid phase and aqueous phase cadmium in
sediments.                        •   • •   .

Titration results—sediments
   A similar titration procedure has been used to
evaluate the behavior of sediments taken from four
different marine environments: the Long Island
Sound and Ninigret Pond sediments used in the
toricity tests; and sediments from Black Rock Har-
bor and the Hudson River (Fig. 5). The binding
capacity for cadmium, (Cd]B, is estimated by ex-
trapolating a straight line fit to the dissolved cad-
mium data. The equation is:
             = max{0,m<[Cd]A-[Cd]B)J   (5)
where [£Cd(aq)] is the total dissolved cadmium,
[Cd]A is the cadmium added, [Cd]B is the bound
cadmium, and m is the slope of the straight line.
The sediments exhibit different binding capacities
for cadmium, listed in Table 3, ranging from ap-
proximately 1 ftmol/g to more than 100 funol/g.
   The possibility  that acid volatile sulfide is a
direct measure of the solid phase sulfide that reacts
with cadmium can be examined in Table 3, which
lists the sediment binding capacity for cadmium
and the measured AVS for each sediment. The sed-
iment cadmium binding capacity appears to be
somewhat less than the initial AVS for the sedi-
ments tested. However a comparison between the
initial AVS of the sediments and that remaining af-
ter the cadmium titration is completed, Tabk 3,
                          CADMIUM TITRATION OF IRON SULFIDE
         < -so
               0.0
                             1.0
                             CADMIUM ADDED (umol Cd/umol FeS)
                  t                                        /
 Fig. 4. Cadmium titrations of amorphous FeS. Abscissa is cadmium added normalized by FeS initially present. Or-
 dinate is cadmium electrode response (left panel) and total dissolved cadmium (right panel). The lines connecting the
 data points are an aid to visualizing the data.

-------
14%
D. M. DlTOROETAl.
                            CADMIUM TITRATION OF SEDIMENTS

                                   DRY WEIGHT NORMAUZATION
                                                10.0
                                                            100.O
                                                                         1000.0
                              CADMIUM ADDED (umol Cd/gm dry wt)
Fig. 5. Cadmium tfrrarion of sediments: Black Rode Harbor, Long Island Sound. Hudson River, Ninigret Pond. Cad-
miura added per unit dry weight of sediment versus total dissolved cadmium.
suggests that some AVS is lost during titration. In
any case the covariation of sediment binding ca-
pacity and AVS is dear in the data in Table 3. This
suggests that AVS is .the proper quantification of
the solid phase sulfides that can be dissolved by
        SEDIMENT TOXIOTY AND AVS
              NORMALIZATION
   The three-sediment toxicity experiment illus-
trated in Figure IB was designed to test the utility
of AVS as a predictor of the cadmium binding ca-
pacity of sediments and therefore a predictor of the
concentration of cadmium that would cause sedi-
ment toxicity. The results are shown in Figure 6 in
which the sediment cadmium is normalized by the
          AVS for that sediment. The averages of the initial
          and final values are used for AVS. Mortality occurs
          at the point where the sediment cadmium begins to
          exceed the sediment AVS on a molar basis. Total
          mortality occurs at [Cd]/[AVSl > 3. The estimated
          LCSOs for the three sediment experiment are 1.7 to
          2.2 pmol Cd/fonol AVS (Table I).  There are no
          significant (95%) differences for  each  pair of
          LCSOs as shown by a t test.
             The critical point is that the sediment AVS can
          be used to normalize the sediment cadmium'con-
          centration in the same way that sediment organic
          carbon is  used  to normalize nonionic  organic
          chemicals. The reason that both methods work is
          that they properly account for the chemical activ-
          ity of the chemical, in both the aqueous and sedi-
                      Table 3. Cadmium binding capacity and AVS of sediments
Sediment
Black Rock Harbor "
Hudson River .
Long bland Sound0
Mixture'
Ninigret Pond'
. Initial AVS
(f/mol/g)*
175.0 (41.0 )
12.6 ( 2.8 )
15.9 ( 3.3 )
5.45( - ) ;
2.34 ( 0.73)
Final AVS
(pmol/g)11
13.9 (6.43)
3.23(1.18)
0.28 (0.12)
Cd binding
capacity
(pmol/g)
114. (12.1 )
8.58 ( 2.95)
4.57 ( 2.52)
1.12(0.42)
             •Average (standard deviation) AVS of repeated measurements of the stock.
             "Average (standard deviation) AVS after the sediment toxicity experiment.
             'From the three sediment experiment.

-------
                             Add volatile sulfide and cadmium toxirity
                                                               1497
                          MORTALITY vs SEDIMENT CADMIUM

                          ACID VOLATILE SULFIDE NORMALIZATION
          |
               100
                80
                60
                4O
                20
• uaouND

• MtXniRE

O MNMMETKND
                 0
                  0.01
        0.10
1.OO
10.00
100.00
                         SEDIMENT CADMIUM (umbl Cd' / umol AVS)
 Fig. 6. Mortality versus AVS normalized sediment ***•»"«" for Lone Island Sound, Ninigret Pond, and a 50/50 (v/v)
 mixture. The sediment cadmium and AVS are the averages of the initial and final concentrations in the control vessels.
 ment phases [8]. Below 1 jonol Cd/pmol AVS the
 cadmium is all precipitated as CdS(s) and the ac-
 tivity of cadmium is low. Above 1 ftmol Cd/pmol
 AVS there exists free cadmium in the interstitial
 water, sorbed cadmium in the sediment phase, as
'well as CdS(s). The activity of cadmium in the sys-
 tem is now high enough to cause mortality. The
 reason is that the additional cadmium added in ex-
 cess of 1 pmol Cd/fimol AVS is large enough to ex-
 ceed the activity of cadmium in the system that
 causes mortality even in the presence of some sorp-
 tion phases (see Fig. 3).
   The normalization of metal concentration by
 AVS is, of course, invalid if the AVS is zero. This
 would be the case for a fully oxidized sediment.
 For sediments with  trace amounts of AVS it is
 likely that other phases would be important as
 well. Hie results of these experiments indicate that
 the lower limit of applicability is an AVS of ap-
 proximately 1 fiinol/g (Table 3) and possibly lower.

     IMPLICATIONS FOR METAL TOXICITY
                IN SEDIMENTS
   lie first order importance of AVS in determin-
 ing the toxicity of cadmium in sediments has im-
 portant implications. These are discussed below.

 A VS in freshwater sediments
   Acid volatile sulfide is commonly found in ma-
 rine sediments (Table 4). Remarkably, it is also a
 common constituent of freshwater sediments. Its
                       presence can be rationalized as follows. Sulfide is
                       produced by the diagenesis of particulate organic
                       carbon, represented as CHjO, with sulfate as the
                       electron acceptor. [24]:
    2CH2O + SOI" -» 2CO2 + SJ

  The precipitation of iron sulfide [23]:
                                                               (10)
                                                               01)
                       forms iron monosulfide which is the majority of
                       the AVS. It might be expected that AVS is signifi-
                       cant only in marine sediments because the concen-
                       tration of sulfate in seawater is 28 mM = 2,700 mg
                       SO4/L.  By, contrast typical river water sulfate
                       concentration is 0.12 mM --11.5 mg SO4/L [28].
                       However, sedimentary organic' matter is present in
                       either locale and the sulfate in fresh water appears
                       to be sufficient to produce a significant quantity of
                       AVS. Surprisingly large values (0.31-112 pmol/g)
                       are found for sediments from the Great Lakes,
                       rivers and other freshwater lakes (Table 4). There-
                       fore the AVS concentration must be considered
                       when addressing cadmium and other metal toxic-
                       ity in freshwater sediments.

                       Application to other metals
                          The  other potentially toxic metals all form
                       metal sulfide precipitates that are more insoluble
                       than iron sulfide (Table 2). The iron and manga-

-------
 1498
D. M. DlTOROETAL.
                            Table 4. AVS in freshwater and marine sediments
Temperature1
Location (°Q
Fresh water sediments
Everglades peat basin
Lake Mendota
Lake Ontario
Lake Erie
Marine sediments
Long Island Sound
NWC
NWC
NWC
DEEP-1
FOAM-1
Sapdo Island
Mud Flat
Mud Flat
Tidal Ck.
Tidal Ck.

—
'• ' —
—
(W)


3.0
13 2
19.0
18.S
20.0

(W)
(S)
(W)
(S)
AVSfcmol/g)
depth interval
(0-1 cm)

— .
—
, 11.6
15.0


0.0
0.60
0.097
0.62
7.50

1.88
3.44
9.69
5.94
,(0-10 cm)

0.31- 1.3
8.7 -112.0
27,1
. 7.5


, 8.35
10.5
10.3
17.4
13,3

14.6
43.2
28.4
31.9
Reference

P6J
P7J
138) '
P9J

{30]

, '



[31J




               •(W) = Winter; (S) = Summer.
 nese sulfides have log(aKv) > -22.3 (above the
 dotted line) whereas the remaining sulfides have
 log(olf metal
          sulfides if their molar sum is less than the AVS.
          For this case no metal toxicity would be expected
          and:
                          [AVS]
<1
                                                    (6)
          where [MeTJi is the total cold acid extractable
          metal concentration in the sediment. Chi the other
          hand if their molar sum is greater than the AVS
          concentration then a portion of the metals with the
          lowest sulfide solubility parameters would exist as
          free metal and presumably exert a toxicity. For this
          case the following would be true:
                           {AVS)
>1
                                                    (7)
           But these two equations are the formulas used to
           determine the extent of metal toxicity in sediments
           assuming additive behavior and neglecting the ef-
           fect of partitioning. Whether the normalized sum
           is less than or greater than one discriminates be-
           tween nontoxic and toxic sediments. The additivity
           does not come from the nature of the mechanism
           that causes toxicity. Rather it results from the

-------
                             Acid volatile sulfide and cadmium toxicity
                                         1499
equal ability of these divalent metals to form metal
sulfides with the same stoichiometric ratio of Me
and S.
   This discussion is predicated on the assumption
that all the metal sulfides behave similarly to cad-
mium sulfide. In addition, it has been assumed that
only acid soluble metals are reactive enough to af-
fect the free metal activity. At present no experi-
mental data to support either of these conjectures
exists so that this discussion is purely speculative.

AVS and sediment qualify criteria
   Because AVS can bind cadmium and presumably
other metals and thereby eliminate their toxicity,
AVS will obviously play a role in the determination
of sediment quality criteria for metals. For fully
oxidized sediments with little or no AVS, AVS nor-
malization would not be appropriate. Partitioning
would be controlled by other sediment phases such
as iron and manganese oxides and organic carbon
19]-
   An estimate of when  partitioning to other
phases can be important can be made using the
proposed sediment quality criteria formula [8]:
                                          (8)
where CSQC 0*8/8) is the sediment quality criteria,
Kp is the partition coefficient (L/g), and CWQC
(jtg/L) is the chronic water quality criteria.  For the
case where there is only one metal competing for
the AVS, the molar equivalent of the AVS would
not be bioavailable. Therefore it should be added
to the allowable concentration so that:
                                          (9)
where [Cspcl is the molar sediment quality crite-
ria (jtmol/g), Jfp is the partition coefficient (L/g),
and [CWQC] is the molar chronic water quality cri-
teria (pmol/L). The range for freshwater chronic
criteria for  the metals named in Table 2  (hard-
ness = 100 mg/L) is 0.0001 to 1.6 janol/L. The
marine criteria for the same metals are 0.0001 to
0.88 fcmol/L {29]. The importance of partitioning
can be judged by comparing the product Kp x
[CWQc] t° the AVS concentration.
    Consider an AVS concentration of 1 pmol/g. If.
the partition coefficient isKp=l L/g then a metal
with a criteria concentration of 1 pmol/L would
have its sediment quality criteria doubled because
of the partitioning. For Kp = 10 L/g the criteria
concentration at which partitioning doubles the
sediment quality criteria drops to  0.1  pmol/L.
Therefore, the effect of partitioning only becomes
significant for relatively low AVS concentrations
(-1 pmol AVS/g) and for the metals with larger
partition coefficients and chronic water quality cri-
teria concentrations.

Vertical and temporal AVS profiles
   The normal method for sediment preparation in
spiked sediment toxicity tests is to produce a uni-
form distribution of chemical and sediment by
careful mixing. For these systems the AVS is uni-
formly distributed and the concentration to be used
for normalization is unambiguous.
   However, the distribution of AVS in intact sed-
iment cores exhibits both  vertical and temporal
variation over the annual cycle [30,31] (Table 4).
There is a seasonal variation in the surface concen-
tration of AVS at the Long Island Sound NWC
station. All stations exhibit a strong vertical gradi-
ent between the surface 1 cm and the average of
the top 10 cm. It appears that intact cores should
be used for sediment toxicity testing if metal tox-
icity is suspected.  Indigenous predators such as
Nephtys indsa {32]  should be eliminated, however,
perhaps by asphyxiation. This method has the ad-
vantage of not affecting the AVS concentration be-
cause anaerobic conditions are maintained.
   The vertical and temporal  variation in AVS
makes it more difficult to  decide what AVS con-
centration should be used in evaluating the poten-
tial toxicity of metals in natural sediments. This is
in contrast to the distribution of sediment organic
carbon which is more spatially uniform and tem-
porally stable. Clearly, further work is required to
understand the effect of the spatial and temporal
variability of AVS on metal  toxicity  in intact
sediments.

Sediment sampling and interstitial
water generation
   Ferrous sulfide oxidizes rapidly in aerobic envi-
ronments. For suspensions, oxidation is virtually
complete within a few hours [33]. A decline in AVS
of up to 50*70 was noted for sediments that were
held for a long period in apparently airtight con-
tainers or that were exposed to air. It is  clear,
therefore, that care should be taken to keep sedi-
ments free from exposure to oxygen before AVS
measurements or toxicity testing.
   The use of elutriates as  a surrogate for intersti-
tial water is also suspect because oxidation of metal
sulfides and release of soluble metals can occur.
Procedures for producing large volumes of "pore"
water by equilibrating suspensions of sediments

-------
 1500
D. M. Di TOKO IT At.
 must be checked for the extent of AVS oxidation
 that occurs.  •


                CONCLUSIONS

    It has been shown that AVS is the proper nor-
 malization parameter for cadmium toxicity in sed-
 iments. The amphipod LCSOs on an AVS normal-
 ized basis, [Cd]/[AVS], is the same for sediments
 with over an order of magnitude difference in dry
 weight normalized cadmium LCSOs. The correla-
 tion between mortality and interstitial water metal
 activity has also been confirmed. Although the fact
 that metals can form insoluble sulfides is well-
 known, it apparently has not been recognized that
 FeS and MnS, quantified as AVS, is a reactive pool
 of solid phase sulfide that is available to bind with
 metals which have sulfide  solubility parameters
 smaller than FeS, for example, nickel, zinc, cad-
 mium, lead, copper and mercury.
   Titrations  of amorphous FeS and MnS  with
 cadmium demonstrate that the displacement reac-
 tion, Equation 4, occurs. Further, Iterations of sed-
 iments with cadmium indicates that an  abrupt
 increase of dissolved cadmium occurs when the
 added cadmium exceeds the measured AVS. How-
 ever, these data are not as certain because AVS ap-
 pears  to be  lost during the titration and  the
 relationship is only approximate (Table 3). Never-
 theless, the AVS normalized toxicity data (Fig. 6)
 demonstrate that the normalization is quantitative.
   Surprisingly, the AVS of freshwater sediments
 is in the same range as marine sediments. There-
 fore, AVS should also be the proper normalization
 for these sediments. The normalization is invalid if
 the AVS is zero, such as for a fully, oxidized sedi-
 ment.  For sediments with trace amounts of AVS it
 is likely that other phases would be important as
 well. The experiments reported in this paper indi-
 cate that the lower limit of applicability is AVS -
 1 pmol/g and possibly lower. The other sorption
 phases are expected to be important only for sed-
 iments with smaller AVS concentrations and for
 metals with large  partition coefficients and large
 chronic water quality criteria.
   We have since determined that using the metal
 concentration in the sediment which  is simulta-
 neously liberated  by the  AVS extraction, rather
 than the total metal concentration of the sediment,
•is the correct procedure for other metals (Ni, Cu,
 Zn). A sample of the solution remaining in the re-
 action vessel after the AVS procedure is complete
 is filtered and analyzed for the metals concentra-
 tions.  For the Cd experiments reported above the
 total metal and simultaneously extracted metal
           concentrations are equivalent. The details will be
           reported subsequently.


           Acknowledgement-Has research was sponsored by, an
           EPA Cooperative Agreement CR812824-01 between
           Manhattan College and EPA Environmental Research
           Laboratory. Narragansett, RI. The assistance and encour-
           agement of Christopher Zarba, EPA Criteria and Stan-
           dards Division; Herbert Allen, Drexel University and our
           research assistants at Manhattan College: Indra Sweeney,
           Paul Morgan, Clare Sydlik, Luisa Milevoj and Christine
           Begley and at the EPA Narragansett Laboratory: De-
           borah Robson and Kathleen McKenna (SAIQ, are grate-
           fully acknowledged.
                           REFERENCES

           t. Adams, WJ., R.A. Kimerfc and R.G. Moshcr. 1985.
              Aquatic safety assessment of chemicals sorbed to
              sediments. In R.D. Cardwell, R. Purdy and R.C.
              Bahner, eds.. Aquatic Toxicology and Hazard As-
              sessment: Seventh Symposium. American Society for
              Testing and Materials, Philadelphia, PA, pp. 429-
              453.
           2. Swartz, R.C., G.R. Dilswortb, D. W. Scbnlls and
              J.O. Lambcfsop. 1985. Sediment toxicity to a marine
              infaunal amphipod: Cadmium and its interaction
              with sewage sludge. Mar. Environ. Res. 18:133-153.
           3. Muir, D.C.G., G.P. Rawn, B.E. Townseud, W.L.
              Lockhart and R. Creenluugh. 1985. Bioconceotration
              of cypermethrin, deltamethrin, fenvalerate, and per-
              methrin by Chironomus tentans larvae in sediment
              and water. Environ. Toxicol. Chem. 4:51-61.
           4. Adams, WJ. 1987. Bioavailability of neutral lipo-
              philic organic chemicals contained in sediments: A
              review. In K.L. Dickson, A.W. Maki and W.A.
              Brungs. eds.. Fate and Effects of Sediment-Bound
              Chemicals in Aquatic Systems.  Pergamon Press,
              Elmsford, NY, pp. 219-244.
           5. Kemp, P.P. and R.C. Swartz. 1988. Acute toxicity of
              interstitial and particle-bound cadmium to a marine
              infaunal amphipod. Mar. Environ. Res. 26:135-153.
           6. Nebeker, A. and G. Scboytema. 1988. DDT/Endrin
              Results. Evaluation of Carbon Normalization The-
              ory. U.S. Environmental Protection Agency Envi-
              ronmental Research Laboratory, Corvallis, OR.
           7. Swartz, R.C., D.W. Sennits, T.H. DeWitl, G.R.
              Ditsworth and l.O. Lambenon.  1990. Toxicity of
              fliioranthene in sediment to marine amphipods: A
              test of the equilibrium partitioning approach to sed-
              iment quality criteria. Environ. Toxicol. Chem. 9:
              1071-1080.
           8. Di Toro, D.M., C. Zarba, DJ. Hansen, W. Berry,
              R.C. Swartz, C.E. Cowan, S.P. Pavlou, H.E. Allen,
              N. A. Thomas and P.R. Paquin. 1991. Technical basis
              for establishing sediment quality criteria for nonionic
              organic chemicals using equilibrium partitioning. En-
              viron. Toxicot. Chem. (in press).
           9. U.S. Environmental Protection Agency. 1989. Brief-
              ing report to the EPA science advisory board on the
              equilibrium partitioning approach to generating sed-
              iment quality criteria. EPA 440/5-89-002. Office of
              Water Regulations and Standards, Criteria and Stan-
              dards Division, Washington, DC.
          10. American Society for Testing and Materials. 1989.

-------
                                Acid volatile sulfide and cadmium toxicity
                                             1501
    Proposed new standard guide for conducting solid
    phase 10-day static sediment toxitity tests with ma-
    rine and estuarinc amphipods. Draft No. 5. Philadel-
    phia, PA;
11.  Hesstein, R.H. 1976. An in situ sampler for close in-
    terval pore water studies. Limnol. Oceanogr. 21:912-
    914.
12.  Carignan, R., F. Rapin and A. Tessier. 1985. Sedi-
    ment porewater sampling for metal analysis: A com-
    parison of techniques. Geochim. Cosmochim. Acta
    49:2493-2497.
13.  Morse, J.W., FJ. Millero, J.C. Cornwdl and D.
    Rickard. 1987. The chemistry of the hydrogen sulfide
    and iron sulfide systems in natural waters. Earth Sri.
    Rev. 24:1-42.                   ;
14.  Scott, KJ. and M.S. Redmond. 1989. The effects of
    a contaminated dredged material on laboratory pop-
    ulations of the tubicolous amphipod, Ampelisco ab-
    dlta.  In U.M. Cowgill and L.R. Williams, eds.,
    Aquatic Toxicology and Hazard Assessment: Twelfth
    Symposium. American Society for Testing and Ma-
    terials, Philadelphia, PA, pp. 289-303.
15.  Hamilton, M.A., R.C. Russo and R.V. Tbnrston.
    1977. Trimmed Spearman-Karbcr method of estimat-
    ing median lethal concentrations in toxicity  bioas-
    says. Environ. Sci. Techno!. 11:714-719. Correction
    12:417(1978).
16.  Borgmann, U. 1983. Metal speciation and toxicity of
    free metal ions to aquatic biota. In J.O. Nriagu, ed.,
    Aquatic Toxicology. John Wiley & Sons, New York,
    NY, pp. 47-72.
17.  Sunda, W. and R.R.L. Guillard. 1976. The rektion-
    ship between cupric ion activity and the toxicity of
    copper to phytoplankton. J. Mar. Res. 34:511-529.
18.  Sunda, W.G., D.W. Engd and R.M. Thuotte. 1978.
    Effect of chemical speciation of toxicity of cadmium
    to grass shrimp, Palaemonetes pugio: Importance of
    free cadmium ion. Environ. Sci. Technol. 12:409-
    413.
19.  Snnda, W.G. and J.M. Lewis. 1978. Effect of com-
    plexation by natural organic ligands on the toxicity
    of copper to a unicellular alga, Monochrysis httheri.
    Limnol. Oceanogr. 23:870-876.
20.  Boulegue, J-, C J. Lord HI and T.M. Church. 1982.
    Sulfur speciation and associated trace metals (Fe, Cu)
    in the pore waters of Great Marsh, Delaware. Geo-
    cftim. Cosmochim. Acta 46:453-464.
21.  Emerson, S., L. Jacobs and B. Tebo. 1983. The be-
    havior of trace metals in marine anoxic waters: Sol-
    ubilities at the oxygen-hydrogen sulfide interface. In
    C.S. Wong, E. Boyle, K.W. Bruland and J.D. Bur-
    ton, eds., Trace Metals in Sea  Water. Plenum Press,
    New York, NY. pp. 579-608.
22. Davies-Colley, RJ., P.O. Nelson and KJ. William-
    son.  1985. Sulfide control of cadmium and copper
    concentrations in .anaerobic  estuarine sediments.
    Mar. Chem. 16:173-186.
23. Berner, R.A. 1967. Thermodynamic stability  of sed-
    imentary iron sulfldes. Am. J. Sci. 265:773-785.
24. Goldhaber, M.B. and I.R. Kaplan. 1974. The sulfur
    cycle. In E.D. Goldberg, ,ed.,  The Sea,  Vol. 5-
    Marine Chemistry. John Wiley & Sons, New York,
    NY,  pp. 569-655.
25. Landers, D.H., M.B. David and MJ. Mitchell. 1983.
    Analysis of organic and inorganic sulfur constituents
    in sediments, soils and water. Int. J. Environ. Anal.
    Chem.  14:245-256.
26. Phillips, H.O. and K.A. Kraus. 1965. Adsorption on
    inorganic materials VI. Reaction of insoluble sulfides
    with metal ions in aqueous media. J. Chromatogr.
    17:549-557.
27. Pankow, J.F. 1979. The dissolution rates and mecb7
    anisms of tetragonal ferrous sulfide (Mackinawite) in
    anoxic aqueous systems. Ph.D. thesis. California In-
    stitute of Technology, Pasadena, CA. "
28. Stnmm, W. and JJ. Morgan. 1970. Aquatic Chem-
    istry. waey-Interscience, New York, NY.
29. U.S. Environmental Protection Agency. 1986. Qual-
    ity criteria for water. EPA 440/5-86-001. Office of
    Water, Washington, DC.
30. Alter, R.C. 1980. Diagenetic processes near the sed-
    iment-water interface of Long Island Sound. I. De-
    composition and nutrient element geochemistry (S,
    N, P). In B. Saltzman, ed., Estuarine Physics and
    Chemistry: Studies in Long Maid Sound, Academic
    Press, New York, NY, pp. 238-350.
31. Reaves, C. 1984. The migration of iron and sulfur
    during the early diagenesis of marine sediments.
    Ph.D. thesis. Yale University, New Haven, CT.
32. Redmond, M.S. and KJ. Scott. 1989. Amphipod
    prcdation by the infaunal polychaete, Nephtys iadsa.
    Estuaries 12:205-207.
33. Nelson, M.B. 1978. Kinetics and mechanisms of the
    oxidation of ferrous sulfide. Ph.D. thesis. Stanford
    University, Palo Alto, CA.
34. Schoonen, M.A.A. and H.L. Barnes. 1988. An ap-
   . proximation of the second dissociation constant for
    HjS. Geochim. Cosmochim. Acta 52:649-654.
35. Byrne, R.H., L.R. Kump and K J. CantreQ. 1988.
    The influence of temperature and pH on trace metal
    speciation in seawater. Mar. Chem. 25:163-181.
36. Altscbnler, Z.S., M.M. Schnepfe, C.C.  SOber and
    F.O. Simon. 1983. Sulfur diagenesis in Everglades
    peat and origin of pyrite in coal. Science 221:221-
    227.
37. Nriagu, J.O. 1968. Sulfur metabolism and sedimen-
    tary environment: Lake Mendota, Wisconsin. Lim-
    nol. Oceanogr: 13:430-439.
38. Nriagu, J.O. and R.D. Coker. 1976. Emission of sul-
    fur from Lake Ontario sediments. Limnol. Ocean-
    ogr. 21:485-489.
39. Matisoff, G., J.B. Fisher and P.L. McCaU. 1981. Ki-
    netics of nutrient and metal release from decompos-
    ing lake sediments. Geochim. Cosmochim. Ada
    45:2333-2347.

                   APPENDIX I

          SOLUBILITY RELATIONSHIPS
             FOR METAL SULFIDES

The behavior of iron sulfide during a titration with cad-
mium can be analyzed using a simplified equilibrium
model of the Cd(II)-Fe(II)-S(II) system £28]. The mass ac-
tion laws for the sulfide solubilities are:
                         [S2-J =
                                             (12)

                                             (13)

where [Cd2+J, [Fe2+] and [S2~] are the molar concentra-
tions; -yCdz+. ?F«2+. and yst- are the activity coeffi-
cients; and KfcS and JCCds are  the sulfide.solubility

-------
   1502
                                            D. M. Di TORO ET AI.
   products. The mass balance equations for total cadmium,
   iron(IJ) and sulfidc are:
                                                        [FeS(s)J. [FeS(s)J, -
                                                                                         •FcS
                                                 (14)

                                                 (15)

             -] + [CdS(s)J + [FeS(s)J = [FeS(s)J,    (16)
  where etc,** = CSCdCaqMCd2*]. aF«*+ = [SFe(aq)]/
  [Fe2+J and ers2- = {£S(aq)]/(S2~] are the ratios of the
  total dissolved Cd, Fe(II) and S0I) to the divalent species
  concentrations, respectively. [CdS(s)] and (FeS(s)J are the
  concentrations of solid nh»w> M«tmi«i— • — • j — —-•" j
  _________ _— v. .*»«• I/IKUC raiimuini ana n*on sulfide;
  |FeS($)]j is the initial iron sulfide in the sediment and
  [Cd]A is the added cadmium.
    The solution of these equations begins with substitut-
 ing Equations 14 and 15 into Equation 16. Noting that
 <*s3-(S2~] = (SS(aq)] < (Gd]A> which states that the to-
 tal dissolved sulfide in the interstitial water is much less
 than the cadmium added, it follows that:
                                                                                                 S

                                                                                                      09)

                                                       where it has been assumed that the activity coefficients
                                                       for Cd*+ and Pe1* are equal, 70!*+ «* TFezr, because
                                                       they are both divalent cations. .
                                                          The relative magnitudes of afei+ KfeS and 00,2+ x
                                                       ACUS determines the behavior of fFeS(s)] and (CdS(s)J as
                                                       cadmium is added to the sediment. For this reason they
                                                       are termed sulfide solubility parameters. Table 2 presents
                                                       reported values. Because the cadmium solubility param-
                                                       eter is  much less than the iron sulfide solubility param-
                                                       eter, that is, aai+Kcds < «Fe2+ *RS. Equations 18 and
                                                       19 become:
                                                                     [CdS(s)J. {CdA]
                                                                                                    (20)
                                                      and:
        2
  T$2.[S  ,
                             fCd]
                                               (17)

Then substituting Equations 12, 13 and 17 into Equa-
tions 14 and IS yields the concentrations of solid phase
sulfides:
[CdS(s)J » [CdJ
.(,-
                                            \
                                            )
                                           S/
           IFeS(s)] - [FeS(s)]i - [OiAJ          (21)

 Therefore, as cadmium is added to this system odntfnpi
 sulfide forms at the expense of iron sulfide. the overall
 reaction is:       ,                                 '


         Cdl+ + FeS(s) -» CdS(s) + Fe»*        (22)

Note that if acaj+*c<»s > «Fc2+^Fes then [FeS(s)] •>
[FeS(s)]i; [CdS(s)J » 0 and no cadmium sulfide would
form.

-------

                                                           Environmental Ibxicology and Chemistry, Vol. 15, No. 12, pp. 2080-2094, 19%'
                                                                                                      Printed in the USA
                                                                                                 0730-7268/96 $6,00 + .00
 PREDICTING THE TOXICITY OF METAL-CONTAMINATED FIELD SEDIMENTS  USING
                       INTERSTITIAL CONCENTRATION OF METALS AND
                           ACID-VOLATILE SULFIDE NORMALIZATIONS
      DJ. HANSEN,*t W.J. BERRY,*f J.D. MAHONY.t W.S. BOOTHMAN.t D.M. Dl TORO,J§ D.L. ROBSON,||
                               G.T. ANKLEY.* D. MA,Tt Q .YANtt and C.E. PescHt
        tU.S. Environmental Protection Agency, National Health and Environmental Effects Laboratory, Atlantic Ecology Division,
                                    27 Tarzwell Drive. Namgansett, Rhode Island 02882
                   ^Department of Environmental Engineering, Manhattan College, Bronx, New York 10471, USA
                            {HydroQual, Inc.. 1 Lethbridge Plaza, Mahwah, New Jersey 07430, USA
                  {Rhode Island Department of Environmental Management, Providence, Rhode Island 02903, USA
                 #U.S. Environmental Protection Agency, Mid-Continent Ecology Division, 6201 Congdon Boulevard,
                                                Duluth, Minnesota 55804      •   ^
            ttPRC Oceanic Administration, Institute of Marine Environmental Protection, Dalian. People's Republic of China

                                    (Received 18 September 1995; Accepted  20 May 1996)


      Abstract—We investigated the utility of interstitial water concentrations of metals and simultaneously extracted metal/acid-volatile
      sulfide (SEM/AVS) ratios to explain the biological availability of sediment-associated divalent metals to bentbic organisms exposed
      in the laboratory to sediments from five saltwater and four freshwater locations in the United States, Canada, and China. Hie
      amphipod Ampelisca abdita or the polychaete Neanthci arenaceodentata were exposed to 70 sediments from the five saltwater
      locations, and the amphipod Hyalella azteca or the oligochaete Lumbricvlia variegatut were exposed to SS sediments from four
      freshwater locations in 10-4 lethality tests. Sediment toxicity was not related to dry weight metals concentrations. Almost complete
      absence of toxicity in spiked sediments and field sediments where metals were the only known source of contamination and where
      interstitial water toxic units' (IWTUs) were <0.5 indicates that toxicity associated with sediments having SEM/AVS ratios <1.0
      front two saltwater locations in industrial harbors was not metals-related as these sediments contained O.S and SEMMVS ratios >1.0 were toxic. Toxicity was observed less often when SEM/AVS ratios
      >1.0 (39%) or IWTUs >0.5 (55%) were used alone. The difference between the molar concentrations of SEM and AVS (SEM -
      AVS) can provide important insight into the extent of additional available binding capacity, the magnitude by which AVS binding
      has been exceeded, and, when organism response is considered, the potential magnitude of importance of other metal binding
      phases. For these reasons, SEM - AVS should be used instead of SEM/AVS ratios as a measure of metals availability, Over all
      published experiments with bom metal-spiked and field sediments. SEM - AVS and IWTUs accurately (99.2%) identified absence
      of sediment toxicity and with less accuracy (79.1%) identified the presence of toxicity.

      Keywords—Sediment     Acid-volatile sulfide    Metals   Toxicity   interstitial water
       t            INTRODUCTION
  Accurate prediction of the bioavailability of metals in sed-
iments requires mechanistic knowledge of the role of sediment
geochemistry in sediment/pore water partitioning, metal form,
and organism-mediated exposure pathways [1], In oxic sedi-
ments, metals availability for hioaccumulation is related to
organic carbon binding and complexation to iron oxyhydroxide
[2]. Other geochemical  processes also control metals avail-
ability and form in oxic sediments [3-5].
  In anoxic sediments, the availability of divalent metals to
organisms living in nearby oxic surficial sediments or.tubes
has been related to acid-volatile sulfide (AVS), principally iron
monosulfide, binding [6] and organic carbon partitioning [7].
Simultaneously extracted metal (SBM), the metal extracted by
the AVS analytical method, not total metal, is the best estimate
   * To whom correspondence may be addressed.
   Contribution 1575, U.S. Environmental Protection Agency (EPA),
National Health and Environmental Effects Research Laboratory, At-
lantic Ecology Division, Namgansett, Rhode Island.
of potentially bioavailable metal concentrations for compari-
son to AVS [8]. Acid-volatile sulfide, hence availability of
metals, can vary by season and sediment depth in response to
sulfur cycles, which are related to temperature and productivity
[9]. Sediments spiked with cadmium, copper, lead, nickel, zinc,
or mixtures of these divalent metals have been shown to con-
tain little interstitial metal and to be nontoxic to saltwater or
freshwater  snails, oligochaetes, polychaetes, or amphipods
when molar concentrations of AVS exceed molar concentra-
tions of SEM (SEM/AVS ratio £1.0) [6,8,10-13]. Toxicity
was often, but not always, observed at SEM/AVS >1.0. Fur-
thermore, field-collected saltwater [13] and freshwater [14-
15] sediments with SEM/AVS ratios £1.0 also were nontoxic.
Most laboratory-spiked  and field-collected, sediments  were
toxic when SEM/AVS ratios exceeded 1.0. Absence of toxicity
in spiked and field sediments  was associated with interstitial
metal concentrations less than those known to be toxic to tested
species in water-only tests. Toxic sediments contained inter-
stitial metal concentrations similar to concentrations toxic in
water-only tests.
                                                        2080

-------
 SEM/AVS ratio and interstitial metals: Held sediment toxicity prediction

   The objective of this article is to further demonstrate  the
 utility of interstitial water concentrations of metals and sedi-
 ment concentrations normalized on the basis of SEM/AVS
 ratios to explain the bioavailability  of sediment-associated
 metals to benthic organisms. The first part presents previously
 unpublished data on die relationship between total metal con-
 centrations,  interstitial, metal concentrations,  and SEM/AVS
 ratios and toxicity to the saltwater amphipod Ampe/uca abdita
 (which was exposed to sediments from five marine sites located
 in Maryland, Massachusetts, and New York, USA; New Bruns-
 wick, Canada; and LJaoning Province, China) together with
 previously published results using sediments from New York
 with the saltwater polycbttte Neanthes arenaceodentata [13],
 Next, data are presented on these relationships with the fresh-
 water amphipod Hyalella azteca and the oligochaete Lum-
 briculus variegatus, which were exposed  to sediments from
 four freshwater field locations. Data from locations in New
 York, Michigan, and Washington, USA, have been published
 previously [8,14-15]; those from Missouri, USA, are new. All
 are herein analyzed collectively. Finally, this article combines
 results from all experiments using field-collected saltwater and
 freshwater sediments with those from all available laboratory
 spiked-sediment tests using a variety of saltwater and fresh-
 water species [10].
                       METHODS
 Saltwater field sites
   Sediment collection, 'storage, and handling. Sediments were
 collected by plastic scoop, shovel, Ponar grab, or modified
.Van Veen grab from Jinzhou Bay, China (September 1992);
 Belledune Harbor. New Brunswick, Canada (August 1990);
 Bear Creek,  Maryland (February 1992);  a  tidal marsh near
 Fairhaven, Massachusetts (March 1991); and Foundry Cove.
 New York (August 1989) (Fig; 1). Samples  consisting of ap-
 pro*. 5 to 10 cm of surficial sediment were homogenized, and
 aliquots  removed for total metal, total organic carbon, and
 grain size analyses. Sediments were transported under ice and
 stored at 4"C in sealed glass jars with limited headspace con-
 taining nitrogen until use. Prior to conducting  toxicity tests,
 sediments were rehomogenized, taking care to limit oxidation
 of metal sulfides.
   Toxicity tests. The 10-d lethality tests with the amphipod A.
 abdita generally  followed  methodologies described by  the
 American Society for Testing and Materials [16], Di Toro et al.
 [6], and Berry et all  [10]. Those with the polychaete N.  aren-
 aceodentata are described by Pesch et al. [13]. Amphipod ex-
 posure chambers consisted of 900-ml glass canning jars, with
 a  1.3-cm-diameter overflow hole covered with 400-|un Nitex*
 mesh. Each chamber contained 200 ml of sediment and 600 ml
 of seawater. Polychaete chambers consisted of 600-ml beakers
 containing 200 ml of sediment One day before the start of the
 test, sediment from each station was placed  into each of four
 (two chemistry, days 0 and 10, and two biology) replicate ex-
 posure chambers. For each experiment with sediments from the
 five  saltwater locations, one or more  treatments consisted of
 four replicate chambers containing sediment from an uncontam-
 inated reference station in central Long Island Sound, New York,
 Narragansett  Bay, Rhode Island, USA, or an uncontaminated
 sediment from a location near the study site. Sediments from
 all stations at Foundry Cove and stations 1 to 10 at the Mas-
 sachusetts salt marsh site had interstitial salinities less than those
 tolerated by A. abdita or N. arenaceodentata; therefore, sedi-
 ments were mixed with brine, to obtain 26 to 32%o interstitial
                     Environ. Toxicol. Chem. 15, 1996    2081

 salinities prior to testing. Diffusion samplers (peepers) were
 placed  in both biology replicates and the day  10 chemistry
 replicate to sample interstitial water. Peepers consisted of 5-mi
 polyethylene vials (21 mm high, 20 mm in diameter) covered
 with a  1-jtrn polycarbonate membrane and filled with 30%«
 salinity water [10]. A plastic strap around the peeper extended
 above the sediment to facilitate recovery. To provide continuous'
 renewal of overlying water, filtered seawater (20°C, 28 to 34%o
 salinity) flowed through each replicate chamber at approx. 30*
 volume additions/d.
   Each exposure began with random placement of 20 amphi-
 pods or 15 polychaetes in the day 10 chemistry replicate and
 in the two biology replicates for each treatment. Sediment from
 the day 0 chemistry replicate was homogenized, and aliquots
 removed and frozen for AVS, SEM, and bulk metal analyses.
 Experimental chambers were checked daily for dead animals
 and water flow. Overlying water was sampled at the beginning
 of every test and at least once thereafter, with samples acidified
 and stored in vials as described  above. On day  10, peepers
 were removed from each sediment, and the water sample acid-
 ified and stored. Sediment from the day 10 chemistry replicate
 was homogenized, and aliquots removed for AVS and SEM
 analyses. Sediments from the biology replicates were sieved
 through a 0.5-mm mesh screen to quantify dead and surviving
 organisms. Samples with more titan 10% of the amphipods
 missing were recounted by a second person. Missing amphi-
 pods were assumed to be dead. For illustrative purposes, sed-
 iments were classified as toxic if mortality was greater man
 24%, as proposed by Mearas et al. [17] from results of sed-
 iment tests with the amphipod Rhepoxynius abronius. Sedi-
 ments having s24% mortality were considered nontoxic.
  Chemical  analyses. Sediment samples were analyzed for
AVS by the cold-acid purge-and-trap technique described by
Allen et al. [18], Cornwetl and Morse [19], and Boothman and
Helmstetter [20]. Simultaneously extracted metal and bulk
metals analyses were performed .using inductively  coupled
plasma emission spectrometry (ICP). For analyses of bulk met-
als, the metals were extracted from freeze-dried sediments by
ultrasonic agitation with 2 M of cold nitric acid (50 ml/5 g
wet sediment) at 60*C overnight followed by centrirugation.
Total metals  analyses of sample blanks and recoveries of
known metal additions demonstrated 85 to 100% recoveries
from sediments, 85 to 115% recoveries from sample extracts,
and an absence of contamination in our analytical procedures.
The SEM concentration reported is the sum of cadmium, cop-
per, lead, nickel, and zinc on a micromole per gram dry sed-
iment basis.  Concentrations  of all metals in sediments ex-
ceeded analytical detection limits.
  Interstitial water from peepers and overlying water  were
analyzed using ICP or graphite furnace atomic absorption spec-
troscopy. Immediately prior to sediment sampling, peepers
 were removed and rinsed to remove sediments. The water
contained within the peepers was removed by pipette and
placed in a 7-ml polyethylene vial and acidified with 50 |U of
concentrated (pH £ 1.0) nitric acid. Detection limits varied as
a function of sample size and method of analysis. Concentre- ,,
tions in water are reported as the sum of the interstitial water 11
toxic units flWTUs) of detectable metal, i.e., the sum of mi- "I
 crograms of metal per liter in interstitial water divided by the
 10-d LC50 in water-only tests (in u,g/L) for each of the five
 metals, where the 10-d LC50 for A. abdita is 36.0 ng Cd/L,
 20.5 w.g Cu/L, 3,020>g Pb/L, 2,400 |tg Ni/L, and 343 jtg
 Zn/L [10] and for N. arenaceodentata is  3,670 u.g Cd/L and

-------
2082    Environ. Toxicol. Cheat.  15, 1996
                                           D.J. Hansen et al.
                                                                                        Beltedune Harbour N.B.
                                                                       Fertilizer      ,9        «8
                       Patapsoo River
                                                                  Wbst
                                                                  Foundry
                       Foundry Cove, NY
                                                                                               Battery
                                                                                               Plant
                                                                                      .".-XJ2 Source
                                                                                                           16
                                                                                                      Foundry
                                                                                                      Brook
                                                                      100m
Fig. 1. Location of field sites and stations sampled in Jinzhou Bay, Belledune Harbor, Bear Creek, Foundry Cove, and a salt marsh in Massachusetts.
16,090 (Ag Ni/L [13]. Thus, if interstitial water is the principal
source of metals toxicity and availability of metals is the same
in water of water-only tests and interstitial water in sediment
tests. 50% mortality would be expected with sediments having
IWTOs of 1.0. In this article we use IWTUs of 0.5 to indicate
sediments unlikely to cause significant mortality because on
average water-only LCO and LC50 values differ by a factor
of approx. 2. This factor is reasonable because in saltwater
tests mortality was always absent in sediments spiked with
metals when IWTUs  were <0.5 [10]. For illustration, a con-
centration of 0.01 IWTU is used to indicate interstitial water
samples that contain no detectable metal.

Freshwater field sites
  Methods used to collect, store, and handle freshwater sed-
iments and to test sediments with the amphipod H. azteca have

-------
.  SEM/AVS ratio and interstitial metals: Field sediment toxicity prediction
                     Environ. Toxicol. Chem. IS, 1996   2083
  been described for samples from Steilacoom Lake, Washing-
  ton, and Keweenaw Watershed, Michigan,, by Ankley  et al.
  [15] and for H. azteca and the oligochaete L variegatus with
  sediments from Foundry Cove, New York, by Ankley  et al.
  [14]. These same procedures were also used with sediments
  from Turkey Creek, Missouri. General biological and chemical
  procedures, as well as the conceptual experimental design,
  were essentially  the same  for saltwater.and  freshwater tests,
  except that for freshwater tests bulk metals analyses were not
  performed  and interstitial water was extracted by centriruga-
  tion instead of diffusion samplers.

                         RESULTS
  Saltwater field sites
    Description of field sites 'and toxicity test results. Jinzhou
 , Bay is located in the northeastern quadrant of the Bohai Sea,
  China (Fig. 1). It has an area of about 150 km2, including 62
  km2 of tideflats,  with an average depth of 3.5 m [21]. A zinc
  smelter located near the mouth of the Wuli River is the largest
  source of. metals to the bay, although other industrial dis-/
  charges are significant contributors. Sediments for this  study
  were collected from seven locations along a 30-km transect
  from the mouth of the river to the northeastern portion of the
  bay. Total concentrations of divalent metals in sediments col-
  lected ranged from 261 to  21,641 u.g/g dry weight (Table 1).
  Zinc constituted  between 78.S  and 86.5% of the total.  Sedi-
  ments from stations 1 and 2 also contained low concentrations
  of polycyclic aromatic hydrocarbons (PAHs) (<12 u.g/g for
  individual PAHs),  polychlorinated biphenyls (PCBs) (<0.03 .
  M-g/g for individual congeners), and chlorinated pesticides
  (<0,03 jxg/g for  any individual pesticide). Concentrations of
  total organic carbon (TOC) ranged from 0.11 to 11.5%;  AVS,
  from 3.0 to 126  u.mol/g; SEM, from 2.9 to 374 u,mol/g; and
  SEM/AVS  ratios, from 0.51 to 8,36. The sum of the IWTUs
  for the five divalent metals ranged from no metal detected
  (<0.01) to  0.58.  The four  sediments with the highest metals
  concentrations were toxic (>24% mortality) to A.  abdita.
  However, only the most  contaminated  sediment contained
  >0.5 IWTU of metals and had an SEM/AVS ratio > 1.0, which
  suggests that metals may  not be the principal cause of the
  toxicity observed in the other three toxic sediments (Figs, 2
  and 3). (The vertical dashed lines located on these figures and
  the figures that follow at an SEM/AVS ratio of 1.0 or an IWTU
  of 0.5  indicate concentrations  below which mortality is not
  expected. The horizontal dashed line at  24% mortality indi-
  cates the approximate mortality at which statistically signifi-
  cant effects are expected.)
    Belledune Harbor, which receives outfalls from a lead smelt-
  er and fertilizer plant, is located in the southwestern portion
  of Chaleur Bay, New Brunswick, Canada (Fig, 1). Harbor
  sediments are particularly enriched, relative to adjacent areas,
  in concentrations of cadmium, lead, and zinc; other metals are
  somewhat elevated [22]. The closure of the lobster fishery due
  to the elevation  of cadmium concentrations in algae, snails,
  mussels, scallops, barnacles, crabs, and lobsters  has been of
  particular concern [23,24], Sediments for our study were col-
  lected by Ponar grab from 10 stations, seven inside and three
  outside the harbor. Total concentrations of divalent metals in
  these sediments  ranged from 277 to 2,200 u,g/g dry weight,
  with 74.7 to 93.5% of the total consisting of lead  and zinc
  (Table 1). Concentrations of TOC ranged from 0.73 to 1.62%;
  AVS, from 5.5 to 102 junol/g; SEM, from 1.9 to 18.4 umol/g;
  and SEM/AVS ratios, from 0.17 to 0.36. The sum of the IWTUs
  ranged from <0.01 to 0.58. None of the sediments were toxic
  (24% mortality) to A abdita. In contrast,
  others with similar concentrations (^ 13,800  u.g/g)  were not
  toxic (£24% mortality). Sediments with SEM/AVS ratios s 1.0
  were always nontoxic, whereas only five of 11  sediments  with
  SEM/AVS ratios >1.0 were toxic (Fig. 2). Sediments with £0.5
  IWTU were always nontoxic. those with  >2,2 IWTUs were
  always toxic, and two of seven sediments with intermediate
  IWTUs (3:0.5 to 2.2) were toxic (Fig. 3). Data on  chemical
  concentrations and JV.  arenaceodentata mortality in tests  with
  Foundry Cove sediments  are not included in Table 1 because
  they have been presented elsewhere by Pesch et al.  [13]. Six
  of 17 sediments tested with this polychaete had SEM/AVS ratios
  <1.0, 16 of 17 sediments had IWTUs <0.5, and none of the
  sediments were toxic (Figs. 2 and 3). Absence of toxicity to N.
  arenaceodentata in the 11 sediments that contained SEM/AVS
  ratios •> 1.0, five of which were  toxic to amphipods, is not
  surprising.  This polychaete is more tolerant to cadmium and
  nickel .than the amphipod and can avoid sediments containing
  toxic concentrations of these metals [13],
    Bear Creek is a tributary of the Patapsco River just east of
  Baltimore, Maryland (Fig. 1). Sediments from this portion of
  Baltimore Harbor are known to be toxic and contain'high
  concentrations of metals, PAHs, PCBs, and other substances
  [26,27] from  many municipal and industrial sources. Sedi-
  ments used in our study were collected from 14 stations using
  a  modified Van  Veen grab. Total concentrations of divalent
  metals in sediments we tested ranged from 44.2 to 2,210 jig/g
  dry weight, with zinc accounting  for approx. 75% of the  total
  concentration  (Table 1). Concentrations of TOC ranged from
  0.13 to 7.38%; silt and clay, from 4 to 99%; AVS, from  0.40
  to 304  u-mol/g;  SEM, from 0.64 to 31.0 >mol/g; and SEM/
  AVS ratios, from 0.10 to 16.7. Seven of the 14 sediments from
  Bear Creek were toxic to A. abdita; these included seven of
  the nine sediments  with  the highest dry weight metals  con-
  centrations (12.5 to 30.6 (jumol/g). Sediments mat were  non-
  toxic contained metals concentrations from 0.6 to 21.0 u.mol/g.
  Both toxic and nontoxic sediments had £0.03 IWTU of metal
  (Fig. 3). Given the absence of detectable interstitial water met-

-------
2084     Environ. Toxicol. Chem.  15, 1996
D.J. Hansen et al.
  Table 1. Summary of sediment characteristics, metals concentrations, and amphipod mortality in sediments from Jinzhou Bay, Belledunc
                        Harbor, Foundry Cove, Bear Creek, Baltimore Harbor, and a salt marsh in Massachusetts
%.
Station TOC
Jinzhou Bay
1 11.5
2 2.0
3 0.37
4 0.50
5 0.26
6 0.11
7 0.17 '
a 	
Belledune Harbor
1 0.98
2 .29
3 .08
4 .62
5 .20
6 .12
7 " .10
8 0.73
9 0.87
10 0.92
• 0.99
Foundry Cove
1 10.2
2 5.20
3 ' 13.0
4 8.13
5 ' 9.37
6 5.03
7 0.79
8 13.6
9 5.82
10 10.9
11 0.55
12 16.4
13 14.6 -
14 7.18
15 4.76
16 1.45
0.88
Bear Creek
I '7.1
2 7.38
3 5.75
4 5.47
5 6.15
6 3.32
7 0.13
8 5.19
9 4.4
10 0.17
11 4.61
12 0.16
13 4.19
14 3.14
3.89
%
Silt/-
clay

_
—
_
-
_
_
-
- .

37
68
76
68
52
42
44
23
31
38
94

-
-
_
_
-
-
—
_
-
_
-
-
—
-
_
_
94

99
97
97
96
80
58 .
7
97
93
5
98
4
94
91
87
Total divalent metals, ng/g
Cd

182
151
9.1
41
2.6
8.4
4.8
0.1

9.7
11
15
13
8.0
6.8
7.4
1.2
1.9
2.0
0.9

38,900
5,920
5,940
9,520
13,100
5,500
66
522
6,230
.19
35
163
88
363
21
10
0.4

8.8
10
5.4
4.8
3.4
4.8
0.0
5.8
4.2
0.2
2.6
0.2
Ii7
1.3
0.8
Cu

1,200
295
39 '
47
6.4
12
13
13

51
101
104
86
58
49
68
15
19
22
46

143
87
81
106
116
101
28
67
74
31
18
58
44
104
93
26
56

206
228
265
207
151
191
3.2
254
241
9.4
140
41
139
97
32
Ni

19
24
10
17
8.6
9.4
7.4
2.3

33
38
40
39
, 37
32
31
29
28
33
21

31,500
5.180
4,160
3,700
7,670
2,340
60
386
3,500
39
45
137
92
227
29
18
26

54
62
60
56
38
50
2.0
49
53
4.9
51
2.8
47
39
28
Pb

3,240
608
113
142 -
22
17
18
32

498
955
1,140
906
539
463
689
94 •
131
149
32

194
157
93
135
156
357
10
98
113
27
6.1
88
48
127
177
48
35

195
.212
209
175
162
173
3.0
250
274
9.4
162
7.4
128
88
13
Zn

17,000
4,440
737
1,320
221
300
239
67

401
783
896
874
581
503
843
137
178
192
122

403
297
278
313
356
303
80
219
246
.101
65
231
142 ,
317
234
124
160

1,576
1,700
1,140
958
566
1.000
36
1,110
978
69.
617
- 43
459
346
141
£|unol

297
77.3
12.7
22.3
3.8
5.1
4.1
1.4 '

10.0
18.9
21.7
19.9 .
13.1
11.3
17.9
,3.3
4.2
4.6
3.1

893
147
130
155
256
96.7
3.3
16.1
121
3.0
2.4
8.6
5.5 '
142
6.6
2.9
4.0

29.3
31.8
23.7
19.7
12.5
20.1
0.6
23.1
21.0
1.3
13.3
1.4
10.6
. 7.9
3.5
SEM
(j-mol/g

380
64.7
11.0
18.8
3.04
3.94
2.97
1.42

6.73
17.5
18.4
16.6
11.2
10.3
17.0
2.00
3,20
1.89
2.72

779
93.5
105
136
168
52.3
0.67
8.64
86.5
1.27
0.42
1.91
1.94
5.4
0.56
1 0.20
0.22

29.3
31.0
20.4
17.6
17.3
16.7
0.64
23.2
19.5
1-.27
. 11.9
0.74
9.96
6.79
2.85
AVS
M.mol/g

44.7
126
17.8
36.6
3.02
5.42
3.56
12.2

27.2
80.3
102
96.6
47.4
38.5
56.1
5.54,
16,7
11.3
16.8

5.60
•18.0
12.2
26.8
64.6
12.4
0.44
20.2
24.7
2.62
0.41
0.40
0.69
37.1
13.1
1.38
12.0

268
304
76.1
70.1
45.3
46.6
0.40
146
89.2
0.45
50.0
0.40
7.20-
0.40
9.75
SEM/AVS
ratio IWTU

8.5
0.52
0.62
0.51
1.0
0.73
0.84
0.11

0.25
0.22
0.18
0.17
0.24
0.27
0.30
0.36
0.19
0.17
0.16

139
5.18
8.6
5.10
2.60
4.20
1.55
0.43
3.51
0.48
1.02
4.83
2.80
0.14
0.04
0.14
0.02

0.11
0.10
0.27
0.25
0.38
0.36
1.60
0.16
0.22
2.82
0.24
1.87
1.38
16.7
0.29

- 0.58
<0,01
<0.01
0.17
0.03
<0.01

-------
SEM/AVS ratio and interstitial metals: Field sediment toxicity prediction

                                                       Table 1.. Continued
                                                                              Environ. Toiicol. Chem. 15, 1996    2085
            %    Sill/
Station     TOC   clay
                                        Total divalent metals, jig/g .
                 Cd
                  Cu
                   Ni
                     Pb
                   Zn
                           -SEM    AVS  SEM/AVS            Mor-
                           (imol/g   M-mbl/g    ratio     IWTU    tality
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
4.39
1.13
2.74
2.45
0.51
1.18
3.13
0.63
2.03
0.13
1.24
1.47
0.99
0.99
16.4
26
10.6
33.2
12.3
 4.5
12.9
27.5
 5.1
17.4
 1.5
17.2
10.9
94
94
1.7
2.3
0.7
0.0
2.6
0.6
0.8
0.4
0.5
0.5
0.2
4.0
0.5
0.1
0.1
508
572
236
639
348
184
179
 47.5
 93
 89
 30
  8.2
 15
 56
 61
42
43
20
28
28
12
12
11
 7.9
 9.5
 4.2
 3.8
 4.0
25
28
 148
 192
  63
 226
  97
  39
  52
. 28
  20
  29
   6.2
   8.1
   8.3
 • 35 -
  39
535
629
272
685
410
172
188
 86
103
152
 42
 25
 28
142
164
17.6
20.3
 8.53
22.1
12.7
 5.92
 6.17
 2.40
 3.27
 4.03
 1.21
 0.62
 0.79
 3.6
 4.1
11.5
14.0
 6.12
16.2
 7.24
 4.59
 3.74
 1.72
 2.10
 2.55
 0.73
 0.40
 0.50
 2.84
 2.55
 85.4
 19.1
 18.6
  2.35
 21.8
  5.86
 38.9
 11.7 .
 18.0
 18.0
  3.24
 12.4
  4.00
 15.5
' 16.1
0.13
0.73
0.33
6.92
0.33
0.78
0.10
0.15
0.11
0.14
0.22
0.03
0.12
0.18
0.16
0.67
1.38
0.58
0.47
0.46
0.45
0.42
0.35
0.37
0.40
0.43
0.31
0.31
0.42
0.21
18
10
 5
 8
 5
13
10
 8
 5
 3
 3
10
 2
 5
 0
' Reference sediments from Long Island Sound, lower Narragansett Bay, or a clean site nearby.
al, it is not surprising that SEM/AVS ratios for sediments from
Bear  Creek were not related to sediment toxicity; i.e., five
sediments having SEM/AVS ratios > 1.0 were not toxic, and
seven of the sediments having SEM/AVS ratios <1.0 were
toxic (Fig. 2). Most toxic sediments released visible oil sheens
when stirred, suggesting that PAHs may ultimately prove to
be a source of the observed sediment toxicity. These obser-
vations support the conclusion that toxicity observed in Bear
Creek sediments was not metal-associated.
                                                           The salt marsh, containing a small tidal creek less than 500
                                                        m long (Fig. 1), is near Fairhaven, Massachusetts, on the west-
                                                        em side of Buzzards Bay. The creek is divided by a hurricane
                                                        barrier into an upper section of low salinity and a lower section
                                                        with higher  salinity. A metal products manufacturer was the
                                                        principal source of metals in the sediments. Sediments for our
                                                        study were collected by plastic scoops from 23 locations, 10
                                                        from the upper side of the hurricane barrier and '13 from the
   JNZHOU
f
	 * 	 	
£ '

80
60
40
20
J
BEU.EDUNE<&0)

*^ * *
                                                                         0.01    0.1
                                                                                          10
                                                                                               100  0.01   0.1
                                                                                                            10   100
                                                                              FOUNDRY COVE OLa.)
80
60

40-
2U
0.1
A *

A
*
A * A • **
)1 0.1 1 10 100 10
                             ' FOUNDRY COVE QkaJ
80
40
20
0.


* *' /**J. A
)l 0.1 1 10 100 10

80
60
40
20,
n

A *
A
	 * .A

A •
                                                                                                        FOUNDRY COVE (Ra)
                                                                         0.01   0.1
                                                                                           10
                                                                                                100
80
60
40
20
nl
0.1



	 A 	
J, 1 *A *
31 0.1 1 10 10
            BEAR CREEK (ia)



%




80

60
4a
20
rt.

A A
A

A
' A :*
% °" *
                                         SALTMAPSHfA.Q.l
                                 100
                                  80

                                  60

                                  40

                                  20
ea
60
40
20
n
41
A
A
4
                                                                                                           SALT MARSH (2u£L)
80
60
40
20
n


A
A**
          0.1   i   10  100 1000   , OBI  0.1    i    10   100  1000
              SEM/AVS
                                            SEM/AVS
Fig. 2. Percent mortality of the amphipod Ampelisca abdita (A.a.)
and the polychaete Neanthes arenaceodentata {N.a.) as a function of
SEM/AVS ratios in sediments from Jinzhou Bay, Belledune Harbor,
Foundry. Cove, Bear Creek, and a salt marsh in Massachusetts.  •
                                                               0.01   0.1     1     10    100  6.01   6.1     1     10    100

                                                                   InteottHol Water Toxic Unih        tntenffld Water Toxic Unlb

                                                        Fig. 3. Percent mortality of the amphipod Ampelisca abdita (A.a.)
                                                        and the polychaete Neanthes arenaceodentata (N.a.) as a function of
                                                        IWTUs of cadmium, copper, lead, nickel, and zinc in sediments from
                                                        Jinzhou Bay. Belledune Harbor, Foundry Cove, Bear Creek, and a •
                                                        salt marsh in Massachusetts.

-------
2086    Environ. Toxicol. Chem. 15, 1996
                                           D.J. Hansen et al.
                   Table 2.  Summary of sediment characteristics, metal concentrations, and amphipod (Hyalella azteca
                     - H.a.) or oligochaete (Lumbriculus variegatus - L.V.) mortality in freshwater sediments from
                           •  Steilacoom Lake, Keweenaw Watershed, Turkey Creek, and Foundry Cove
SEM 0
Station Day 0
Steilacoom Lake
1 0.89
2 3.05
3 1.93
4 2.84
5 1.25
6 0.60
7 0.66
8 1.33
9 1.95
10 1.27
11 2.90
e _
imol/g)' AVS (iimoUg)
Day 10

1.36
2.00
1.85
3.78
3.56- ,
2.02
1.11
2.68
3.91 '
2.01
2.05
-
Day 0

4.01
2.89
0.30 .
1.94
2.06
4.16
1.48
1.60
0.39
2.17
0.65
-
Day 10

5.65
1.02
<0.02
0.92
0.29
0.81
1.44.
<0.05
<0.03
1.60
0.41
-
Average
SEM/AVS

0.23
1.51
>49.5
2.79
6.45
1.32
0.61
:>27.2
>67.5
0.93
4.11
-
IWTU*
H.a. (Lv.)

<0.22
<0.22
<0.22
<0.22 '
<0.22
<0.22
<0.22
<0.22
<0.22
<0.22
<0.22
-
% Mortality
H.a. (Lv.)

0
5
0
10
5
15
0
. 0
10
" 5-
0
0
Keweenaw Watershed .
1
2
3
4
5
6
7
Q —
9
10
11
c _
Turkey Creek
U ' -
. 2
3 ,
4
5
6
7
c _
Foundry Cove
1 789
2 66.4
3 43.8
4 92.2
5 50.1
6 176
7 0.29
8 9.23
9 92.9
10 0.31
11 0.38
12 3.02
13 2.02
14 3.34
15 1.78
16 0.01
4.68
26.4
62.6
15.1
, 19.6
5.65
8.49
1.74
28.1
0.36
10.8
_

67.2
51.5
85.4
47.6
50.1
82.1
94.5
' -

703
115
. 915
106
74.8
210
0.50
14.0
58.5
0.31
0.52
2.20
1.79
7.86
0.44
0.05
_
-
-
-
-
_
_
-
-
'—
-
- •

_
_
_
_
-
_
-
_

3.12
9.39
15.3
10.4
7.59
46.9
0.09
5.12
6.73
0.92
0.39
1.92
1.18
20.0
9.07
0.94
11.6
<0.006
0.03
0.08
0.06
0.01
0.12
0.01
0.46
0.09
0.02
-

28.1 ' •
52.2
30.1
48.4
44.2
- 38.2
78.2
-

5.65
13.8
13.6
19.0
9.83
31.2
0.10
5.20
14.7
1.17
0.16
6.15
0.64
10.0
9.92
0.49
0.40
>4,440
2,090
189
332
565
70.7
17,400
61.1
4.0
674
_

2.39
0.99
2.84
0.98
1.13
2.15
1.21
-

189
7.69
4.78
7.24,
7.11
5.25
4.11
2.25
8.90
0.32
2.11
0.97
2.25
0.31
0.12
0.05
0.41
1.19
9.97
4.52
2.81
1.71
5.45
19.6"
3.19
0.52
3.10
-

0.73
0.44
1.11
0.38
1.83
1.31
0.49


18.8 (0.50)
41.5 (1.54)
94.6 (2.44)
7.29(0.61)
4.58(0.18)
11.93(0.55)
- (3.29)
77.3 (6.49)
3.16(1.53)
2.43 (0.27)
- (0.54)
32.7 (1.35)
110. (1.13)
2.03 (3.02)
0.40 (0.32)
- (0.38)
20
55
100
100
85
75
95
100
95
35
90
10

20
15
45
0
5
20
45
5

100(87)
100(0)
100(0)
100(0)
80(0)
100(0)
100(0)
40(0)
100(24)
0(0)
100(0)
60(-)
80(0)
20(0)
0(0)
0(0)
                  • Simultaneously extracted metal is SEM copper for Steilecoom Lake and Keweenaw Watershed, SBM
                   zinc for Turkey Creek, and SEM cadmium plus nickel for Foundry Cove.  .
                  b Interstitial water toxic units are calculated using 10-d water-only LCSOs for H. azteca of 2.8 tig/L for
                   cadmium, 31 pg/L for copper, 780 u,g/L for nickel, and 436 M-g/L (hardness 330 mg/L) for zinc, and
                   for L variegatus of 158 (ig/L for cadmium and 12,200 u,g/L for nickel.
                  e Reference sediments from uncontaminated West Bearskin Lake, Minnesota, USA.
lower section. Total concentrations of divalent metals in these
sediments ranged from 82.6 to 3,320 fig/g dry weight (Table
1). Zinc and copper were the principal metals on a dry weight
basis in these sediments. Concentrations of TOG ranged from
0.13 to 4.39%; silt and clay, from 1.5 to'61.5%; AVS, from
0.44 to 419  nmol/g; SEM, from 0.73 to 31.8 iwnol/g; and
SEM/AVS ratios, from 0.10 to 6.90. Only one of 23 sediments
from the salt marsh was toxic to A. abdita (Figs. 2 and 3).

-------
SEM/AVS ratio and interstitial metals: Field sediment toxicity prediction
   100


    80


    60


    40


    20
                     10             100
                        Copper (ug/L)
1,000
Rg. 4. Toxicity of copper to Hyalella azteca versus copper concen-
trations in a water-only exposure (open symbols) and interstitial water
in sediment exposures using Keweenaw Watershed sediments (closed
symbols). (Modified from Ankley et al..[15].)         I
The SEM/AVS ratio for the toxic sediment was 4.49. and the
IWTU was 1.51. All other sediments had SEM/AVS ratios
£1.0 and IWTUs £0.64 and were nontoxic.

Freshwater field sites
  Description of field sites and toxicity test .results. High con-
centrations of copper in sediments from Steilacoom Lake, Wash-
ington, originated principally from attempts to control aquatic
vegetation using copper sulfate. Copper SEM concentrations in
sediments from  11  locations tested ranged from 0.60 to 3.91
(nmoVg (38 to 248  jig/g); AVS,  from <0.02  to 5.65 junol/g;
and SEM/AVS ratios, from 0.23 to >67.5 (Table 2) [15]. Eight
of the 11 sediments tested had  SEM/AVS ratios > 1.6.  No
copper was detected in interstitial water (IWTU < 0.22), and
no sediments were  toxic  to H. azteca. Absence of toxicity in
sediments having SEM/AVS ratios > 1.0 and the  lack of de- -
tectable copper in the interstitial water are likely consequences
of the presence of other sediment binding phases [15].
  In contrast, 10 of 11 sediments from Keweenaw Watershed,
Michigan, were toxic to H. azteca [15]. Mining-derived copper
concentrations in sediments ranged from 0.36 to  174 |unol/g
(22.9 to 11,000 |*g/g); AVS, from <0.006 to 11.6 jimol/g; and
SEM/AVS ratios, from 0.4 to  17,400 (Table 2). The one sed-
unent not toxic to amphipods  had 0.41 toxic units of copper
in interstitial wafer and an SEM/AVS ratio of  0.40. Toxic
sediments had 6.52 to 19.6 IWTUs of copper and SEM/AVS
ratios ^4.0. Acid-volatile sulfide concentrations in the 10 toxic
sediments were extremely low (<0.01 to 0.46 (imol/g) with
comparatively high copper concentrations (0.36 to 1.74 u,mol/
g); nine SEM/AVS ratios were.£61. Amphipod  mortality in
response to copper concentrations in water-only tests was sim-
ilar to amphipod mortality as  a  function of interstitial water
copper concentration in sediment tests (Fig. 4 [15]). The 10-d
LC50 (95% confidence limits) for amphipods exposed to cop-
per in water-only tests did not differ from the LC50 on the
basis  of interstitial 'dissolved copper concentrations and am-
phipod mortality from tests with  Keweenaw sediments, 31 (28
to 35) versus 28 (21 to 38) p,g/L [15].
   Sediments from Turkey Creek, Missouri, contained high and
relatively uniform concentrations of zinc (47.6 to 94.5 umol/g,
3,110 to 6,180 jtg/g) and AVS (28.1 to 78.2 junol/g, Table 2)
originating from strip mine tailings. Therefore, SEM/AVS ra-
tios (0.98 to 2.84)  and  IWTUs (0.44 to 1.83) varied  little in
the seven sediments tested. The two sediments having SEM/
                    Environ. Toxicol. Chem. 15, 1996    2087

AVS ratios £1.0 were nontoxic and had £0,44 IWTU of zinc.
The SEM/AVS ratios of the five remaining sediments ranged
from 1.13 to 2.84, and IWTUs ranged from 0.49 to 1.83. Two
of these sediments were toxic.
  Sediments from Foundry Cove, New York, tested with salt-
water A. abdita and N. arenaceodentata were also tested using
the freshwater amphipod H. azteca and the oligochaete L.
variegatus by Ankley et al. [14], Sediments contained ap-
proximately equimolar concentrations of cadmium and nickel,
with the sum of the SEM concentrations of these metals from
freshwater tests ranging from <0.01 to 789 ftrnol/g; AVS, from
0.09 to 46.9 umol/g; and SEM/AVS ratios, from 0.05 to 189
(Table 2). Four of five sediments with SEM/AVS ratios £1.0
were not toxic  to amphipods, while all sediments having SEM/
AVS ratios >1.0 were toxic to amphipods. Only the two sed-
iments with the highest SEM/AVS ratios (8.90 and 189) were
toxic to the oligochaete; 14 of 16 sediments were not toxic to
oligochaetes. Sediments with IWTUs &3.16 were toxic to am-
phipods;  when 0.40 to 2.43  IWTUs were present, no toxicity
was observed. Interstitial molar concentrations of nickel al-
most always exceeded those of cadmium by one to three orders
of magnitude (data not shown). However, cadmium was most
likely the cause of both amphipod and oligochaete mortalities
because cadmium is over 250 times more toxic than nickel to
H. azteca,  as evidenced by  the 10-d water-only LC50 of 2,8
|ig/L for cadmium and 780 jig/L for nickel. Similarly, cad-
mium is about 80 times more toxic to L. variegatus than nickel,'
with 10-d water-only LC50s of 158 |tg/L for cadmium and
12,200 (ig/L for nickel. Cadmium contributed from 88.6 to
99.9% of the total IWTUs of metals.

                      DISCUSSION
Saltwater field sites
  Bulk metals  concentrations in saltwater sediments cannot be
used to,causally relate metals concentrations to the acute re-
sponse of amphipods and polychaetes (Fig.  5). Mortality of
amphipods in  70 sediments from five saltwater locations or
polychaetes in 16 sediments from Foundry Cove was not re-
lated  to  the sum of the molar  concentrations of cadmium,
copper, lead,  nickel, and zinc on a dry weight of sediment
basis. Sediments having dry weight metals concentrations from
9.50 to 885 fLinol/g from 17 stations in Jinzhou Bay, Bear
Creek, Foundry Cove, and the marsh in Massachusetts were
toxic (>24% mortality). In contrast, dry weight metals con-
centrations from 0.20 to 885 fi.mol/g were nontoxic  (£24%
mortality), an  overlap of two to three orders of magnitude in
metals concentration.
   Normalizing metals concentrations in these sediments using
SEM/AVS ratios,  without insight into mortality caused by
co-occurring toxic substances, also does not permit accurate
causal predictions of metal toxicity in sediments from the field
(Fig. 6). Of the 59 sediments with SEM/AVS ratios £ 1.0 (Table
 1) from the five locations, 49 (83%) were not toxic, and 10
(17%) were toxic. These 10 toxic sediments were from Jinzhou
Bay and Bear Creek, both highly industrial locations. Of the
37 sediments  with SEM/AVS ratios >1.0, only seven were
. toxic. Absence of toxicity when SEM/AVS ratios are >1.0has
 commonly been observed.  However,  when SEM/AVS ratios
 are £l.O, toxicity has been observed  in only one of 92 sedi-
 ments spiked  with metals [10] and one of 15 sediments from
 freshwater field sites [ 14,15] (Table 3). For these two sediments
 the true SEM/AVS ratios may have been >1.0, as concentra-
 tions were within the precision expected in AVS and SEM

-------
2088    Environ. Toxicol. Chem. 15, 1996

                    lOOi	:	
                                                                                                       D.J. Hansen et al.
                      60
                 g
                 £
                     20
                                                10                      100
                                             Bulk Metal Qjmol/g dry wt)
                                                                                                1000
Fig. S. Percent mortality of the' ampbipod Ampelisca abdita (A.a.) and the polychaete Neanthes arenaceodentata (N.a.) in sediments from
saltwater locations in a salt marsh (A - A.a.), Belledune Harbor (*  = A.a.), Bear Creek (• = A.a.), Foundry Cove (• = A.a.; it = N.a.), and
Jinzhou Bay (+ = A.a.) as a function of the sum of the concentrations of cadmium, copper, lead, nickel, and zinc in micromoles divalent metal
per gram dry weight sediment.                                                                     •                        '
analyses and they had 0.4 and 1.4 IWTUs of metal. Given the
fact that field sediments from highly industrialized locations
contain many substances other than metals and are often toxic,
nonmetals-associated toxicity should always be suspected. If
toxic sediments have SEM/AVS ratios £1.0, we might suspect
the cause to not be metals; with SEM/AVS ratios > 1.0, toxicity
may or may not be related to metals.
  Metals concentrations expressed on a sum of the IWTU basis
(Fig. 7) can provide insight mat in  part may explain apparent
anomalies between SEM/AVS ratios and the observed toxicity
of these sediments. In spite of the presence of very high dry
weight metals concentrations, 56 of 70 sediments had <0.5
IWTU .of metal. Of the  10 toxic sediments having SEM/AVS
ratios < 1.0, none had >0.5 IWTU of metal. This suggests that
metals are unlikely to be the cause of the toxicity. Seven of
these  sediments (most, of which released oil when agitated)
                                                              were from Bear Creek, and the rest were from Jinzhou Bay.
                                                              The absence of toxicity. in many sediments having SEM/AVS
                                                              ratios > 1.0 is understandable because most (66.7%, or 12 of
                                                              IS) of these nontoxic sediments had 1.0 (one
                                                              each from Jinzhou Bay and the salt marsh and five from Foun-
                                                              dry Cove) all had >0.5 IWTU of metals. Furthermore, inter-
                                                              stitial metal concentrations are likely to overestimate the con*
                                                              centration of available metal because of differences in metal
                                                              form, greater binding to dissolved organic carbon  or ligands
                                                              in interstitial water [28], release of bound metal during sam-
                                                              pling or analytical procedures [18],  or organism avoidance of
                                                              metal exposure [13],               ,
                                                                We believe it is inappropriate to include in this article data
                                                              from locations having sediments with toxicities almost cer-
                                                              tainly not due to metals. This decision is additionally justified
                     100
                      80
                 6   60-


                j   40


                      20^
                       0.01
                              A ••
                             - A  *
                                                               A
                                                               N
                                      0.1
1
10
100
1000
                                                        SEM/AVS
Fig. 6. Percent mortality of the amphipod Ampelisca abdita (A.a.) and the polychaete Neanthes arenaceodentata (N.a.) in sediments from
saltwater field locations as a function of the ratio of the sum of the molar concentrations of cadmium, copper, lead, nickel, and zinc simultaneously
extracted (SEM) with AVS to the molar concentration of AVS (SEM/AVS ratio), (See Fig. 5 legend for definitions of symbols.)

-------
SEM/AVS ratio and interstitial metals: Field sediment toxicity' prediction
                    Environ. Toxicol. Chem. 15, 1996    2089
                   Table 3. Accuracy of prediction of the toxicity of sediments from using saltwater and freshwater
                         field locations, spiked-sediment tests, and combined field and spiked sediment tests
                            as a function of SEM/AVS ratios, IWTUs, and both SEM/AVS and IWTUs
Study
type
Saltwater field




'
Freshwater field
1


'

Laboratory spike
(freshwater and
saltwater) -
-


All





% of Sediments
Parameter
SEM/AVS

IWTU :

SEM/AVS, IWTU

SEM/AVS

IWTU
,
SEM/AVS, IWTU

SEM/AVS

IWTU

SEM/AVS. IWTU

SEM/AVS

. IWTU

SEM/AVS, IWTU

Value '
as 1.0
>1.0
<0.5
2:0.5
s;l.0,<0.5
>1.0, aO.5
£1.0
>1.0
<0.5
sO.5
=sl.O, <0.5
-> 1.0, £0.5
si.O
>1.0
<0.5
sO.5
sl.0,<0.5
> 1.0, £0.5
, sl.0 '
>1.0
<0.5
S0.5
sl.0, <0.5
>I.O, «s0.5
n
42
31
59
15
39
11
15
48
20
38
10
34
92
83
88
76
76
65
149
162
167
129
120
110
Nontoxic"
100.0
80.6
100.0
53.3
100.0
45.5
93.3
47.9
95.0
42.1
100.0
29.4
98.9
26.5
98.9
22.4 '
98.7
12.3
98.7
43.2
98.8
31.8
99.2
20.9
Toxic1
0.0
19.4
0.0
46.7
0.0
54.5
6.7
52.1
5.0
57.9
0.0
70.6
1.1
73,5
1.1
77.6
1.3
87.7
1.3
56.8
1.2
68.2
0.8
79.1
                 • Nontoxic sediments, <24% mortality; toxic sediments, >24% mortality.
because in experiments with metal-spiked sediments [10], only
one of 76 sediments having IWTUs <0.5 and SEM/AVS ratios
< 1.0 were toxic. Therefore, data from Bear Creek and Jinzhou
Bay are not included in the text, figures, and tables that follow.
These data were included above to demonstrate the value of
both SEM/AVS ratios and IWTUs in discriminating between
metals-associated and nonmetals-associated toxicity in sedi-
ments and to demonstrate the point that toxicity in field sed-
iments is common even when SEM/AVS ratios are < 1.0  but
that when this occurs it is unlikely metals-related. For the data
from saltwater  field locations in Belledune Harbor, the salt
marsh and Foundry Cove, all 42 sediments with SEM/AVS
ratios £1.0 were not toxic (Fig. 8, Table 3). Of the 31 sedi-
ments that had  SEM/AVS ratios > 1.0, only  six sediments
(from Foundry  Cove and the salt marsh) were toxic, and all
had >0.50 IWTU (Table 3). Of the 25 nontoxic sediments with
SEM/AVS  ratios >1.0, 71.4% (10  of 14) of  the sediments
tested with amphipods and 90.9% (10 of 11) of the sediments
                     100
                         I

                      80


                  8  60
                  >•

                  B -40;.
                        0.01
0.1                1
      Interstitial Water Toxic Unit
                  10
100
Fig. -7. Percent mortality of amphipod Ampelisca abdita (A.a.) and the polychaete Neanthes arenaceodentata (Ma.) as a function of IWTUs of
metals in sediments from saltwater field locations. The IWTUs are the sum of metal-specific interstitial water concentrations per 10-d LC50 for
cadmium, copper, lead, nickel, and zinc. Interstitial water concentrations with nondetectable metal are plotted at 0.01 IWTU. (See Fig. 5 legend
for definitions of symbols.)

-------
2090    Environ. Toxicol. Chem. 15, 1996
                                          D.J. Hansen el al.
luu-
80-
~ 60-
£
o 40-
o
20-



-

A
* v *vi* "V."
* * ^S*^" *A^* jj^ *
31 0.1 1 10 100 10
SEM/AVS





00

Fig. 8. Percent mortality of the amphipod Ampclisca abdita (A.a.) and the polychaete Neanthes arenaceodentata (Mo.) in sediments from three
saltwater field locations in a salt marsh (A = A.a.), Belledune Harbor (• = A.a.), and Foundry Cove (• = A.a.; it - N.a.) as a function of
the SEM/AVS ratio.
tested with polychaetes had 1.0,  31 of 79 (39.2%) were toxic (Table
3). Therefore, we believe that SEM/AVS ratios sl.O can ac-
curately predict field sediments likely to not be acutely toxic
due to metals. Use of an SEM/AVS ratio > 1.0 alone to predict





g
^
t:
2
-



IUU


80

60
40

20
4
<
<
rv
o o
i O


0.01          0.1            1            10
                    Interstitial Water Toxic Units
                                                                                100
                               1000
Fig. 9. Percent mortality of the amphipod Hyalella azteca (H.a.) and the oligochaete Lumbriculus variegalas (L.v.) as a function of interstitial
water toxic units (IWTUs) of metals in sediments from freshwater field locations in Foundry Cove (o = H.a.; * = L.V.), Steilacoom Lake {•$
= #.«.), Keweenaw Watershed {o = H.a.), and Turkey Creek (A = H.a.). Interstitial water concentrations with nondetectable metals are plotted
at 0.01 IWTU.     .                                                            .

-------
SEM/AVS ratio and interstitial metals: Field sediment toxicity prediction

                           	,—D-ODO	~.
                                                                                  Environ. Toxicol. Chem. 15, 1996    2091
                      80-
                g
                     40
                     20
                                                                    -a—

                                                                     *.
                                             A  A
                                                a
                                       D»      A
                                            60
                                             .    O            O
                                            <&0     O
                                   «WC«D-^Air0.5, 68.2% of 129 sediments
100-
80-
g 60-
>-
"o
B 40-
20
H
0.

oo
' .*• "
PB Q
e
o
a
' A
a A
••• • i' • • - • • ^ ^ ,^..-- 	 ,T.,,.
* ASlv— j Cftv* Jp • *k
3i o.i i 10 loo 1000 ia
SEM/AVS





XJO

Fig. 11. Percent mortality of the amphipods Ampelisca abdita (A.a.) and Hyalella azteca (H.a.); oligochaete Lumbriculus variegatus (L.v.); and
polychaete Neanthes arenaceodentata (N.a.) exposed to sediments from saltwater locations (solid symbols) as a function of the SEM/AVS ratio.
(See Fig. 8. and 9 legends for definitions of symbols.)

-------
2092    Environ. Toxicol. Chem. 15, 1996
                                                                                                          D.J. Hansen et al.
     100
      80
      20
                                                . ^
	*»	»'	T,'-*^	;»tVv» •

                 •'•  «   * mmm  • • • -   •
      0.001     0.01      0.1       1       10      100      1000

                   Total Metd Of SEM (jjmol/g)

Fig. 12.- Percent mortality as a function of total dry weight metals
concentrations of the oligochaete Lumbriculus variegatus (L.v.), pol-
ychaetes Capitella capitata (C.c.) and Neanthes  arenaceodentata
(Ma.); harpacticoid Amphiascus tenuiremis (A.t.); amphipods Am-
pelisca abdita (A.a.) and Hyalelia azteca (H.a.); and snails Hetisoma
sp. (H.sp.) exposed to sediments from saltwater field locations, fresh-
water field locations, and sediments spiked with individual metals or,
mixtures.
were toxic (Table 3). Given the effect on toxicity or bioavail-
ability of the presence of dissolved organic carbon or ligand-
associated metal in interstitial water, water quality (hardness
or salinity), and organism behavior,  it is not surprising that
many sediments having IWTUs >0.5 are not toxic.
  Organism response in sediments whose concentrations are
normalized on an SEM/AVS basis is consistent with metal-
sulfide binding on a mole to mole basis as first described by
                                              Di Toro et al. [6] and in recommendations for assessing the
                                              bioavailability of metals in sediments proposed by Ankley et
                                              al. [29]. Sediments spiked with metals and field sediments from
                                              saltwater and freshwater locations with SEM/AVS ratios £1.0
                                              were uniformly (98.7% of 149 sediments) nontoxic (Fig. 14,
                                              Table 3). The majority (56.8%) of 162 sediments having SEM/
                                              AVS ratios > 1.0 were toxic. Use of both IWTUs and SEM/
                                              AVS ratios did not improve the accuracy of predictions of
                                              sediments that were nontoxic (99.2% of 120 sediments; Table
                                              3). However, it is noteworthy that toxic sediments were pre-
                                              dicted with 79.1 % accuracy in 110 sediments when both SEM/
                                              AVS >1.0 and IWTUs 2:0.5 were used jointly  as decision
                                              parameters (Table 3). This approach is, therefore, very useful
                                              in identifying sediments of concern.
                                                Because AVS can bind divalent metals on  a mole to mole
                                              basis, and other metals in proportion to their molar concen-
                                              trations, normalizing concentrations of SEM in sediments from
                                              the field as the difference of SEM -  AVS instead of the con-
                                              ventional SEM/AVS ratio can provide important insight into
                                              the extent of available additional sulfide binding capacity or
                                              the extent to which AVS binding has been exceeded (Fig. 15).
                                              Furthermore, absence of organism response when AVS binding
                                              is exceeded can indicate the potential magnitude of importance
                                              that other binding  phases may have in controlling bioavail-
                                              ability. This  insight into the additional binding capacity of
                                              AVS and  other  sediment phases and the magnitude of ex-
                                              ceedance of binding are important advantages for normaliza-
                                              tion of the concentration of metals in sediments on an AVS
                                              basis over that of  interstitial water concentration. For most
                                              nontoxic  saltwater and freshwater field sediments we have
                                              tested,  1 to 100  |unol of additional metal would be required
                                              to exceed the sulfide binding capacity; i.e., SEM - AVS =
0.01
                                  0.1             1.            10            100
                                            Interstitial Water Toxic Unit
                                                                                    1000
10000
Fig. 13.' Percent mortality as a function of IWTUs of metals for saltwater and freshwater benthic species, including the oligochaete Lumbriculus
variegatus (Lv.); polychaetes Capitella capitata (C.c.) and Neanthes arenaceodentata (N.a.); harpacticoid Amphiascus tenuiremis (A.t.); am-
phipods Ampelisca abdita (A.a.) and Hyalelia azteca (H.a.); and snails Helisoma sp. (H.sp.) exposed to sediments from saltwater field locations
(salt marsh [A = A.a.}, BeUedune [» = A.a.}, and Foundry Cove [• = A.a.; * = Ma.]), freshwater field locations (Foundry Cove [D = H.a.;
A = Lv.], Steilacoom Lake [^ •> H.a.], Keweenaw Watershed [O = H.a.], and Turkey Creek [A = H.a.]), and sediments spiked with individual
metals or mixtures (Saltwater: Cd, O « A.a.; Cu, O  = A.a.; Ni, © - A.a.; Pb, O = A.a.; Zn, © = A.a.; mix, 
-------
SEM/AVS ratio and interstitial metals: Field sediment toxicity prediction

               100-
                                                                                    Environ. Toxicol Chem. 15, 1996    2093
                 0.01
1             10

      SEM/AVS
                                                                                            1000
                                                                                    10000
Fig. 14. Percent mortality as a function of the SEM/AVS ratio for saltwater and freshwater benthic species including the oligochaete Lumbriculus
variegatus {L v.); polychaetes Capitella capitata (C.c.) and Neanthes arenaceodentata (N.a.)\ amphipods Ampelisca abdita (A.a.) and Hyalella
azteca (H.O.); and snails Helisomasp. (H.s) exposed to sediments from saltwater field locations, freshwater field locations, and sediments spiked
with individual metals or mixtures. (See Fig. 13 legend for definitions of symbols.)
-1 to -100 fLmol/g. In contrast, most toxic field sediments
contained 1.0 to 1,000 junol of metals beyond the binding
capacity of sulfide alone. Data on nontoxic field sediments
whose sulfide binding capacity is exceeded (SEM - AVS is
>0.0 (Linol/g) provide the best indication of magnitude and
importance of nonsulfidic binding phases. This is particularly
true for locations such as Steilacoom Lake and the Keweenaw
Watershed, where AVS concentrations were  low, resulting in
high SEM/AVS ratios with little difference between SEM con-
                                                               centrations and sulfide binding potentials (SEM -  AVS is
                                                               numerically low, whereas SEM/AVS ratios are high). Other
                                                               field sediments we tested frequently contained 1.0 to 1,000
                                                               |unol of metal over that bound by sulfide yet were not toxic.
                                                               This indicates that the role of other sediment phases in metals
                                                               bioavailability has great significance. Therefore, accurate pre-
                                                               diction of sediments likely to cause toxicity will require es-
                                                               timates of partition coefficients and binding strengths of these
                                                               sediment phases.
                    100
                     80-
                6  60
                 O   40
                                                                                           ^
                                                                                       am
-100        -10        -i    .     o         i
                         . SEM-AVS (pmol/g dry wt)
                                                                                         100
                                                                                                   1000
Pig.  15. Percent mortality of amphipods Ampelisca abdita (A.a.) and Hyalella azteca (H.a.); oligochaete Lumbriculus variegatus (Lv.); and
polychaetes Capitella capitata (C.c.) and Neanthes arenaceodentata (N.a.) exposed to sediments from three saltwater and four freshwater field
locations as a function of the sum of the molar concentrations of SEM minus the molar concentration of AVS (SEM - AVS). The vertical
dashed line at SEM - AVS = 0.0 indicates the boundary, bet ween sulfide-bound unavailable metal and potentially available metal. (See Fig. 8
and 9. legends-for definitions of symbols.)                                                     <

-------
2094     Environ. Toxicol. Chem. 15,, 1996
                                                                                                             D.J. Hansen et al.
                        SUMMARY

  We believe that results from tests using sediments spiked with
metals and sediments from the field in locations where toxicity
is metals-associated demonstrate the value of sediment concen-
trations normalized by SEM/AVS ratio and IWTUs, instead of
dry weight metals concentrations, -in explaining the biological
availability of metals. Importantly, data from spiked sediment
tests strongly indicate that metals are not the cause of the toxicity
observed in  field sediments when both SEM/AVS ratios  are
< 1.0 and IWTUs are . Metcalfe, L.E. Burrldge and B.A.
    Waiwood. 1980. Distribution of cadmium in-marine biota in the
    vicinity of Beiledune. Cadmium pollution of Beiledune Harbor,
    New Brunswick, Canada. Can. Tech.'Rep. Fish. Aquat. Sci. 963:
    11-34.
25.  Bower, PM., HJ. Simpson, S.C. Williams and Y.H. Li. 1978.
    Heavy metals in the sediments of Foundry Cove, Cold Spring,
    NY. Environ. Sci. Technol. 12:683-687.     •
26.  Bieri, R., et al. 1982. Toxic substances. In E.G. Macalaster, D.A.
    Barker and M. Kasper, eds.. Chesapeake Bay Program Technical
    Studies:  A  Synthesis. U.S.  Environmental Protection Agency,
    Washington, DC, pp. 263-375.
27.  Pinkney, A.E., G. Gowda and E. Rzemien.  1991. Sediment
    toxicity testing of the Baltimore Harbor and C & D Canal ap-
    proach channels with the amphipod, Leptocheirus plumulosus.
    'Department of Natural  Resources, Annapolis, MD, USA.
28.  Green, A.S., G.T. Chandler and E.R. Blood. 1993. Aqueous-.
    pore-water-, and sediment-phase cadmium: Toxicity relationships
    for a  meiobenthic copepod. Environ. Toxicol. Chem.  12:1497-
    1506.
29. Ankley, G.T., N.A. Thomas, D.M. Dl Toro, D J. Hansen, J.D.
    Mahony, WJ. Berry,  R.C. Swartz and RA. Hoke. 1994. As-
    sessing potential bioavailability of metals in sediments: A pro-
    posed approach. Environ. Manage. 18:331-337.

-------

-------
                                                           Environmental Toxicology and Chemistry. Vol. IS. No. 12. pp. 2067-2079, 19%
                                                                                                       Printed in the USA
                                                                                                  0730-7268/96 $6.00 + .00
      PREDICTING THE TOXICITY OF METAL-SPIKED LABORATORY SEDIMENTS
                     USING ACID-VOLATILE SULFIDE AND INTERSTITIAL
                                        WATER NORMALIZATIONS


         W.J. BBRRY.*t DJ. HANSEN.f J.D. MAHONY.t D.L. ROBSON,§ D.M. Dl TORO,| B.P. 'SHIPLEY,*
                               B. ROGERS.ft J.M. CORBINiJ: and W.S. BOOTHMANf
    tU.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division,
                                     27 Tarzwell Drive, Narragansett. Rhode Island 02882
              ^Department of Environmental Engineering, Manhattan College, Riverdale, New York, New York 10471, USA
         {Rhode Island Department of Environmental Management, 291 Promenade Street. Providence, Rhode bland 02908. USA
                               plydroQual. 1 Lethbridge Plaza, Mahwah, New Jersey 07430, USA
                         ffSpringbom Laboratories, 790 Main Street, Wareham, Massachusetts 02571, USA
           ttScience Applications International Corporation, 165 Dean Knauss Drive, Narragansett, Rhode Island 02882, USA
     .   tflexas Natural Resource Conservation Commission, Water Quality Standards, Box 13087. Austin, Texas 78711-3087. USA

                                    (Received 18 September 1995; Accepted 2 July 1996)


     Abstract—Numerous studies have shown that dry weight concentrations of metals in sediments cannot be used to predict toxicity
     across sediments. However, several studies using sediments from both freshwater and saltwater have shown that interstitial water'
     concentration or normalizations involving acid-volatile sulflde (AVS) can be used to predict toxicity in sediments contaminated
     with cadmium, copper, nickel, lead, or zinc across a wide range of sediment types. Six separate experiments were conducted in
     which two or three sediments of varying AVS concentration were spiked with a series of concentrations of cadmium, copper, lead,
     nickel, or zinc or a mixture of four of these metals. The amphipod Ampelisca abdita was then exposed to the sediments in 10-d
     toxicity tests. Amphipod mortality was sediment dependent when plotted against dry  weight metals concentration but was not
     sediment dependent when plotted against simultaneously extracted metal (SEMVAVS or interstitial water toxic units (IWTUs).
     Sediments with SEM/AVS ratios <1.0 were seldom (2.3%) toxic (i.e.. caused >24% mortality), while sediments with SEM/AVS
     ratios >1.0 were frequently (80%) toxic. Similarly, sediments with <0.5 IWTU were seldom toxic (3.0%). while sediments with
     >O.S IWTU were toxic 94.4% of the time. These results, coupled with results from related studies, demonstrate that an understanding
     of the fundamental chemical reactions which control the availability of cadmium, copper, lead, nickel, and zinc in sediments can
     be used to explain observed biological responses. We believe that using SEM/AVS ratios and IWTUs allows for more accurate
     predictions of acute mortality, with better causal linkage to metal concentration, than is possible with sediment evaluation tools
   •  which rely on dry weight metal concentrations.

     Keywords—Acid-volatile sulfide     Interstitial water    Metals     Sediments    Toxicity
                    INTRODUCTION
  Di Tore et al.  [1] showed that the toxicity of cadmium-
spiked marine sediments was linked to metals/acid-volatile
sulfide (AVS)  ratios and interstitial water (IW) metals con-
centrations. Since .then several, studies using freshwater and
saltwater sediments spiked with cadmium, copper, lead, nickel,
and zinc [2-5] have demonstrated the utility of these param-
eters in causally Unking toxicity to metals in sediments. Kemp
and Swartz [6] maintained constant IW concentrations in cad-
mium-spiked sediments  amended with varying quantities of.
organic carbon and found mat mortality was correlated with
IW concentration but not total sediment  concentration. The
utility of metals/AVS ratios and  IW concentrations has also
been demonstrated in studies conducted at field sites contam-
inated with copper [7] and a mixture of cadmium and nickel
[3,8,9]. Two colonization experiments with cadmium-spiked
sediments, one conducted in a freshwater lake [10] and a sea-
water colonization test conducted in the laboratory [11], also
support the use of mis approach for predicting the toxicity of
   * To whom correspondence may be addressed.
   Contribution 1622, U.S. Environmental Protection Agency, Na-
tional Health and Ecological Effects Research Laboratory, Narragan-
sett, Rhode Island.
these metals in sediments. The results of these studies have
been reviewed by Ankley et al. [12], who proposed practical
sediment assessment methodologies using simultaneously ex-
tracted metal (SEM)/AVS ratios and IW metals concentrations
to evaluate cadmium, copper, nickel, lead, and zinc contami-
nation in sediments. The success of this approach for predicting
the. bioavailability of these metals in sediments is in direct
contrast to the lack of success in using  dry weight metals
concentrations for this purpose [1,3,13].
  The theoretical foundation  for  equilibrium  partitioning
(EqP) theory-based SEM/AVS predictions of metal toxicity is
that the sulfides of cadmium, copper, nickel, lead, and zinc all
have lower sulfide solubility product constants than do the
sulfides of iron and manganese, which are formed naturally in
sediments as a product of the bacterial oxidation of organic
matter [14]. As a result, these metals will displace manganese
and iron whenever they are present together with manganese
and iron monosulfides [3]. Because the solubility product con-
stants of these sulfides are small, sediments with an excess of
AVS will have very low metal activity in the IW, and no
toxicity due to these metals should be observed in the sedi-
ments.
  The results of the studies cited above were consistent with
the following predictions based on EqP theory: (1) when sed-
                                                        2067

-------
 2068   Environ. Toxicol. Chem. 15,  1996
                                           W.J. Berry et al.
 iments have a metals/AVS ratio of <1.0, sediments will not
 be toxic, and little or no metal will be present in the IW; and
 (2) when sediments have a metals/AVS ratio of >1.0, 'AVS
 binding potential will be exceeded, and metals will be present
 in the IW or available to bind with other sediment phases (i.e.,
 total organic carbon) [1]. Nontoxic sediments having metals/
 AVS  ratios >1.0 may have low IW concentrations, less than
 those known to be toxic in  water-only tests, suggesting the
 importance of additional metal binding phases in sediments
 [4,7,15]. Hie appropriate  fraction of metals to use for AVS
 normalization is referred to as simultaneously extracted metal
 (SEM). This is die metal which is extracted in the cold acid
 used in the AVS procedure. This fraction is appropriate be-
 cause some metals form sulfides which are not fully labile in
 the short time required for the AVS extraction (e.g., Ni and
 Zn) [3]. If a more rigorous extraction were used to increase
 the fraction of metal extracted which did not also capture the
 additional sulfide extracted, then the sulfide associated with
 the additional metal release would not be quantified. This
 would result in an erroneously high metal/AVS ratio [3].
   An analysis of a simple chemical equilibrium model for the
 system M(II), Fe(D), S(-II), where M(D) are the divalent met-
 als that form sulfides, shows that
where

  {M} = the activity of M in die IW,
  [M]T= die total metal concentration,
  KM = the solubility product for die metal sulfide (MS), and
     ; = the solubility  product of FeS.
The ratio AW^F* is 10-'*, 10-«>, 10-'", 10-"*, and 1Q-|W
for M = nickel, zinc, cadmium, lead, and copper, respectively
[3]. If the metals are present in excess of the sulfides (SEM/
AVS >1.0) and there are no other sediment phases capable of
binding the metals, e.g., dissolved organic carbon (DOC) or
total organic carbon (TOC), metal will be present in the IW,
and die sediment may be toxic.
  Toxicity predictions based on sulfide binding for sediments
contaminated with mixtures of metals which form insoluble
sulfides would use the sum of the molar concentration of SEM
for the divalent metals present (i.e., Cd, Cu, Ni, Pb, and Zn)
for comparison with the molar concentration of AVS in  the
sediment If the sum of the SEM  is greater than that of AVS,
metals may occur in the IW in sufficient concentrations to be
toxic. If the toxicity of the cationic metals in TW is assumed
to be additive [16], it should be possible to predict the toxicity
of the sediments in  the same  way as in the  individual metal
experiments, using the sum of the interstitial  water toxic units
(IWTUs).  The divalent metals should appear in the IW in
reverse order of the solubilities  of their sulfides [1]. Thus,
nickel should appear first in the IW in sediments with SEM/
AVS ratios slightly  >1.0, followed, by zinc, cadmium, lead,
and copper as the concentration of metals increases relative
to that of AVS.
  In this article we  summarize the available data from acute
laboratory sediment toxicity tests with spiked sediments which
can be used to test the utility of SEM/AVS ratios and IWTUs
in predicting sediment toxicity. Fust, the methodology and
results from a series of toxicity tests using saltwater sediments
spiked with cadmium, copper, lead, nickel,  or  zinc and one
test using an equimolar mixture of cadmium, copper, nickel,
and zinc are examined in detail. These tests are highlighted
 because they serve as an example from a single laboratory
 (U.S. Environmental Protection Agency, National Health and
 Environmental Effects Research Laboratory, Narragansett, RI)
 of results from a series of sediment spiking experiments with
 metals which followed a consistent methodology performed
 with a relatively.sensitive species, the amphipod Ampelisca
 abdita. Ampelisca abdita is an  estuarine, tube-building, in-
 faunal amphipod commonly used in sediment toxicity testing
 [17]. Data from the A. abdita tests with nickel and cadmium
 have been published elsewhere [1,3], but this is the first time
 the data from tests using this species with copper, lead, and
 zinc have been reported: This is also the first published report
 of a toxicity test hi which sediments spiked with a mixture of
 metals have been used to assess the usefulness of AVS and
 IW metals concentrations in the prediction of sediment tox-
 icity. Published results, from spiked sediment tests in which
 SEM/AVS and/or IW metals concentrations were measured,
 including those using polychaetes [5,9] and  copepods [4] in
 saltwater sediments and oligochaetes and snails [2] in fresh-
 water sediments, are also discussed.

     MATERIALS AND METHODS: AMPHIPOD TESTS

 Organism collection and acclimation

   Ampelisca abdita were collected from tidal flats in the Pet-
 taquamscutt (Narrow) River, a small estuary flowing into Nar-
 ragansett Bay, Rhode Island. Surface sediment containing the
 amphipods was either sieved in the field or transferred to the
 laboratory within Vt h and then sieved through a 0.5-mm mesh
 screen. In the laboratory amphipods and amphipod tubes were
 vigorously'sieved in a tub of seawater, then the sieve was
 quickly lowered into the water, and the amphipods were col-
 lected from the water surface. The amphipods were maintained
 for 3 to 7 d in the laboratory in sieved collection-site sediment
 and flowing filtered seawater in 4-L glass jars and acclimated
 to me test  temperature at the rate of 2 to 4°C/d. During ac-
 climation amphipods were fed the laboratory-cultured diatom
 Phaeodactylum tricornutum ad libitum.
   One sediment from Ninigret Pond, Charlestown, Rhode Is-
 land, used  in the cadmium experiment, was tested using the
 amphipod Rhepoxynius  hudsoni. Rhepoxynius hudsoni was
 collected using collection and acclimation methods similar to
 those used for A. abdita, except that after collection R. hudsoni'
 was washed directly from the sieve into sorting dishes.

 Water-only tests

   Ten-day static renewal tests were conducted with A. abdita
 to determine water-only LCSOs for cadmium, copper, nickel,
 lead, and zinc in seawater. Animals were exposed, unfed, to
 five concentrations of metal and a control, with two replicates
 per concentration. Amphipods were exposed in 900-ml glass
 canning jars that contained 800 ml of water.  Acclimated am-
 phipods were sieved from the holding jars, sequentially dis-
 tributed to  100-ml plastic cups (10 amphipods/cup), then ran-
 domly added to the exposure chambers. Seventy-five to 100%
 of me water in each replicate was renewed every other day,
 depending on die experiment. Water was sampled at least once
 during the test (usually twice, once near the beginning and
 once near the end of the test) to  determine die concentration
 of metal. In some experiments, aliquots from die two replicates
"were pooled prior to analysis. Exposure chambers were cov-
 ered with black plastic. The exposure chambers were checked
 daily, and amphipods which appeared dead were removed and

-------
 Predicting the toxicity of metal-spiked sediments
                    Environ. Toxicol. Chem. 15, 1996    2069
 examined under a dissecting microscope. Live animals were
 returned to the test, and dead animals were recorded and dis-
 carded.

 Spiked sediment tests
   Amphipods were exposed to control and  metal-spiked sed-
 iments in 10-d tests with continuous renewal of overlying wa-
 ter. In all experiments two sediments of different AVS con-
 centration were used: one from Ninigret Pond (AVS = 1.18-
 2.25 uM S/g) and one from Long Island Sound (AVS = 9.72-
 19.9 jxM S/g) (Table 1). In the cadmium test, a  1:1 mixture
 of these two sediments was also used (AVS = 4.34 uM S/g).
 The nominal treatments used in most experiments, expressed
 as the molar ratio of metal to  AVS, were 0.0 (control), 0.1,
 0.3,1,3, 10, and 30 (Table 1). There were four replicates per
 treatment in each test: two "biological" replicates were used
 to assess mortality, and two "chemical" replicates were used
 for metal and AVS analyses of the sediment at test initiation
 and termination. Twenty (30 in the cadmium test) amphipods
 were added to each biological replicate and  the day 10 chem-
 ical replicate at the start of the test.
  The Long  Island Sound sediment was collected from an ,
 uncontaminated site in central Long Island Sound (40°7.95'N,
 72°52.7'W) with a Srnith-Mclntyre grab sampler, returned to
 the laboratory, press-sieved wet through a 2-mm-mesh, stain-
 less-steel screen, homogenized, and stored  at 4°C. Two sep-
 arate collections of Long Island Sound sediment (LIS1 and
 L1S2) were made. The TOC for L1S1 was 0.88%, and the grain
 size composition was 5% sand, 71% silt, and 24% clay. Grain
 size data are not available for LIS2, but the  TOC  was 0.99%.
  Sediment was also collected in Ninigret Pond. The upper 5
 to 10 cm of sediment was collected with a shovel, returned to
 the laboratory, sieved wet through  a 2-mm stainless-steel
 screen, rinsed several times to remove high-organic fine par-
 ticles, homogenized, and stored at 4°C. Sediment was collected
 from two different sites in Ninigret Pond (MINI  and NTN2),
 but both sediments had TOC values of 0.15% and were 100%
 sand (NIN2 was composed mostly of sand which would pass
 through a 0.5-mm sieve, while most of the NIN1  sediment
 was retained on a 0.5-mm sieve).
  Sediments were  spiked with metal chloride or  nitrate salts
 in 1-gallon glass jars. Methods  differed slightly from experi-
 ment to experiment, but typically 1,200 ml of wet sediment
 was added to 2,000 ml of 20°C seawater that contained the
 desired weight of metal salt. The spiked sediments were stirred
 with a nylon stirrer attached to an electric  drill until homo-
 geneous, then the overlying air in each jar was replaced with
 nitrogen, and the jars were capped and rolled for 1 h. Jars of
 sediment were held at 20°C for 8 to  10 d before the start of
 the test Excess water and any precipitate was siphoned off
 the sediment surface, and the sediment rehomogenized before
 it was  added  to the exposure chambers.
  The exposure chambers were 900-ml  glass canning jars,
 each with a 1.3-cm-diameter overflow hole (covered with 400-
 u,m Kitex* mesh)  11.7 cm from the bottom of the jar. Each
jar contained 200 ml of sediment and approx. 600 ml of sea-
 water over the sediment. Each jar was covered with-an 8-cm-
 diameter glass Carolina dish with a 17-mm-diameter hole for
 the seawater delivery tube  and  air line, which consisted of a
 2-ml glass pipette.
  Diffusion  samplers (i.e., peepers) were constructed from
 polyethylene  vials (21 mm high, 20 mm diameter, 5 ml ca-
 pacity) with 1-um polycarbonate membranes in the cap [9].
 For the mixed metals experiment each diffusion sampler con-
 sisted of two vials attached back-to-back to double the volume
 of IW collected. The peeper design used in the cadmium ex-
 periment was described by Di Tore et al. [1].
  Sand-filtered Narragansett Bay water, heated to 20 ± 1°C,
 with a mean salinity of 30 ppt (28 to 34 ppt) was used in the
 experiments. Seawater flowed into each exposure chamber
 from  a distribution system consisting of chambers with self-
 priming siphons and splitter chambers. Flow rate for each ex-
 posure chamber was approx. 28 to 35 volume additions/d (ex-
 cept in the cadmium experiment, in which it was approx. 10
 volume additions/d). Exposure chambers were placed in 20°C
 water baths to maintain temperature. The exposure chambers
 were kept under constant light to help keep  the amphipods in
 the sediment, thus maximizing exposure.
  The test was started by placing a diffusion sampler in each
 exposure chamber and adding 200 ml of sediment to just cover
 the diffusion sampler. Seawater was allowed to flow through
 the chambers for 1 d. Amphipods were removed  from the
 holding containers as described above, distributed sequentially
 to I00-ml plastic cups until there were 20 amphipods/cup (30
 in the cadmium experiment), then one cup of amphipods was
 added randomly to each exposure chamber.  The seawater de-
 livery system was turned off for 1 h, and any amphipods that
had not burrowed into the sediment in that time were replaced,
 except in  those replicates where there was  an  obvious dose
response (i.e., where there were a greater than average number
 of unburrowed individuals in all replicates).
  Samples of sediment were taken from the day 0 chemistry
replicates  for metals and AVS analyses. All but about 1 cm
of overlying water was removed from each day 0 chemistry
replicate with a vacuum pump and pipette tip. The sediment
 and a small amount of remaining seawater was homogenized
with a stainless-steel spatula. Approximately  half the sediment
was placed in an acid-stripped polyethylene jar for total metals
analysis, while  the remainder was placed in a  100-ml poly-
ethylene specimen cup for AVS and SEM analysis.  Each jar
was capped, and the samples  held in the dark at 4°C until
 analysis.
  The experimental chambers were checked daily,  and am-
phipods which  appeared dead were removed and examined
under a dissecting microscope. Live animals were returned to
the test, and dead animals were recorded and discarded. The
volume of water delivered to each exposure container was
 measured before each test, and the total flow rate to the system
 was measured and adjusted daily. Temperature of the water
bath and salinity of the incoming seawater were measured
 daily.  The overlying water in each biological  replicate was .
 sampled for metals concentrations at least once near the be-
ginning and once near the end of each test. In some tests the
 samples from the two replicates were pooled. Each overlying
water sample was placed in an acid-stripped, 7-ml polyeth-
ylene  vial  and acidified with 50 \i\ of concentrated nitric acid
(pH < 1).
  At the end of the test, the diffusion samplers were carefully
removed from each replicate. Any sediment  remaining on the
cap or membrane portion of the sampler was rinsed off using
clean  seawater. The membrane was then punctured with an
 acid-stripped, 5-ml disposable pipette tip, and the'contents of
the sampler removed by pipette. The IW collected from each
diffusion sampler was added to an acid-stripped, 7-ml poly-
 ethylene vial, acidified with 50 u,l of concentrated nitric acid
 (pH < 1), and stored for metals analysis. The sediment from

-------
2070    Environ. Toxicol. Chem. 15, 1996
                                         W.J. Berry et al.'
the chemistry replicates was sampled for metals and AVS con-
tent as described for die day 0 chemistry replicates. The con-
tents of each biology replicate were sieved through a 0.5-mm
screen. Material retained on the sieve  was examined imme-
diately or preserved with Rose Bengal  stain for later sorting.
Amphipods were counted, and any missing animals were as-
sumed to have died and decomposed. Any replicate in which
10% or more amphipods were not found was recounted by
another investigator as a quality assurance check.

Chemical analyses
  Sediment samples were analyzed for AVS .by a cold-acid
purge-and-trap technique described by Di Toro et al. [1,3].
Simultaneously extracted metal analyses for the copper, nickel,
and zinc experiments were performed using graphite furnace
atomic absorption  spcctrometry (AAS). Simultaneously ex-
tracted metal analyses for the lead and the mixed metals ex-
periments were performed using inductively coupled plasma
emission spectrometry (1CP). Simultaneously extracted metal
was not measured in title cadmium experiment because the
importance of SEM versus total metal was not understood at
that time. However, cadmium does hot form sulfides which
are insoluble in the AVS procedure [3],  so acid-extractable
and SEM cadmium concentrations are interchangeable. In the
mixed-metals experiment the acid-extractable and  SEM cad-
mium  concentrations were similar (W. Berry,  unpublished
date). Simultaneously extracted metal  for only the metal(s)
under study was measured in the individual chemical exper-
iments. However, die sum of'die  SEM for all  the cationic
metals averaged only 3.2 |iM/g for unspiked Long Island
Sound sediment and 0.081 jjJvt/g for unspiked Ninigret Pond
sediment and was thus  of little importance in, die SEM/AVS
ratio for the level of metal spiking used.
  To allow for comparisons with other metal toxicity studies,
total (dry weight) metals analyses  were also performed. For
this analysis, metals  were extracted from freeze-dried sedi-
ments by ultrasonic agitation with 2 M cold nitric acid (SO ml
to 5 g wet sediment). The supernatant containing the extracted
metals was separated from die sediment residue by centrifu-
gation and analyzed by ICE
  The IW samples from the copper experiment were analyzed
for total copper using graphite furnace AAS. The IW samples
from die cadmium experiment were analyzed for Cd2* using
a cadmium ion-specific electrode. Total  cadmium was esti-
mated by multiplying die measured Cd2* by 20, which is the
ratio of total cadmium'to Cd2* in seawater [1]. All omer in-
terstitial and overlying waters were analyzed for trace metals
by ICP. The IW samples from die mixed metals experiment
were  diluted fivefold with 2 M HNO, in order to, provide
sufficient solution for analysis.

Data handling
  Ten-day LC50 values for die water-only tests were calculated
by die trimmed Spearman-Karber method [18}. Detection lim-
its were calculated for all chemical analyses on die basis of
instrument detection limits and sample size. In those instances
where a mean concentration is a summation of measured data
and data below die limit of detection,  one-half die detection
limit was used for those values below  die limit of detection.
Means for which there are no'measured values above the de-
tection limit are indicated by "n.d." in die appropriate tables
and graphs.
   For illustrative purposes, sediments which caused >24%
mortality were classified as toxic. Mearns et al. [19] found
that sediments which caused <24% mortality in tests with die
amphipod R. abronius were not consistently classified as toxic.
This criterion is similar to die "80% of control survival" cri-
terion used in die Environmental Monitoring and Assessment
Program (EMAP) program for sediment tests with A. abdita
[20]. Sediments having £ 24% mortality were classified as
nontoxic.
  Many of die interstitial and overlying water concentrations
in this article are expressed as IWTUs. An IWTU is die mea-
sured interstitial water concentration divided by the water-only
LC50 concentration for that particular compound for die test
organism. For example, a sediment with an IW concentration
equal to die water-only LCSO concentration for die test or-
ganism would have 1.0 IWTU. When more than one toxic
metal is present, IWTUs are calculated as die sum of die toxic
units of die individual metals, e.g.,

           IWTUcd+jc = (IW concn. Cd/LCSOo,)
                       +  (IW concn. Ni/LC50w)

Thus! if IW is die principal source of metal toxicity, and avail-
ability of metals is  die same in water of water-only tests and
IW in sediment tests, 50% mortality would be expected with
sediments having IWTUs of 1.0. In this article we use IWTUs
of <0.5 to indicate sediments unlikely  to cause significant
mortality because, on average, water-only LCD and LCSO val-
ues differ by a factor of approx. 2 [21] and  because die data
in our experiments support this value as a break point between
toxic and nontoxic sediments. Calculation of IWTUs was based
solely on detectable metal concentrations.

       RESULTS: SALTWATER AMPHIPOD TESTS
Water-only tests
  The LCSO values from die water-only tests are summarized
in Table 2. No 10-d water-only LCSO value for cadmium was
available for R. hudsoni (it  is very difficult to conduct 10-d
water-only tests with amphipods from this genus because of
problems with excessive control mortality), so die 10-d LCSO
value for A. abdita (36 u.g/L) was substituted for die calcu-
lation of toxic units for tins species. This substitution is rea-
sonable because these amphipods have similar sensitivities,
i.e., die 4-d Cd LCSOs for*, hudsoni (640 pg/L) and A. abdita
(340 fig/L) differed by less than a factor of 2 [1].'

Sediment chemistry
  Day 0 versus day 10 chemistry values. Acid volatile sulfide,
SEM, and  dry weight metals  sediment  chemistry measure-
ments varied somewhat from day 0 to day 10, but die variation
was generally within 20% and did not show a definite time-,
concentration-,  or  metal-dependent pattern.  Therefore, all
AVS, SEM, and dry weight metals sediment chemistry data
are reported as die mean of day 0 and day  10 value (Day 0
and day 10 values are available on request.)
  Interstitial water metal versus SEM/AVS ratio. The  IW
metal concentrations in all  experiments were usually below
die limit of detection in sediments with SEM/AVS ratios below.
1.0 (Fig. 1 and Table 1).  In the cadmium and mixed  metals
experiments, die IW metals concentrations, while appearing
large, are unknown because of die large detection limits in
these experiments.  In die cadmium experiment, one-half die
detection limit of die cadmium electrode was 1.33 jimol/L.
For die mixed metals experiment, die sum of one-half of die
detection limits of die four metals spiked in  this test was 1.S4

-------
Predicting the toxicity of metal-spiked sediments
Environ. Toxicol. Chem. 15, 1996    2071
                  Table t. Summary of AVS, metal concentrations and amphipod (A. abdita) mortality in 10-d toxicity
                         tests with sediments spiked with cadmium, copper, lead, nickel, zinc, and a mixture of
                                                 cadmium, copper, nickel, and zinc
Metal
Cadmium1

















Copper













Lead

'



,



-



Nickel







. - -


'




Zinc


Sedi-
ment
US



'

NIN

-



Mix




*
US



'


NIN






US






NIN





i
US







NIN







US

,
Nominal
SEM/AVS
Control
0.1 x '
0.3X
IX
3X
10x
Control
0.1 X
0.3X
IX
3X
10X
Control
0.1 x
0.3X
IX
3x
10X
Control
0.1 X
0.3X
IX
3x
10X
SOX
Control
0.1 X
0.3X
IX
3x
10X
30X
Control
O.IX
0.3X
IX
3x
10X
SOX
Control
O.IX
0.3X
IX
3x
10X
30X
Control
O.IX
0.3X
IX
3X
10X
30x
100X
Control
O.ix
0.3X
IX
' 3x
10X
30X
100X
Control
O.IX
0.3X
SEM
metal1
(jtmol/g)
0.00
1.57
4.85
16.7
51.7
177
0.00
0.15
0.64
2.57
5.90
' 24.3
0.00
0.30
1.75
9.64
20.7
48.4
0.27
1.00
1.57
11.6
47;0
176
306
0.00.
0.05
0.09
0.43
2.08
5.40
10.4
0.23
1.25
4.14
14.5
28.3
67.9
78.2
0.02
0.20
0.60
1.70
7.10
16.6
20.2
0.70
1.12 .
2.90
9.80
26.4
70.8
266
573
0.17
0.35
0.24
0.54
1.21
2.62
7.62
22.9
1.20
2.80
5.50
AVS
(|unol/g)
14.9
14.9
14.9
14.9
14.9
14.9
1.31
1.31
1.31
1.31
1.31
1.31
4.34
4.34
4.34
4.34
4.34
4.34
13.3
12.2
4.44
1.21
1.94
1.67
1.84
1.22
1.42
1.08
0.635
0.323
0.345'
0.63
19.9
18.6
12.8
16.4
14.9
15.5
14.2
1.20
1.92
2.23
3.10
5.75
4.08
3.37
12.6
12.0
10.6
6.17
3^89
4.01
2.70
1.90
1.93
1.88
1.84
1.02 <
0:60
0.65
0.90
0.50
11.2
11.7
13.4
SEM/AVS
ratio
0.00
0.10
' 0.33
1.12 to
3.50
11.9
0.00
0.12
-0.50
1.95
4.34
18.5
0.00
0.10
0.40
2.22
4.80
11.2
0.02
0.10
0.35
9.60
24.3
105
166
0.00 .
0.04
0.08
0.68
6.46
15.67
16.58
0.01
0.07
0.32
0.89
1.90
4.38
5.49 ,
0.02
0.09
0.26
0.60
1.24
4.08
5.97
0.05
0.10
0.30
1.60
6.80
17.7
99.1
303
. 0.10
0.20
0.13
0.53
2.02
4.02
^8.72
49.3
0.10
0.24
0.41
IWTU"
<8.3
<8.3 '
<8.3
<8.3
2,410
13,400
<8.3
<8.3 .
<8.3 -
<8.3
264
81.1
<8.3
<8.3
<8.3
<8.3
967
3,280
0.10
0.10
0.13
0.87
264
1.320
1.980
0.10
0.30
0.14
0.39
2.08
415
7,920
<0.016
<0.016
•C0.016
0.20
0.30
1.10
18.4
<0.02
0.02
<0.02
0.03
0.21
2.50
43.3
<0.02
<0.02
<0.02
<0.02
9.44
164
844
2,980
<0.02
<0.02
<0.02
- 0.06
1.50
55.2
272
763
0.12
0.10
0.10
%
Mortality
1.65
8.35
16.7
10
100
88.4
. 5
12.5
12.5
40
95
100
16.7
11.7
23.4
46.7
100
85
12.5
7.5
17.5
100
100
100
100
22.5
" 5
15
30
100
100
.100
10
5
123
7.5
22.5
V 42.5
100
10
17.5
15
5
17.5
55
92.5
0
. 7.5
2.5
10
95
100.
100
100
. 5
7.5
'2.5
7.5
2.5
97.5 -
100
97.5
15
7.5
17.5

-------
 2072   environ. Toxicol. Chem. 15. 1996
                                                                                      WJ. Bony et al.
                                                   Table 1. Continued.
Metal











Mixture
(Cd, Cu, Ni, Zn)
'




-




Sedi- Nominal
ment SEM/AVS
IX
3X
10X
30X
NIN Control
O.ix
0.3X
IX
3X
,10x
30x
L1S Control
O.IX
0.3X
IX
3X
10X
NIN Control
O.ix
0.3X
IX
3x
10X
SEM
metal*
(iunol/g)
20.3
74.3
155
140
0.01
0.30
0.70
1.50
2.00
4.13
8.82
3.40
4.60
8.85
19.6
53.8
116-
0.10
0.21
0.30
0.62
2.44
3.00
AVS
(lunol/g]
15.1
18.2
15.0
14.0
2.25 .
2.45
3.00
2.73
1.82
1.31
1.94
9.72
11.2
6.93
3.35
1.72
0.21
1.99
1.34
1.43
0.55
0.90
0.09
SEM/AVS
I ratio
1.34
4.09
10.3
9.96.
0.00
0.11
0.23
0.54
1.09
3.15
4;54
0.35
0.41
1.30
5.90
31.3
538
0.04
0.15
0.20
1.13
2.82
33.1
IWTU*
0.02
10.2
345
8.280
<0.014
<0.014
<0.014
0.03 '
23.3
755
2.910
<3.2
<3.2
<3.2 '
1.91
142
16,200
<3.2
<3.2
<3.2-
5.53
23.7
2,010
*;
Mortality
15
77.5
100
100
5
12.5
12.5
5
35
95
100
. 5
2.5
2.5
15
100
100
5
17.5
5
22.5
30
100
                 • Simultaneously extracted metal concentrations repotted are only for die meul(s) spiked in the particular
                  experiment.
                 * Interstitial water toxic unite were calculated using the water-only LCSOs in Table 2.
                 ' Simultaneously extracted metal concentrations were not available for the cadmium experiment, so bulk
                  measurements for cadmium were substituted.
junol/L. Above an SEM/AVS ratio of 1.0 the IW concentration
increased up to five orders of magnitude with increasing SEM/
AVS ratio (Table 1). In each experiment there were usually
one or more  sediments with SEM/AVS ratios slightly >1.0
with IW concentrations below or near detection limits. The
presence of low IW metals concentrations in sediments with
SEM/AVS ratios >I.O may in part be due to analytical vari-
ability but is also due to the presence of other binding phases.
In some sediments spiked with copper, nickel, and a mixture
Of metals, AVS decreased with increasing metals concentration
(Table 1), presumably due to the formation of copper and
nickel sulfides not completely soluble in the short tune required
for the AVS extraction. This apparent decrease in AVS is an-
other example of the importance of using SEM (as opposed
to total  metal) in the calculation of metals/AVS ratios.  ,
  The relationship between IW metal concentration and SEM/
AVS ratio in the mixed metals experiment was similar to that
in.the individual metal experiments when the molar concen-
trations of an the metals are summed (Fig. I). Further insight
into the partitioning of the metals in the IW from  the mixed
metals experiments can be gained by plotting the IW concen-
trations for each individual metal in mis experiment (Fig. 2).


     Table 2. Ten-day water-only LC50 values for A. abdita
Metal
LC50
95% Confidence limits
Cadmium
Copper
, Lead
Nickel
Zinc
36.0
20.5
3,020
2,400
343
Not reliable
16:5-25.5
1.980-4,610
2.050-2,820
291-405
                                           In the Long Island Sound sediment, all four metals were below
                                           the limit of detection in treatments with SEM/AVS ratios of
                                           1.25 or lower (Fig. 2). As the SEM/AVS ratio of the treatments
                                           increased, detectable concentrations of metal began to appear.
                                           The most soluble sulfide (Ni) appeared first and at the highest
                                           IW concentration. As SEM/AVS ratios increased, the other
                                           metals appeared in the order of their solubility product con-
                                           stants. The metal with the least soluble sulfide (Cu) appeared
                                           last and at the lowest concentration. The relationship between
IUUUUULT
100000

10000'
1000
100
10
]i

0.1-
n.m-

*
• **
"** A +
V $ +
'* *
l'*L V
*h-
2f~"* +«%Hdrfr4|^«
i* *JRwyrB# « 4 .
&* »* A* «* •
*•*• * *' "*
                                                    0.001  •  0.01
                                                  0.1
                                                                                  10
                                              100
1000
                                                                                          SEM/AVS
Fig; 1. Interstitial water metals concentration (jxM/L) as a function
of SEM/AVS ratio. Data from the mixed metals experiment represent
the molar sum of cadmium, copper, nickel, and zinc. All experiments
combined. Data below the IW detection limit are plotted at one-half
the detection limit, indicated by arrows. All data in the copper ex-
periment were above the limit of detection. Data below the SEM
detection limit are plotted at SEM/AVS = 0.001. • = Cd; A = Cu;
* = Ni; 4  =  Pb; • = Zn; and + = mixed metals.

-------
 Predicting the toxicity of metal-spiked sediments

 g 10000

 I  1000

 •8   100


 I   •"


 1   0.1
    0.01
      0.01
                                                    1000
   10000
    100°
     10°
      10


     0.1
    0.01
       0.01
0.1
                          1
10
                                           100
                                    1000
                           SEM/AVS
 Fig. 2. Individual IW metals (fiM/L) concentration in the mixed met-
 als experiment as a function of SEM/AVS ratio. The top panel rep-
 resents data from the Long Island Sound sediment, the bottom panel
 data from the Ninigret Pond sediment. Data below the IW detection
 limit are plotted at one-half the detection limit, indicated by arrows.
 • = Cd; A = Cu; * = Ni; and • *= Zn.
 IW concentration and SEM/AVS ratio in the Ninigret Pond
 sediments was similar to that in the Long Island Sound sed-
 iments (Fig. 2). In die sediment treatments with SEM/AVS
 ratios of <1.0 there was no detectable metal in the IW. In one
 sediment with an SEM/AVS ratio slightly >1.0  (1.12) there
 were small, but measurable, zinc and cadmium concentrations
 in the IW. In the sediment treatment with the next higher SEM/
 AVS ratio, there was measurable nickel, zinc, and cadmium,
 with the metal concentrations decreasing in that order. Only
 in the sediment with the highest SEM/AVS ratio was mea-
 surable copper found in the IW.

 Sediment toxicity
   Hie mortality of ampbipods as a function of  total metals
 concentrations followed a similar pattern in each of the five
 individual metal and the mixed metals toxicity tests. Mortality
 appeared sediment dependent when plotted on a total metals
 basis (Fig. 3). Mortality increased with increasing metals con-
 centration  (ng/g dry weight) for each sediment, but in each
 experiment there were treatments in low AVS sediments (Ni-
 nigret Pond) which caused  100% mortality at dry weight met-
 als concentrations which did not cause appreciable mortality
 in treatments-from the high AVS  sediment (Long Island
 Sound). Thus, although mortality is concentration dependent
. for each sediment, the concentration-response curves do not
 overlap. Therefore, it is not possible to predict amphipod mor-
 tality in different sediments on the basis of total metals con-
 centrations alone (Figs. 3 and 4a).
                     Environ. Toxicot. Chent. 15, 1996    207i

   Mortality did not appear to be metal specific when plotted
 on a molar dry weight metal basis (Fig. 4a).  Some factor, in
 the sediment appeared to affect the toxicity of all five metals
 similarly because within a sediment the results for  all'five
 metals were very similar (Fig. 4b).
   Mortality in the individual and mixed metals experiments
 was sediment independent when plotted on an SEM/AVS ratio
 basis (Fig. 5). Sediments with an SEM/AVS ratio < 1.0 did
 not cause mortality significantly different from the control (i.e.,
 >24%). In sediments with an SEM/AVS ratio >I.O, mortality
 increased with increasing SEM/AVS ratio, although  in each
 experiment there were usually one or two sediments with SEM/
 AVS ratios slightly >1.0, and in one instance 5.8  (Fig. 5f),
 which did not cause significant mortality. This indicates that
 there are other binding phases in the sediment. Thus, it is
 possible  to predict with accuracy which sediments  will be
 nontoxic (cause <24% amphipod mortality) and, with slightly
 less accuracy,  which sediments will be toxic (cause >24%
. amphipod mortality). When the results of all of the experiments
 are plotted together the mortality of amphipods as a function
 of SEM/AVS ratio appears to be metal and sediment indepen-
 dent for the five individual metals and for a mixture of metals
 (Fig. 6).
   Mortality was also not sediment or metal specific when plot-
 ted against IWTUs. Little amphipod mortality occurred in sed-
 iments with IWTUs of <0.5 (Fig. 7).  (The data from those
 sediments in which IW cadmium was not detected in the cad-
 mium experiment and the data from the mixed metals expcr
 iment in which IW metal was not detected are not included
 in Figs. 7 and 9 and Table 3 because the detection limits in
 these experiments were greater than a toxic unit, making these
 data uninterpretable.) Amphipod mortality was higher in sed-
 iments with >0.5 IWTU and increased with increasing IWTU
 value. As was the case when mortality was  plotted  against
 SEM/AVS and ratios exceeded 1.0, there were several sedi-
 ments with IWTU values >O.S which did not cause mortality.
This was especially true in the range of IWTU values >0.5
 but <10,0 (Fig. 7). This may be due in part to the variability
 inherent in water-only tests but may also indicate that not all
 of the IW metal is bioavailable. Thus, for both SEM/AVS ratios
 and IWTUs, sediments not likely to cause amphipod mortality'
 can be predicted with near certainty, but predicting which sed-
 iments are likely to cause amphipod mortality is less accurate.
 The lack of metal specificity and the results of the mixed metals
 experiment ( Fig. 7 and Table  1) indicate that the sum of the
 IWTUs can be used to make predictions about amphipod mor
 tality of any combination of metals tested in these experiments.
   When the results of the individual metals and the mixed.
 metals tests are taken together, 97.7% of the 43 sediments with
 SEM/AVS ratios <1.0 were not toxic (i.e., caused mortality
 <24%).  Of  the  45 sediments with SEM/AVS ratios > 1.0,
 80.0% were toxic (i.e., caused mortality s=24%). Ninety-seven
 percent of the 33 sediments with an IWTU value of <0.5 were
 not toxic, while 94.4% of the 36 sediments  with an IWTU
 value of >O.S were  toxic. When both SEM/AVS  ratio and
 IWTU were combined, the predictive ability was not improved
 over the use of IWTU alone. Of 29 sediments with SEM/AVS
 ratios <1.0 and IWTUs <0.5, 96.6% were nontoxic, while
 94.4% of 36 sediments with SEM/AVS ratios >1.0 and IWTUs
 >0.5 were toxic (Table 3).
                      DISCUSSION
   Our results demonstrate that it is not possible to predict the
 toxicity of sediments spiked with metals using the total metal

-------
 2074    Environ. Toxicol. Chem. 15, 1996
                                                                                 WJ. Beny et al.
                                                              100

                                                               80

                                                               60

                                                               40

                                                               20
100
80

>.
•o 60
"§
^ 40
20
C






•-^S
••-•*-« <*•*•*
I
I
t
I
i
I •
I

                                                              100

                                                               80

                                                               60

                                                               40]

                                                               20
0.01    0.1     1     10    100   1000    0.01   0.1     1
       Dry Wt. Metal 
-------
 Predicting the toxichy of metal-spiked sediments
                                                                             Environ. Toxicol Chem. 15, 1996    2075

100
80
| 60
2
«* 40
20

o.c

a






)1

• Cd
A OJ
* HI
• Zn
+ Mi

' ~f • '-
**•£
0.1

• it ttfri * **»* *
mm -*4
. • .-

•
A *"

** * *• +^R» **
1 10 100 IOC
                      DryWt. MetotftJmd/fl)
100.


80-

60

40-

20
                                *" «.;
                                           •
 0.01      0.1.        1        10
                   Dry Wt. Metal (pmol/a)
                                           100
1000
Fig. 4.  Percentage mortality of A. abdita (R. hudsoni in the Ninigret
Pond sediment in the cadmium experiment) as a function of the sum
of the concentrations of cadmium, copper, lead, nickel, and zinc (nM
metal/g dry weight sediment). All experiments combined. Upper panel
plots data by metal; * «= Cd; A » Cu; * - Ni; * = Pb; • = Zn;
and + = mixed metals. Lower panel plots data by sediment. A =
Long Island Sound sediment; • = Ninigret Pond sediment; and • =
mix.-
of the metal sulfide [1], and little or no metal is present in the
IW [IS]. When the binding capacity of the sulfide is exhausted,
partitioning is controlled by sorption [1], and in most of our
test sediments the IW concentrations of metal increased sharp-
ly enough that nearly 100% mortality resulted. The effect is
similar to the "throwing of a switch" at SEM/AVS ratios > 1.0.
This multiple order of magnitude increase in IW concentration
with a factor of 2 or 3 increase  hi sediment concentration
explains, why the chemistry of the anaerobic sediments controls
the toxicity of metals to organisms living in aerobic sediment
microhabitats (e.g., the amphipods living hi their burrows in
our experiments). It also explains why the toxicity of different
metals in the sediments in our experiment was similar on an
SEM/AVS  basis (Fig. 6) even though their toxicities differ
markedly hi water-only toxicity tests (Table 2).  Finally, the
observed sequence in appearance of nickel, zinc,  cadmium,
and copper into sediment IW  is expected because this is the
order of the sulfide solubility products: nickel = -27.98, zinc
= -28.39,  cadmium = -32.85, and copper = -40.94. there-
fore, metals availability and organism response  observed in
  our experiments were controlled by fundamental chemical re-
  actions [1,3]. Knowledge of these reactions allows prediction
  of biological effects and can also be useful in explaining which
  metal(s) might be causing sediment toxicity.
    Available data from other spiked sediment experiments also
  support our observations on the use of SEM/AVS ratios and
  IWTUs to predict mortality in sediment toxicity tests. Our data
  and data from freshwater  tests using oligochaetes and snails
  exposed to sediments, spiked with cadmium [2] and saltwater
  tests using polychaetes exposed to sediments spiked with cad-
  mium, copper, lead, nickel, or zinc [5,9] and copepods exposed
  to sediments spiked with cadmium [4] all yield similar results
  when mortality  is plotted against SEM/AVS ratio (Fig. '8).
  These data describe tests  with  six freshwater and saltwater
  species and sediments from seven sites, with AVS concentra-
  tions ranging from 1.9 to 65.7 umol/g dry weight and TOC
  ranging from 0.15 to 10.6%. Mortality in these experiments
  was sediment independent when plotted on an SEM/AVS ratio
  basis. With the combined data, 98.9% of the 92 metals-spiked
  sediments with SEM/AVS ratios <1.0 were nontoxic, while
  73.5% of the 83 sediments with SEM/AVS ratios >1.0 were
  toxic.
   The presence of additional binding factors may account for
  the fact that not all sediments  with  SEM/AVS ratios >1.0
  caused increased mortality. In addition, organism behavior, in
  a toxicity test can control exposure and limit the impact of
  metals in sediments. Many of the sediments which had the
 highest SEM/AVS ratios in excess of 1.0 that produced little
 or no mortality were from experiments using the polychaete
 Neanthes arenaceodentata (Fig. 8). The polychaetes did not
 burrow in most of these sediments and presumably were not
 fully exposed to the metals in the sediment (see Fig. 3 in Pesch
 et al. [9]). This same phenomenon may also explain the  low
 mortality of snails, Heliosoma sp., in freshwater sediments
 with high SEM/AVS ratios (Fig. 8). These snails are epibenthk;
. and also have the ability to avoid contaminated sediments (O.
 Phipps,  personal commication).  Increased mortality was al  '
 ways observed in sediments with SEM/AVS ratios >5.9 in
 tests with the other four species  tested.
   The combined data from all available freshwater and salt-
 water tests also follow the same pattern as our saltwater  am
 phipod data when plotted on an IWTU basis (Fig. 9). Mortality
 was not sediment specific when it was plotted against IWTU,
 and sediments with IWTUs of «C0.5 were generally nontoxic.
 Of the 88 sediments with IWTUs <0.5, 98.9% were nontoxic,
 while 77.6% of the 76 sediments with IWTUs >0.5 were toxic.
 When SEM/AVS ratio and IWTU were combined, our ability
 to predict which sediments would be toxic was improved. Of
 the 71 sediments with SEM/AVS ratios <1.0 and IWTUs <0.5,
 98.6% were nontoxic, while 87.7% of the 65 sediments with
 SEM/AVS ratios >1.0 and IWTUs >0.5 were toxic (Table 3).
 This close relationship between IWTUs and sediment toxicity
 was also observed in other studies with metal-contaminated
 field sediments  [11,24] as well  as in studies with nonionic
 organic chemicals both in  the field [25,26] and in the labo-
 ratory [27-29]:
   One limitation to the overall spiked  sediment data we sum-
 marize above is that all tests used acute exposures, and the
 interpretation of these, results may not be applicable to chronic
 exposures. Bioaccumulation of metals was also not measured,
 except in two cases [2,9]. The applicability of AVS and IWTU
 normalizations to chronic exposures and bioaccumulation in
 benthic organisms are discussed elsewhere [10,1 1,30-32]. Fur-

-------
 2076    Environ. Toxicol. Chem. 15. 1996
                                                                                    W.J. Beiry et al.
                    100
                     80

                 f  oO

                     <°
                     20;
                        a
                                       -I-
0.001  0.01   0.1    1    10
             SEM/AVS: Cd
100  1000
                                         100

                                          80

                                          60

                                          40

                                          20
                                                                                      •*•*   AA
                                                               0.001  0.01   0.1     1     10    100  1000
                                                                            SEM/AVS: Cu
                    100

                     80

                     60-1

                     40

                     20
                        • A  .A A
                                        i
                                 *«.•'*
                                A*" *   1 •
                     0.001  0.01   0.1     1     10    100   1000
                                  SEM/AVS:NI
100
80
60
40
20
O.C

d A
• ' B

•
A
v; v 	
*' * -1
01 0.01 0.1 1 10 100 10
SEM/AVS:R>





00

100
80
60
40
20
O.C

e ..*,
A

fe
*iA*
. * ",
01 0.01 0.1 1 10 100 1C
SEM/ AVS: Zn





00

100
80
60
40
20
n
f * A

•
• •
A
	 ! 	 '4 A '.
                                                                0.001 0.01  0.1    1     10   100  1000
                                                                       SEM/AV&Cd + Cu+NI + Zn
Fig, 5. Percentage mortality of the amphipod A. abdita (A. hudsoni in the Ninigret Pond sediment in the cadmium experiment) as a function of
the ratio of the sum of the molar concentrations of cadmium, copper, lead, nickel, and zinc simultaneously extracted with AVS to the molar
concentration of AVS (SEM/AVS) in three sediments: Long bland Sound (LIS), Ninigret Pond (N1N), and a 50/50 mixture of these two sediments
(Mix). Each panel represents data from a separate experiment. Data below the SEM detection limit are plotted at SEM/AVS = 0.001. A = Long
Island Sound sediment; • = Ninigret Pond sediment; and • - mix.
thermore, these data are from tests conducted in the laboratory
with homogenized spiked sediments. Therefore, conclusions
from these studies must be carefully applied when conducting
risk assessments with field sediments.
  While our results also show that SEM/AVS and IWTU are
accurate predictors of the absence of mortality in a sediment
toxicity test, predictions of which sediments might be toxic
are less accurate. The fact that a significant number of .sedi-
ments (20%) tested had SEM/AVS ratios of > 1.0 but did not
cause increased mortality indicates that other binding phases,
such as organic carbon [33], may also control bioavailability
in anaerobic sediments. While the SEM/AVS model of bio-
availability accurately predicts which sediments will not be
toxic, a model which utilizes SEM/AVS ratios or SEM - AVS
[23] but incorporates other binding phases might more accu-
rately predict which  sediments will be toxic [34].
  Similarly, a significant number of sediments with IWTUs
>O.S were nontoxic. This is likely the result of IW ligands.
                                         which reduce the bioavailability and toxicity of dissolved met-
                                         als, sediment avoidance by polychaetes or snails, or meth-
                                         odological problems in contamination-free sampling of IW.
                                         Ankley et al. [8] suggested that differences between the hard-
                                         ness of the IW and that of the water in the water-only tests
                                         might affect the accuracy of prediction of sediment toxicity
                                         using  IWTUs in freshwater unless the IWTUs are hardness
                                         corrected. Furthermore, Green et al. [4] and Ankley et al. [8]
                                         hypothesized that increased DOC in the IW reduced the bio-
                                         availability of cadmium in their sediment .exposures relative
                                         to the water-only exposures. Green et al. [4] found that the
                                         LC50 value for cadmium in an IW-only exposure was more
                                         than twice that in a water-only exposure and that the  LC50
                                         value for cadmium in IW associated with sediments was more
                                         than three times that in a water-only exposure.' A significant
                                         improvement in the accuracy of toxicity predictions  using
                                         IWTUs might be achieved if DOC binding in the IW is taken
                                         into account.

-------
 Predicting the toxicity of metal-spliced sediments
                                                                                    Environ. Taxicol. Cheat.  15, 1996    2077
 I
    100

     so

     60

     40

     20
                      .v/
                                       * ***+ * * * +
                            .   \
                   >****£**
         1   0.01     0.1       1       10      100    1000
                          SEM / AVS

Fig. 6.  Percentage mortality of A. abdita (R. hudsoni in the Ninigret
Pond sediment in the cadmium experiment) as a function of the ratio
of die sum of the molar concentrations of cadmium, copper, lead,
nickel, and zinc simultaneously extracted with AVS to the molar con*
centration of AVS (SEM/AVS). All experiments combined. Data be-
low the SEM detection limit are plotted at SEM/AVS - 0.001. • -
Cd; A = Cu; * = Ni; * = Pb; • = Zn; and += mixed  metals.
  The results from our experiments and those published by
others as summarized in this article demonstrate that using
SEM, AVS, and IW metals concentrations to predict which
sediments that contain cadmium, copper, lead, nickel, and zinc
will not be toxic is quite certain. This is very useful, because
the vast majority of sediments found in the1 environment have
SEM/AVS ratios. < 1.0 [22,23,35]. An important consideration
here is that most existing data from field sites are from sed-
iment samples collected in the summer, when the  seasonal
cycles of AVS produce the maximum binding potentials
[36,37]. Hence, sampling at seasons and conditions when AVS
is at minim"'  values is  a must in evaluations of specific sed-
iments and in establishing the true level of overall concern
about metals in. sediments. Predicting which of the remaining
sediments (those with SEM/AVS ratios >1.0) will be toxic is
presently less  certain, although the correct classification rate


£
Mortal
*


100

80
60
40
20
ri
* t » *• **•» **** ' *
* * • •

^ •
*
•
* : A*
*******
101    0.1
                   1
                          10    100   1000  10000 100000
                   Intefstttiol WotGf Toxic Units
Fig. 7. Percentage mortality of A. abdita (R. hudsoni in the Ninigret
Pond' sediment in the cadmium experiment) as a function of IWTUs.
In the individual metal experiments IWTU equals the IW concentra-
tion of the individual metalM. abdita LCSO for that metal. The IWTU
for the mixed metals experiment is the sum of (IW Cd concn./Cd
LC50 for A. abdita) + (IW Cu concn^Cu LCSO for A. abdita) + (IW
Ni-concn Ml LCSO for A. abdita) + (IW Zn concnTZn LCSO for A.
abdita). All experiments combined. Data below the detection limit
are plotted at IWTU « 0.01.  • = Cd; A = Cu; * =  Ni; *  = Pb;
* "r Zn; and += mixed metals.
                                                                100


                                                                80


                                                                60
•*» = L.V.). copper (A = A.a.\ A - C.c.), lead (* = AM.;
                                                                0 =  C.c.), nickel (* = A.O.; •& = Ma.), zinc (• = AM.; a = C.c.),
                                                               or a mixture of metals (+ = A.a.) as a function of the SEM/AVS
                                                               ratio. Data below the SEM detection limit are plotted at SEM/AVS
                                                               = 0.001.          ,                -
                                                               seen in these experiments (better than 70% of sediments pre-
                                                               dicted to be toxic were toxic) is probably better than that of
                                                               any other available method. An SEM/AVS ratio >1.0 should
                                                               trigger additional tiered assessments, which might include tox-
                                                               icity tests, measurements of IW metal, toxicity identification
                                                               evaluations (TIE),  and characterization  of the spatial (both
                                                               vertical and horizontal) and temporal distribution of chemical
                                                               concentration and toxicity. In  this context, the. SEM, AVS,
                                                               IWTU approach should be viewed as only one of many sed-
                                                               iment evaluation methodologies.
                                                                 Of course.  SEM, AVS, and IWTUs can only predict toxicity
                                                               or the lack of toxicity due to metals in sediments. They cannot
                                                               be used alone to predict the toxicity of sediments contaminated
                                                               with toxic concentrations of other contaminants,  e.g., poly-
                                                               cyclic aromatic hydrocarbons. However, SEM/AVS ratios have
                                                               been used in  sediment assessments to rule out metals as prob-
                                                               able causative agents of toxicity [22]. Also, the  SEM/AVS
                                                               approach to predicting the biological availability and toxicity'
                                                                    100


                                                                    80


                                                                    60
                                                                                 A     «
                                                                                 i--ft--*"
                                                                             01
                                                                                    1      10     100    1000
                                                                                     htenhUcd Water Toxic Units
                                                                                                              10000   10QQQG
                                                               Fig. 9. Percentage mortality of saltwater and freshwater benthic spe-
                                                               cies, including oligochaetes (L variegatus, Lv.), polychaetes (C. cap-
                                                               itata, C.c., and N. arenaceodentata, N.a.), ampbipods (A. abdita,
                                                               A.a.), harpacticoids (Amphiascus tenuiremis, A./.), and snails (Heli-
                                                               soma &p.,H.s.) exposed to sediments spiked with cadmium (<$ - A.t.\
                                                               see Fig. 8 legend for definitions of other symbols), copper, lead, nickel,
                                                               zinc, or a mixture of metals as a function of IWTUs. Data below the
                                                               detection limit are plotted at IWTU = 0.01.

-------
 2078    Environ. Toxicol. Chem. 15, 1996
                                             WJ. Berry et al.
                     ' liable 3. Accuracy of predictions of the toxicity of sediments from spiked sediment tests with
                          saltwater amphipods and combined freshwater and saltwater spiked sediment tests as
                               a function of SEM/AVS ratios, IWTUs, and both SEM/AVS and tWTUs
% Sediment •
Study type.
Amphipods
(saltwater)




All species (fresh-
water and salt- •
water)



Parameter
SEM/AVS

WTO

SEM/AVS, IWTU

SEM/AVS

IWTU

SEM/AVS, IWTU

Value
asl.0
>1.0
<0.5
ssO.5
S1.0, <0.5
> 1.0, ^0.5
£1.0
>1.0
<0.5
2:0.5 .
SJ1.0,<0.5
>,1.0.s0.5
' n
43
45
. 33
36
29
36
92
83
88
76
71
65
Nontoxic
97.7'
20.0
97.0
5.6
96.6
5.6
98.9
26.S
98.9
22.4
98.6
12.3
Toxic
2.3
80.0
3.0
94.4
3.4
94.4
1.1
73.5
1.1
77.6
1.4
87.7
of cadmium, copper, lead, nickel, and zinc is applicable only
to anaerobic sediments that contain AVS, because in aerobic
sediments binding factors other then AVS control bioavail-
ability [38,39]. Measurement of IW metal may be useful for
evaluations of these metals in  aerobic sediments and other
metals in sediments in general [12]. Even with theses caveats,
we believe that the use of SEM, AVS,  and interstitial mea-
surements in combination are superior to all other currently
available sediment evaluation procedures to causally assess
the implications of these five metals associated with sediments.
Acknowledgement—We thank S. Benyi, A. Kuhn, and C. Schlekat
for graphics and database support and R. Burgess, R. Hoke, K. Scott,
and two anonymous reviewers for comments that improved the manu-
script W. Berry,' J. Corbin, D. Robson, B. Rogers, and B. Shipley
were supported under U.S. Environmental Protection Agency (EPA)
Contract 68-C1-0005 to Science Applications International Corpo-
ration. DJtf. Di Tbro and J.D. Mahony were supported under U.S.
EPA Cooperative Agreement R812824010 with Manhattan College.
This document has been reviewed in accordance with U.S. EPA, Na-
tional Health and Environmental Effects Research Laboratory, policy
and approved for publication. The contents of this publication do not
necessarily reflect the views of the U.S. EPA. Mention of trade names
or commercial products does not constitute endorsement or recom-
mendation for use.

                      REFERENCES
 1. IN Toro, D.M., JJ>. Mahony, DJ. Hansen, KJ. Scott, MA.
    Hicks, S.M. Mays and MS. Redmond. 1990. Toxicity of cad-
    mium in sediments: The role of acid volatile sulfide. Environ.
    Toxicol.  Chem. 9:1489-1504.
 2. Carlson, A.R, G.L. Phipps, V.R. Mattson, PA. Kosian and
    AAI. Cotter. 1991. The role of acid-volatile sulfide in deter-
    mining cadmium bioavailability and toxicity in freshwater sedi-
    ments. Environ. Toxicol Chem. 10:1309-1319.
 3. Di Toro, D.M., JJ>. Mahony, DJ.  Hansen, KJ. Scott, A.R.
    Carlson and G.T. Ankley. 1992. Acid volatile sulfide predicts
    the acute toxicity of cadmium and nickel in sediments. Environ.
    Sci TechnoL 26:96-101.
 4. Green, A.S., G.T. Chandler and E.R. Blood. 1993. Aqueous-,
    pore-water-, and sediment-phase cadmium: Toxicity relationships
    for a meiobenthic copepod. Environ. Toxicol. Chem.  12:1497- "
    1506.
 5. Casas, AAi. and EA. Crecellus. 1994. Relationship between
    acid volatile sulfide and the toxicity of zinc, lead, and copper in
    marine sediments. Environ. Toxicol Chem. 13:529-536.     '  '
 6. Kemp, P.F. and R.C. Swartz. 1988. Acute toxicity of interstitial
    and particle-bound cadmium to a marine infaunal amphipod. Mar.
    Environ. Res. 245:135-153.
 7. Ankley, G.T., V.R. Mattson,  E.N. Leonard, C.W. West and
    J.L. Bennett 1993. Predicting the acute toxicity of copper in
    freshwater sediments: Evaluation of the role of acid-volatile sul-
    fide. Environ. Toxicol Chem. 12:315-320.
 8. Ankley, G.T., et al. 1991. Acid-volatile sulfide as a factor me-
    diating; cadmium and nickel bioavailability in contaminated sed-
    iments. Environ. Toxicol Chem. 10:1299-1307.
 9. Pesch, OR, DJ. Hansen, W.S. Boothman, WJ. Berry and
  ' JJX Mahony. 1995. The role of acid-volatile sulfide and inter-
    stitial  water metal concentrations in determining bioavailability
    of cadmium and nickel from contaminated sediments to the ma-
    rine polychaete, Neanthes arenaceodentato.  Environ. Toxicol
    Chem. 14:129-141.
10. Hare, L., R. Carlgnan and MA. Huerta-Diaz.  1994. A field
    experimental study of metal toxicity and accumulation by benthic
    invertebrates: Implications for  the acid  volatile sulfide (AVS)
    model. Umnol. Ocemogr. 39:1653-1668.
11. Hansen, D J., JO). Mahony, W J. Berry, S J. Beoyl, JM. Cor-
    bln, SJ>. Pratt, DJM. DI Toro and MJJ. Able. 1996. Chronic
    effect of cadmium in sediments on colonization by benthic marine
 .   organisms: An evaluation of the role of interstitial cadmium and
    acid-volatile sulfide in biological availability. Environ. Toxicol
    Chem. 15:2126-2137.
12. Ankley, G.T., NA. Thomas, D.M. Di Ton, D J. Hansen, JJ>.
    Mahony, WJ. Berry, R.C. Swartz and RA. Hoke. 1994. As-
    sessing potential bioavailability of metals in  sediments: A pro-
    posed  approach. Environ. Manage. 18:331-337.
13. Luoma, S.N. 1989. Can we determine the biological availability
    of sediment-bound trace elements? Hydrobiologia 176/177:379-
    396.      ,
14. Goldhaber, MA. and I.R. Kaplan. 1974. The sulfur cycle. In
-    E.D. Goldberg, ed., The Sea, Vol. S—Marine Chemistry. John
    Wiley  & Sons, New York. NY, USA. pp. 569-655.
15. Gonzales, A.M., JJX Mahony and DJK. DI Toro. 1992. The
    role of organic carbon in the toxicity of anoxic sediments con-
    taminated with copper and other metals: An experimental study.
    Abstracts, 13th Annual Meeting. Society of Environmental Tox-
    icology and Chemistry, Cincinnati, OH. USA, November 8-12,
    p. 162.
16. Spehar, RX. and Jf.T. Fiandt.  1986. Acute and chronic effects
    of water quality criteria-based metal mixtures on three aquatic
    species. Environ. Toxicol Chem. 5:917-931.
17. Scott, KJ. and MS. Redmond. 1989. The effects of a contam-
    inated dredged material on laboratory populations of the tubi-
    culous amphipod, Ampelisca abdita. In U.M. Cowgill and L.R.
    Williams, eds., Aquatic Toxicology and Hazard Assessment, Vol-
    ume 12. STP 1027. American Society for Testing and Materials,
    Philadelphia, PA, USA, pp. 289-303.
18.-Hamilton, MA.,  R.C.  Russo  and R.V.  Thurston.  1977.
    Trimmed Spearman-Karber method for estimating median lethal
    concentrations in toxicity bioassays. Environ. Sci. fechnol 11:
    714-719; Correction 12:417 (1978).
19. Mearns, AJ., R.C. Swartz, J.M. Cummins, PA. Dlnnel, P.
    Plesha and VM. Chapman. 1986. Inter-laboratory comparison
    of a sediment toxicity test using the marine amphipod, Rhepox-
    ynius abronius. Mar. Environ. Res.. 19:13-37.

-------
Predicting the toxicity of metal-spiked sediments

20  VS. Environmental Protection Agency. 1994. Statistical sum-
    mary, EMAP—Estuaries, Virginian Province—1991. EPA/620/
    R-94/005. Technical Report. Office of Research and Develop-
    ment, Washington, DC.
21. Stephan, C.E., DJ. Mount, DJ. Hansen, J.H. Gentile, GA.
    Chapman and WA. Brungs. 1985. Guidelines for deriving nu-
    merical national water quality criteria for the protection of aquatic
    organisms and their uses. PB85-227049. National Technical In-
    formation Service, Springfield, VA,  USA.
22. Wolfe,  DJS., S.B. Bricker,  E.R. Long, KJ. Scott and G.B.
    Thursby. 1994. Biological effects of toxic contaminants in sed-
    iments from Long  Island Sound arid environs. Technical Mem-
    orandum NQS  ORCA 80. National Oceanic and Atmospheric
    Administration, Silver Spring, MD, USA.
23. Hansen, DJ., cfaL 1996. Predicting the toxicity of metal-con-
    taminated field sediments using interstitial concentration of metals
    and acid-volatile sulfide normalizations. Environ. Toxicol. Chem.
    15:2080-2094,
24. Swartz, R.C.; G.R. Dittsworth, D.W. Schults and J.O. Lam-
    berson. 1985. Sediment toxicity to a marine infaunal amphipod:
    Cadmium and its interaction with sewage sludge. Mar. Environ.
    Res. 18:133-153.
25. Hoke, RA., G.T. Ankley, \M. Cotter, T. Goldstein, PA. Re-
    gion, GX. Phlpps and F. VanderMelden. 1994. Evaluation of
    equilibrium partitioning  theory for predicting acute toxicity of
    field-collected sediments contaminated  with DDT, DDE, and
    DDD to the amphipod, Hyalella azteca. Environ. Toxicol. Chem.
    13:157-166.
26. Swartz, R.C, FA. Cole, J.O. Lamberson, S.P. Ferraro, D.W.
    Schults, WA. DeBen, Henry Lee H and RJ. Ozretlch. 1994.
    Sediment toxicity, contamination, and amphipod abundance at a
    DDT- and dieldrin-contaminated site in San Francisco Bay. En-
    viron. Toxicol.  Cheat. 13:949-962.
27. Di Toro, D.M., et al. 1991. Technical basis for establishing sed-
    iment quality criteria for non-ionic organic chemicals using equi-
    librium partitioning. Environ. Toxicol.  Chem. 10:1541-1583.
28. Adams, WJ., RA. Kimerle and R.G. Masher. 1985.  Aquatic
    safety assessment of chemicals sorbed to sediments. In R.D. Card-
    well, R. Purdy  and R.C. Banner, eds.. Aquatic Toxicology and
    Hazard Assessment: Seventh Symposium. STP 854. American
    Society for Testing and  Materials, Philadelphia. PA, USA, pp.
    429-453.
29. Swartz, R.C., D.W. Schults, TJL DeWltt, G.R. Dltteworth and
    J.O. Lamberson. 1990.  Toxicity of fluoranthene in sediment to
    marine amphipods: A test of the equilibrium partitioning approach
                     Environ. Toxicol. Chem. 15, 1996    2079

    to sediment quality criteria. Environ. Toxicol. Chem. 9:1071-
    1080.
30. Ankley, G.T. 1996. Evaluation of metal/acid-volatile sulfide re-
    lationships in the prediction of metal bioaccumulation by benthic
    macroinvertebrates. Environ. Toxicol. Chem. 15:2138-2146.
31. Sibley, P.K., G.T. Ankley, A.M. Cotter and E.N. Leonard.
    1996. Predicting chronic toxicity of sediments spiked with zinc:
    An evaluation of the acid-volatile sulfide model using a life-cycle
    test with the midge Chironomus tentans. Environ. Toxicol. Chem.
    15:2102-2112.
32. DeWitt, T.H., R.C. Swartz, DJ. Hansen, D. McGovern and
    WJ. Berry. 1996. Bioavailabitity and chronic toxicity of cad-
    mium in sediment to the estuarine amphipod, Leptocheirus plu-
    mulosus. Environ. Toxicol Chem. 15:2095-2101.
33. Mahony, J.D., D.M. Dl Toro, A.M. Gonzalez, WJ. Berry and
    G.T. Ankley. 1991. A sediment component in addition to acid
    volatile sulfide that may further control toxicity of metals. Ab-
  . streets, 12th Annual Meeting. Society of Environmental Toxi-
    cology and Chemistry, Seattle, WA, USA, November 3-7, p. 91.
34. Mahony, JJ>., D.M. Dl Toro, AJtf. Gonzalez, M. Curio, M.
    Dilg, Li). De Rosa and LA. Sparrow. 1996. Partitioning of
    metals to sediment organic carbon. Environ. Toxicol. Chem. IS:
    2187-2197.
35. Leonard, E.N, G.T. Ankley and RA. Hoke. 1996. Evaluation
    of metals in marine and freshwater surficial sediments from the
    Environmental Monitoring and Assessment Program relative to
    proposed sediment quality criteria for metals. Environ, Toxicol.
    Chem. 15:2221-2232.
36. Boothman, W.S. and A. Heknstetter. 1992. Vertical and seasonal
    variability of acid volatile sulfides in marine sediments. EPA 600/
    X-93/036. U.S. Environmental Protection Agency, Narragansett,
    RJ.
37. Leonard, E.N., V.R. Mattson, DA. Benoit, RA. Hoke and G.T.
    Ankley. 1993. Seasonal variation of acid volatile sulfide in sed-
    iment cores from  three northeastern Minnesota  lakes. Hydro-
    biologia 271:87^95.
38. Di  Toro,  DAI, F. Harrison, E. Jenne, S. Karickoff and W.
    Lick. 1987. Synopsis of discussion session 2: Environmental fate
    and compartmentalization. In K.L. Dickson, A.W. Maki and WA.
    Brungs, eds., Fate and Effects of Sediment-Bound Chemicals in
    Aquatic Systems. Pergamon, New York, NY, USA, pp. 136-147.
39. Tessier, A., Y. Coufflard, P.G.C. Campbell and J.C. Auclalr.
    1993. Modeling Cd partitioning in oxic lake sediments and Cd
    concentrations in the freshwater bivalve Andonata grandis. Um-
    nol. Oceanogr. 38:1-17.

-------
f
  if

-------

                                                                Environmental Toxicology and Chemistry, Vol. 18, No. 1, pp. 30-39, 1999
                                                                                                          O 1999 SETAC
                                                                                                        Printed in the USA
                                                                                                  0730-7268/99 $9.00 + .00
                                                 Annual Review

  SILVER TOXICITY TO CHIRONOMUS TENTANS IN TWO FRESHWATER SEDIMENTS


         DANIEL J. CALL,*! CHRISTINE N. PoLKiNGHORNE,t THOMAS R MARKEE,! LARRY T. BROOKE,!
                   DIANNE L. GEiGER.t JOSEPH W. GoRsucH.I and KENNETH A. ROBILLARD^
                 fLake Superior Research Institute, University of Wisconsin-Superior, Superior, Wisconsin 54880, USA
                             tEastman Kodak Company, Rochester, New York, New York 14652 USA

                                       (Received 27 May 1998; Accepted 24 July 1998)


      Abstract—Sediment collected from two freshwater lakes. West Bearskin Lake (Cook, MN, USA) and Bond Lake (Douglas, WI,
      USA), was characterized for grain size, total organic carbon, (TOC), acid-volatile sulfides (AVS), simultaneously extracted metals
      (SEM), and iron (Fe). Both sediments had low levels of TOC (<1.0 percent) and AVS (<1.1 |unol/g). West Bearskin Lake sediment
      contained more small (silt and clay) particles than Bond Lake, which was 95% sand. West Bearskin Lake also had higher SEM
      (X SEM = 0.815 vs 0.074 u-mol/g) and had an Fe content that was approximately 30-fold greater than that of Bond Lake. These
      sediments were amended with AgNO, in a series of concentrations, some of which were intended to exceed the total silver (Ag)-
      binding capacity of the sediments, allowing for the appearance of dissolved Ag in pore water (PW). Sediment toxicity tests were  .
      then designed such that the AgNO, amendment levels would result in PW concentrations that bracketed the 10-d concentration
      causing 50% lethality (LC50) for dissolved Ag of O.OS7 mg/L, as determined in a toxicity test in water alone (i.e., without sediment
      present)! The 10-d LC50 values for Chironomus tentans, based upon nominal  additions of Ag to the sediments, were 2.7S and
      1.17 g Ag per kilogram dry sediment for West Bearskin and Bond Lake sediments, respectively. Ah LC50 value based upon
      dissolved Ag in the  PW was determined only for Bond  Lake sediment and was approximately 275 times greater than the water-
      only LC50 value. This indicated that a high proportion of the dissolved fraction was not readily bioavailable to cause lethality. A
      reduction in PW pH and the displacement of other metals from sediment into PW with Ag additions to the sediment likely contributed
      to the observed mortalities and weight losses, particularly at the higher exposure levels. The concentrations of Ag in these sediments
      that resulted  in biological effects are considerably higher than levels reported in  the environment.

      Keywords—Stiver    Toxicity    Freshwater    Sediments     Chironomus tentans
                    INTRODUCTION
   It has been estimated that one fourth to two fifths of the
silver (Ag) released to the environment enters the aquatic com-
ponent [1,2]. Silver is a highly reactive element that may exist
in the aquatic environment in a number of chemical forms.
However, only the free ionic  form of Ag (Ag*) is reportedly
toxic to freshwater organisms [3-8]/This free ionic  form is
present only at very low concentrations due to the fact that it
readily reacts with different inorganic and organic ligands that
are present in the freshwater environment.  Prior to the use of
ultraclean analytical  techniques, measured concentrations of
dissolved Ag in freshwater samples in the United States were
usually reported as <0.2 jxg/L- [9]. Studies in which ultraclean
sampling and analytical procedures were used have reported
even lower concentrations (<0.001 u-g/L) in water [10].
   Most of the Ag-ligand complexes that form in freshwater
[11] settle to the bottoms of streams or lakes to become a pan
of the  sediment [12,13]. Concentrations of Ag in sediment
from highly industrialized areas range from 1 to  ISO mg/kg,
whereas background concentrations in nonurban areas are usu-
ally <0.1  mg/kg [9], The surface sediments act as a  biogeo-
chemical reactor, controlling the return of metals to the over-
lying water or their retention in the underlying sediment re-
pository [14]. In the sediment, the dynamics of Ag speciation
are dependent upon a number of factors, including the geo-
logical characteristics of the sediment, the composition of nat-

   * To whom correspondence may be addressed
(dcall@staff.uwsuper.edu). Contribution 106, Lake Superior Research
Institute.
ural ligands in the overlying and pore water (PW), seasonal
changes that influence biological and chemical cycling (e.g.,
pH. redox status, acid-volatile suffide [AVS] content, organic
carbon content), and physical  influences upon the sediment
(e.g., bioturbation, storm events, and dredging activities). Be-
cause it is the ionic form of the metal in PW that is of toxi-
cological concern to benthic organisms [15-17], it is important
to obtain an understanding of the interactions between Ag and
sediments in determining the concentrations of dissolved Ag
that might exist in a bioavailable and toxic form in PW. This
paper is one in  a series  of papers on Ag in this issue. It ex-
amines the current state  of knowledge regarding Ag in fresh-
water sediments and reports our research results on the toxicity
of Ag-spiked sediments  toward larvae of the freshwater Dip-
teran insect Chironomus tentans. Berry et a!, [18] have applied
the data from this study in evaluating PW  toxicity and si-
multaneously extracted metals (SEM)/AVS relationships.

              MATERIALS AND METHODS

Lake sediment collection, handling, and properties

   Two lakes were selected that were known to contain  sed-
iments of relatively low  total organic carbon (TOC) and AVS
levels. Low TOC and AVS sediments were selected for study
based upon the rationale that these sediments would'represent
worst case conditions for their capacities to bind Ag regarding
the TOC and AVS binding phases. Sediment was collected
from West  Bearskin Lake (Cook, MN, USA) and Bond Lake
(Douglas, WI, USA) with Eckman and Ponar  samplers. It was
placed into polyethylene containers, and refrigerated (~4°C)
                                                          30

-------
 Silver toxicity in freshwater sediments
                      Environ. Toxicol. Chem. 18, 1999    31
                                     Table 1. Physicochemical properties of test sediments
                                                                               Sediment
                  Property
                                                                West Bearskin Lake
                         Bond Lake
Particle size (%)
Gravel
Sand
Silt
Clay
Dry weight '{%) '
Total organic carbon (%) '^
Acid-volatile sulfide (ixmol/g)1
Simultaneously extracted metals (SEM, |imol/g)b
Cadium
Copper
Lead
Nickel
Silver :
Zinc
2SEM (M-mol/g)'
Iron (nmol/g)"
B = 1
0.0
54.2
37.6
8.2
59.7 (n = 16)
0.87 (n = 1)
<0.1 (n = 8)
n = 2
0.003
0.085
0.042
. 0.121
0.004
0.560
0.815 (n ~ 2)
165 (n = 1)
n = 1
- 1.3
95.3
3.4
0.0
76.1 (n = 10)
0.22 (a = 1)
- <0.1 (n = 15)
n = 4 "
O.OOi
0.015
0.007
0.013
0.006
0.035
0.077 (n = 4)
5.8 (n = 7)
                  •Determined during toxicity tests.
                  11 Mean as determined from control sediment during toxicity tests.
                  c SEM - simultaneously extracted metals.
                  " Determined by SEM extraction procedure.  .
 until mixed. Shortly after collection, the sediment was thor-
 oughly homogenized in a stainless-steel drum with a com-
 mercial drill equipped with a mortar paddle, apportioned back
 into the polyethylene buckets, and refrigerated until used in
 tests. The sediment was stored for 10 and 4 to 5 months prior
 to the initiation of Ag equilibration and toxicity tests for West.
 Bearskin and  Bond  Lakes,  respectively.  Sediment samples
 were remixed prior to use in each type of sediment test. Sam-
 ples were analyzed for particle size  distribution, dry weight,
 and TOC by a contract laboratory. Acid-volatile sulfide, SEM
 and iron (Fe) content were also determined (Table 1), for which
 methods are described below.

 Silver toxicity test in water
    A 10-d water-only toxicity test was conducted with 11-d-
 old larvae of  C. tentans exposed to AgNO,. The exposure
' chambers were 300-ml high-form  beakers  containing two
 mesh-covered holes (12-mm diameter) in the walls of the beak-
 ers to allow for an outflow of water into the aquarium in which
 they were placed. Ten larvae were exposed in duplicate to five
 exposure concentrations, plus a control. The nominal total Ag
 concentrations were 0,12:5, 25,50,100, and 200 jig/L. Mean
 measured dissolved Ag concentrations of <5.0, <5.0, <5.0,
 13.0, 66.0, and 155 n-g/L for the control and five treatments
 were provided by pumping test solutions directly from stock
 containers with a multihead peristaltic pump (Cole-Farmer In-
 strument, Chicago, IL, USA) to the bottom of each exposure
 beaker at the rate of approximately  1.0 ml/min. The AgNO3
 stock solutions were prepared in laboratory water, which was
 dechlorinated municipal water that had been passed through
 an ion-exchange resin column for removal of metal ions.
    A thin layer of quartz sand was added to each beaker prior
 to the addition of larvae  to reduce stress. Each test chamber
 was supplied  with 6.0 mg of Tetrafin® slurry daily  for the
 larvae to feed upon. Daily measurements were made  at each
 exposure concentration of temperature, dissolved oxygen, and
 pH in alternating replicates. Hardness, alkalinity, and conduc-
 tivity were measured twice during the test near the beginning
 and end in control, low, medium, and high treatments. The
overall means for test temperature, dissolved oxygen, and pH
were 23.0 ± 0.2°C,  7.7 ± 0.4 mg/L, and 7.4 ± 0.1  units,
respectively (n = 78). Mean values for  hardness (n  - 8),
alkalinity (n - 8), and conductivity (n = 24) were 52.1  ± 4.8
mg/L, 45.0 ± 1.4 mg/L, and 147 ± 5.41 funhos/cm, respec-
tively. Total, dissolved, and free ionic Ag were measured in
all replicates on days 0 and 10 of the exposure and in alternate
replicates on days 3 and 6. Total Ag was the amount measured
by atomic absorption (AA) spectroscopy in unfiltered acidified
samples. Dissolved Ag was  procedurally defined as the Ag
that passed  through a 0.2-fun polyethersulfone membrane
(Gelman Sciences, Ann Arbor, MI, USA) and was measured
by • AA  spectroscopy. Free Ag* was  the form that passed
through the 0.2-u.m membrane and was measured by ion-spe-
cific electrode.

Silver binding capacity experiments
   Preliminary spiking experiments without animals present
were conducted to determine the binding capacity of each
sediment for Ag. AgNO3 was added to the sediment in a series
of concentrations, and dissolved Ag concentrations were mea-
sured in the  PW. It was assumed that dissolved Ag would not
be measurable in the PW until the sediment binding capacity
was reached, at which point it would appear in the PW in
increasing concentrations with increasing AgNO3 spiking lev-
els. The objective was to ascertain  the spiking levels needed
to yield concentrations of dissolved Ag in the PW that would
bracket  the  10-d dissolved Ag LC50 value for C. tentans as
determined in the water-only toxicity test. These experiments
were conducted by spiking AgNQ3 solutions into homogenized
sediment in 2-L fluorinated polyethylene containers. The
spiked sediment was stirred  with a motorized stainless steel
stirrer for 5 min under a nitrogen atmosphere. An acrylic plas-
tic peeper containing two 6.0-ml chambers covered with a 0.2-
|jun pore size polyethersulfone membrane and containing de-
gassed,  deionized water was inserted into the sediment, and
the container was flushed with nitrogen, covered with a screw
lid, and placed into  a refrigerator at 4"C. The peepers were
withdrawn from the sediment at weekly intervals, with a new

-------
           32    Environ. Toxicol. Chem. 18, 1999
                                                                                                        D.J. Call et al.
t
peeper inserted each time to replace the withdrawn peeper. The
water within the peeper chambers was withdrawn with an Ep-
pendorf pipette, placed into Teflon* bottles cleaned with am-
monium  hydroxide and  nitric acid, acidified to pH s2 with
concentrated nitric acid (trace metals grade), and analyzed for
Ag and other metals. Concentrations of Ag in the peeper cham-
ber water (i.e., sediment PW) were procedurally defined as
dissolved, having passed through the 0.2-|un pore size mem-
brane. The spiked sediment samples were analyzed over 2 to
3 weeks  to determine if an equilibrium had t>een established
based upon constant concentrations of dissolved Ag in the
peepers on successive sampling dates.

Silver toxicity tests in sediment

   West Bearskin Lake. Based upon results from the water-
only toxicity test and the Ag binding  capacity study, West
Bearskin Lake sediment was spiked with AgNO, at nominal
concentrations of 1.7, 2.2,3.0,4.0, 5.4,7.2, and 9.6 g Ag per
kilogram dry  sediment. The test was performed in a modified
system based upon that  of Benoit et al. [19], with test con-
ditions as described in  standard procedures [20].  Sediment
control and sand performance controls were also tested, with
mean survival rates of 80.0 and 93.3%, respectively. In the
sand performance controls, the mean dry weights averaged 0.8
mg per individual/  indicating adequate growth for  test ac-
ceptability  [20]. Exceptions to standard procedures were that
the system provided for renewal of overlying water to the
aquaria at a rate equivalent to approximately four volume ad-
ditions daily rather than two, and only three biology replicates
were used rather  than four. Screened exposure beakers (300
ml) were used for both biological and chemical testing at each
exposure level, with four beakers used for chemistry mea-
surements.  Each beaker received approximately 100 ml of ei-
ther AgNOj-amended sediment, control sediment,  or quartz
sand (used as a test performance control). The chemistry beak-
ers each received a two-chambered peeper covered with a 0.2-
|un polyethersulfone membrane, positioned to a depth within
the sediment such that the lower chamber was buried below
the surface of the sediment to collect PW and the upper cham-
ber was in  the overlying water. The beakers containing sedi-
ment were  placed into the aquaria (30.S X  15.0 X 8.0-10.0
cm, length  X width X water depth) of the toxicity test system
7 d prior to the introduction of test animals, with renewal of •
overlying .water occurring during  this time period. This al-
lowed for elimination of any excess nitrate ion concentration
in the overlying water. Temperature, dissolved oxygen, and pH
were measured in all test chambers on days 0 and  10 and-in
at least one replicate per treatment on intermediate days (n =
199). The  overall respective test means (±SD) were  23.3 -
(0.6)°C, 7.5 (0.7)  mg/L,  and 7.3 (0.4) units.
   Ten second-third instar C. tentans larvae (10-11 d old)
were randomly added to each replicate, including the chemistry
beakers, 39 d after Ag addition to the sediment. Each replicate
was fed the equivalent of 6.0 mg of dry solids daily, as hi the
water-only  test. Chemistry beakers were sampled on days 0,
5, and 10 of the toxicity test. A peeper was removed from one
replicate at each exposure concentration, and the peeper cham-
ber water was analyzed  for pH and Ag. A subsample of the
peeper water was also analyzed for concentrations of zinc (Zn),
nickel (Ni), lead (Pb), copper (Cu), total ammonia and  non-
purgeable dissolved organic carbon. Cadmium (Cd)  was not
analyzed due to the very  low concentration in the SEM extract.
The overlying water was also sampled and measured for Ag
and other metals.
   On day 10, the biology test beakers were sieved. Survivors
were counted and placed into preweighed pans for each rep-
licate and dried in an oven at 105*C for ~24 h. Survival,
individual dry weight, and total biomass per replicate were
statistically analyzed.
   Bond Lake. Two  10-d toxicity tests were  performed with
Bond Lake sediment, due to the fact that the PW-dissolved Ag
concentrations in the first test were less than expected based
on the initial binding study, and the exposures were not lethal
to C. tentans.  The AgNO3 additions to the bulk sediment in
the first test resulted in nominal concentrations of 0.005,0.01,
0.03, 0.08, 6.20, and 0.50 g Ag per kilogram dry sediment.
The AgNO3 additions in the second test yielded nominal con-
centrations of 0.03, 0.08, 0.20, 0.50, 1.2, and 3.1 g Ag per
kilogram dry sediment. Each test also contained sediment and
sand performance controls, with respective mean survival rates
of 93.3  and 100% for test 1 and 96.7 and 83.3% for test 2.
The tests were performed as described for West Bearskin Lake.
Larvae (2nd/3rd instar) were added to the sediment 22 d after
the AgNOj additions in both tests. For test 1, the overall means
(±SD) for temperature, dissolved oxygen, and pH (n = 176)
were 23.1 '(0.3) "C, 7.6 (0.8) nig/L, and 7.4 (0.2) units, re-
spectively. For test 2, the respective means (n = 176) were
23.1 (1.0) °C, 7.2 (1.0) mg/L, and 7.4 (0.2) units, respectively.
In Bond Lake, PW was analyzed for Ag, Zn, and iron (Fe).
Nickel, Pb, Cu, and Cd were not analyzed due to their presence
only at very low concentrations in the SEM extract.
   Analytical chemistry. Total and dissolved' Ag concentra-
tions were measured using a Varian SpectrAAlOO AA spec-
trophotometer. Samples were acidified to a pH £2 using trace
metal grade nitric acid. Flame AA was utilized in all cases,
with the exception of the first Bond Lake toxicity test, where
graphite furnace AA was used to measure the dissolved Ag
concentrations due to the fact mat the concentrations were
below the flame detection limit. Free ionic Ag* was measured
using an ion-specific electrode (Orion 9616BN). For total, dis-
solved and free ionic Ag measurements, a calibration curve
was developed from a minimum of four standards bracketing .
sample concentrations. If necessary, samples were diluted With
deionized water to bring them within the range of the stan-
dards. Measurements of additional dissolved metals in peeper
samples were analyzed using flame AA (Varian SpectrAA200).
The metals that were quantified included Cu,  Zn, Fe, Ni, and
Pb.
   Nonpurgeable dissolved organic carbon was analyzed using
a Shimadzu TOC 5050A organic carbon analyzer according
to U.S. Environmental Protection Agency (U.S. EPA) method
415.1 [21]. Samples (6.5 ml) were collected from peepers, 0.5
ml of deionized water was added to provide an adequate sam-
ple volume, and the  samples were preserved with 5 uJ of 2
M HO. Upon  analysis, each sample was sparged with air for
1 min to eliminate any inorganic carbon. The concentrations
of organic carbon were then quantified relative to a standard.
curve developed from four standards.
   Acid-volatile  sulfide and SEM analyses  were conducted
using the HC1. gas train method [22]. A measured mass  of
frozen sediment was placed into a reaction flask with 100 ml
of deionized water and purged with nitrogen, and 20 ml  of
6M HC1 was added. Nitrogen gas was bubbled into the reaction
flask and then passed through a series of glass impinger bottles.
The first impinger bottle was filled with KHP solution and the

-------
 Silver toxicity in freshwater sediments

 final two were filled with 0.1M AjgNO,. Sulfide present in a
 sample was converted to hydrogen  sulfide, which passed
 through the first impinger bottle and reacted with the AgNO3
 in the second to form insoluble Ag2S precipitate. The Ag2S
 precipitate was filtered and weighed to determine AVS. The
 HC1 extract containing the SEMs was filtered to separate the
 extract from the sediment. This extract was analyzed using
 flame AA for Ag, Cd, Cu, Ni, Pb, Zn, and Fe.
   Ammonia was measured both  in the overlying water and
 PW collected from peeper samples as  well as overlying H2O
 collected  with an Eppendorf pipettor. In both cases,  samples
 were treated with 10 M NaOH just prior to analysis. Overlying
 water samples were analyzed using an ion-specific electrode
 (Orion  9512), and peeper samples were analyzed using an
 ammonia  microelectrode (MI-740 Microelectrodes, London-
 derry. NH, USA).
   Statistical  Analyses. Estimates of concentration  causing
 50% lethality (LC50) were obtained using the trimmed Spear-
 man-Karber program (version 1.5) from a U.S. EPA statistical
 software package [23]. Larval survival, individual dry weight,
 and  total  biomass per replicate were  analyzed by one-way
 analysis of variance, and a comparison of control and treatment
 means was performed by Bonferroni's t-test, as provided in a
 TOXSTAT* software package  (University of Wyoming, Lar-
 amie, WY, USA).

              RESULTS AND DISCUSSION
                                               ;
 Silver toxicity in water               ' • -.  *
   Ten-day LC50 values (95% confidence intervals) for C.
 tentans larvae exposed to Ag in the absence of sediment were
0,063 (0.036-0.108), 0.057 (0.030-0.108), and 0.035  (0.017-
 0.075) mg/L for total Ag, dissolved Ag, and free Ag*, re-
spectively. Dry weights of surviving larvae were less than half
of control larval weights  at the two highest exposures.where
the mean free Ag* concentrations, were 0.041 and 0.110 mg/
L. Based upon significant (p = 0.05) reductions of 82.4 and
95.1% hi  dry  weight at the two highest exposures of 0.066
and 0.155 mg/L dissolved Ag,  respectively, but not at 0.013
mg/L, the no observable effect concentration (NOEC) and low-
est observable effect concentration (LOEC) range was 0.013
to 0.066 mg/L of dissolved Ag. Dry weight was only  slightly
reduced (11.1%) at 0.013 mg/L. The observed sensitivity of
C. tentans to Ag in our laboratory water is lower than the 10-
d LC50 of 0.259 mg/L obtained by Rodgers et al. [24]. They
reported a 10-d NOEC value of 0.125 mg/L of Ag. The lab-
oratory  water used in our study may have had fewer ligands
to bind  me Ag* ion than  the pond water used by Rodgers et
al. [24], and some of the larger colloids that may have passed
through their 0.45-jun filter may have been retained'by our
0.2-|un filter. However; our LC50 value was greater than a 48-
h LC50 value of 0.010 mg/L of total Ag for third instar larvae
of this species in an unfed test in which the Ag concentrations
were not measured [25].           ...

Silver binding capacity of sediments   '
   The preliminary Ag equilibration test results (Table 2) in-
dicated that PW concentrations in AgNO3-amended West Bear-
 skin Lake sediment had come  into equilibrium within 15 d.
 Pore-water Ag concentrations were .generally similar  on days
 15, 22, and 29. However,  the day 8 concentrations at the three
 highest Ag,additions were much higher-than later measure-
 ments. From this study, it was determined that a minimal equil-
                       Environ. Toxicol. Chan. 18, 1999   33

  Table 2. Concentrations of "dissolved" silver(Ag) in pore water from
  peeper chambers buried  within West Bearskin Lake sediment in
                preliminary equilibration studies
Nominal
sediment
Ag
(g Ag/kg)
Control
1.7
2.2
3.0
4.0
5.4
7.2
9.6
Dissolved Ag concentration (mg/L)
DayS
<0.005
<0.005
<0.005
0.005
<0.005
1.03 .
'4.09
13.9
Day IS
<0.005
0.010
0.007
0.010
0.012
0.019
0.022
0.087
Day 22
<0.005
<0.005
0.005
0,008
0.021
0.025
0.041
0.080
Day 29
<0.005
0.006
<0.005
0.006
0.010
0.013
0.017
0.092
 ibration period of 15  days was required for West Bearskin
 Lake sediment. It was deduced that this equilibration time
 would also be adequate for Bond Lake sediment, with fewer
 Ag binding sites in the form of AVS, TOG, and total surface
 area of particles in this sediment. Silver equilibration periods
 of 39 and 22 d were used for West Bearskin and Bond Lake
 sediments, respectively, before animals were added in the sed-
 iment toxicity tests. These sediments had been stored • under
 refrigeration for up to 10 months prior to use and likely un-
 derwent some chemical changes during this storage period.
 Plots of the PW Ag concentrations during both the equilibra-
 tion and toxicity testing periods are shown in Figure 1.
    The concentrations of Ag and other heavy metals analyzed
' in the peeper chamber water during the tests at the various
 AgNOj spiking levels (Tables 3 and 4) show that West Bearskin
 Lake sediment- had a greater binding capacity for Ag+ .than
 Bond Lake sediment.  In West Bearskin  Lake sediment, no
 dissolved Ag was measured in the PW at spiking levels up to
 2.2 g Ag .per kilogram dry sediment. At 3.0 g Ag per kilogram
 dry sediment, dissolved Ag in the PW reached a maximum,
 as did toxicity. It is unclear as to why this intermediate con-.
 Generation had the highest PW concentration of Ag. Silver
 displaced the other metals as the spiking levels increased, with
 particularly high concentrations of Z& and Ni released into the
 PW. The pH decreased with increased additions of AgNO3.
    In Bond Lake sediment, no appreciable concentrations of
 dissolved Ag were measured in the peeper chambers at spiking
 levels up to 0.08 g Ag per kilogram dry sediment. However,
 at spiking levels of 0.20 g Ag per kilogram dry sediment 'and
 higher, dissolved Ag concentrations increased with increasing
 AgNOj additions. A ^dramatic increase in, dissolved PW Ag
 occurred between the 0.20  and 0.50 g Ag per kilogram sedi-
 ment spiking levels. Concentrations of Zn in PW also increased
 as with West Bearskin Lake sediment and, at comparable spik-
 ing levels, were displaced to the PW in greater concentrations
 in Bond Lake sediments than hi West Bearskin Lake sediments.
'The concentrations'of Fe in PW decreased dramatically  be-
 tween the spiking levels of 0.50 and  1.2  g Ag per kilogram
 sediment. This may indicate that more Fe oxides had reacted
 with the additional Ag* to form either a precipitate or a large
 (i.e., >0.2 uin) soluble Ag-Fe oxide complex that could not
 pass through the 0.2-u,m peeper membrane.
    These data suggest that  West Bearskin Lake sediment can
 effectively bind 2.2 g  Ag per kilogram dry sediment before
 any dissolved Ag occurs in the PW. By comparison,,Bond
 Lake sediment can effectively bind only 0.08 g Ag per kilo-
 gram dry sediment. This difference in binding capacity may

-------
34     £hvirgAg/kgDiy6«l
•(.•gAg/kgDiySad
   0
    0.01 •
   0.001
       10
                    20           JO
                          Days Post-Spiking
                                                                            Days Post-SpiWno
Fig.  1. Pore-water silver concentrations over the course of the equilibration and toxicity test periods. Arrow represents the start of the 10-d
toxicity test.      .        .          -            -

be due in pan to the slightly higher AVS and TOC of West
Bearskin Lake sediment as well as to differences in particle
size distribution and basic geochemical composition. Prior to
the addition of AgNO3 to die sediments, AVS levels were 1.1
and <0.1  umol/g for West Bearskin and Bond Lakes, respec-
tively. During the toxicity tests, AVS levels were <0.1 |unol/
g in both  sediments. West Bearskin Lake sediment contained
a higher percentage of fine particles  (silt and clay) than Bond
Lake sediment, providing more surface area for adsorption. In
eight different sediments, Rodgers et al. [26] observed that the
silt and clay portion of the sediments contained a much greater
proportion of Ag than the sand fraction following amendment
with AgNOj. West Bearskin Lake  sediment also  contained
more Fe than Bond Lake sediment as shown  in Table 1. Both
Fe sulfides and Fe oxides may react with Ag to decrease the
                                                    amount of free ion present in water in ahoxic and oxic con-
                                                    ditions, respectively [27]. Rodgers et al. [26] concluded that
                                                    sediments vary widely in their-affinity for Ag and their ability
                                                    to bind Ag. They also  stated that no  single parameter was
                                                    predominant in influencing the partitioning and bioavailability
                                                    of Ag but that partitioning and bioavailability were a result of
                                                    the interaction of several sediment characteristics. Berry et al..
                                                    [18] determined the toxicity of two marine sediments of dif-
                                                    ferent AVS levels amended with Ag to the estuarine amphipod
                                                    Ampelisca abdita. Amphipod mortality was sediment specific
                                                    when expressed on a dry weight basis but not when based on
                                                    interstitial water toxic units or the difference between SEM
                                                    and AVS (i.e., SEM - AVS). These studies'are in agreement
                                                    with the current study  on the variable nature  of sediments
                                                    regarding their capacities to bind  Ag.
Table 3. Mean overlying water (OW) and pore-water (PW) pH and dissolved metal concentrations and survival organism dry weight, and total
      biomass of Chironomus tentans larvae in a toxicity test with West Bearskin Lake sediments amended with silver (Ag) as AgNO,
Nominal •
dry
sediment
concn.
(gAg/kg)
Control
1.7
2.2
3.0
4.0
5.4
7.2
9.6 ,
OW/
PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
Dissolved metal (mg/L)*
PH
7.67
7.07
7.78
6.68
7.62
6.44
7.27
6.16
7.39
6.02
6.72
5.46
6.32
5.09
5.76
4.26
Ag
<0.005
<0.005
<0.005
<0.005
0.006
<0.005
0.158
. 0.299
0.031
<0.005
0.011
0.042
0.057
0.010
0.209
0.059
Zinc
<0.005
<0.005
<0.005
0.005
<0.005
0.023
0.007
0.080
<0.005
0.139
0.027
0.438
0.123
1.230
. 0.374
3.940
Nickel
<0.022
<0.022
<0.022
<0.022
<0.022
0.026
<0.022
0.077
<0.022
0.118
0.024
0.281
0.070
0.665
0.177
1.640
Copper
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
<0.006
0.011
0.044
0.090
Lead
<0.018
<0.018
<0.018
<0.018
<0.018
0.022
<0.018
0.037
<0.018
0.043
<0.018
0.049
<*).018
0.067
0.049
0.076
• Mean%(SD)
survival
80.0 (10.0)
76.7 (5.77)
63.3 (20.8)
10.0 (10.0)"
46.7 (40.4)
26.7 (25.2)"
33.3 (5.77)
40.0 (43.6)
Mean (SD)
organism dry
weight (mg)
2.30 (0.05)
1.82 (0.06)
0.94(0.11) . f
0.65 (0.47)
0.76 (0.40)*
0.61(0.01)
1.66 (1.28)
2.05(1.09)
\
Mean (SD)
replicate total
biomass (mg)
15.4
14.0
5.83
(3-84)
(1.27)
(0.804)k
0.810 (0.170)" '
4.47
2.44
6.01
5.02
(2.68)"
(0.64)»
(3,24)"
(1.26)"
* Dissolved is proceduraUy defined as having passed through a 0.2-uin pore size membrane filter.
b Significantly reduced from survival, dry weight, or total biomass of control animals (p s 0.05).

-------
 Silver toxicity in freshwater sediments
                      Environ. ToxicoL Chem. 18, 1999    35
 Table 4. Mean overlying water (OW) and pore-water (PW) pH and dissolved metal concentrations and survival, organism dry weight, and total
         biomass of Chironomia tentans larvae in two toxicity tests with Bond Lake sediment amended with-silver (Ag) as AgNO,
Nominal
dry
sediment
concn.
(gAg/kg)
1
Dissolved metal (mg/L)*
OW/PW pH
Ag
Zinc
Iron
Mean (SD)
% survival
Mean (SD)
organism dry
weight (mg)
Mean (SD)
replicate total
biomass (mg)
. , Test 1
Control
0.005
0.01
0.03
0.08
0.20
0.50
OW
PW
OW
-PW
OW
PW
OW
PW
OW
PW
OW
PW
OW
PW
7.65
6.82
7.62
6.82
7.62
6.82
7.51.
' 6.84
7.68
6.91
7.52
6.90
7.50
6.21
0.0008
0.0002
0.0002
0.0006
0.0002
0.0010
0.0010
0.0027
0.0061
0.0060
0.016
0.034
0.013
0.127
0.017
<0.005
<0.005
<0.005
<0.005
0.009
<0.005
<0.005
<0.005
<0.005
<0.005
<0.005
0.012
0.007
1.012
17.3
<0.065
13.4
<0.065
14.2
<0.065
13.9
<0.065
9.04
<0.065
5:16
<0.065
1.20
93.3
100
, 100
100
83.3
86.7
(11.5)
(0.0)
(0.0)
(0.0)
(15.3)
(5.77). .
83.3 (5.77)
2.66
2.29
1.99
1.92
2.53
1.98
1.85
(0.40)
(0.26)
(0.11)
(0.44)
(0.56)
(0.16)
(0.26)*
22.3
20.4
17.9
16.4
f
18.4
17.2
15.3
(6.95)
(1.33)
(1.21)
(2.89)
(2.53)
(2.46)
(1.79)°
                                                         lest 2
Control
0.03
0.08
0.20
0.50
1.2
3.1
OW
PW
OW
PW
OW
PW'
OW
FW
OW
PW
OW
PW
OW
PW
7.56
6.89
7.50
6.95
7.50
6.89
7.49
. 6.90
7.66
6.32
7.46
4.41
6.89
3.57
<0.005
<0.005
. <0.005
<0.005
<0.005
0.033
' 0.009
0.019
0.017
1.044
0.156
7.979
34.61
1,389
<0.005
<0.005
<0.005
0.006
0.019
<0.005
<0.005
<0.005
0.077,
0.025
<0.005
0.311
0.031
1.49
<0.065
12.1
0.113
9.64
0.275
8.74
0.167
6.42
0.065
4.71
<0.065 • :
1.15
<0.065
0.634
96.7 (5.77)
100 (0.00)
93.3 (5.77)
93.3 (5.77)
90.0 (lOJO)
' . ! •'
53.3(11.5)*
0.0(0.0)"
" 2.96(0.16)
3.03 (0.16) .
2.98 (0.11)
2.56(0.55) '
2.29 (0-iO)*
0.36 (0.037)>
. _*
27.6 (3.97)
29.2 (1.38)
25.8 (2.17)
22.2 (2.12)
20.5 (1.5iy
1.89 (0.311)=
~~?
 " Dissolved is procedurally defined as having passed through a 0.2-|un pore size membrane filter.  ,
 * Significantly different from survival or dry weight of control animals in the same test.
 ' Significantly reduced from total biomass of control animals with test 1 and 2 data pooled (p -s 0.05).
 cNo data as there were no survivors.     •     '.                       •
 Silver toxicity in sediments
    West Bearskin Lake. During the toxicity tests with C. ten-
 tans, no measurable Ag was present in the PW in the two
 lowest exposures of 1.7 and 2.2 g Ag per kilogram dry sed-
 iment (Table 3). At the 1.7-g Ag per kilogram exposure, PW
 concentrations of Zn, Ni, Cu, and Pb were either at or below
 analytical detection limits. At 2.2 g Ag per kilogram dry sed-
 intent, PW concentrations of Zn, Ni, and Pb were detectable,
 but Cu remained below detection. In fact, Cu was measurable
 only at the two highest exposures. Dissolved Ag concentrations
• behaved erratically in the  PW at  exposures of 3.0 g Ag per
 kilogram dry sediment and higher. A reddish orange-colored
 precipitate, presumably oxides of iron, was visible  at spiking
 levels of 4.0 g Ag per kilogram dry sediment and higher.
 Amorphous iron oxides formed at the higher  AgNO3 spiking
 levels may have scavenged the freely dissolved Ag to varying
 degrees, thereby affecting dissolved Ag concentrations.
    The biological  responses of C.  tentans showed that the
 greatest reduction  in survival occurred at the 3.0 g Ag per
 kilogram sediment spiking level, which also  had the highest
 mean measured PW dissolved Ag concentration of 0.299 mg/
L. Mean survival was improved at the higher exposures but
remained below 50% in all treatments higher than 3,0 g Ag
per kilogram dry sediment. Mean organism dry weights were
most greatly reduced between the exposure levels of 2.2 and
5.4 g Ag per kilogram dry sediment;  however, only the 4.0-
g/kg exposure resulted in a significant (p - 0.05) reduction
(Table 3). The total replicate dry weight biomasses of surviving
larvae were significantly (p s 0.05) decreased at all Ag amend-
ment levels  except for the lowest. This endpoint  integrates
both survival and individual organism dry weight. The PW-
dissolved Ag concentrations were below detection (<0.005
mg/L) at the two lowest spiking levels, so it would appear that
Ag was not  directly responsible for the observed biomass re-
duction at the 2.2 g Ag per kilogram sediment amendment
level. The NOEC-LOEC range based  upon total biomass and
nominal bulk sediment additions of Ag was 1.7  to 2.2 g Ag
per kilogram dry sediment. An LC50  value of 2.75 g Ag per
kilogram dry sediment was estimated based on nominal bulk
sediment additions. Due to the erratic dissolved Ag concen-
trations in the PW, it was not possible  to estimate a PW LC50
value for dissolved Ag.

-------
 36    Environ. Toxicol. Cheat.  18, 1999
                                            D.J. Call et al.
                 Table S.  Summary of biological responses of Chironomus tentans larvae exposed to silver (Ag) in water
                                                     and sediments
Biological responses

Test medium
Laboratory water
West Bearskin
Lake sediment ,
Bond Lake
sediment

Matrix*
PW
S
PW
S

10-d LC50"
0.057
2.75
15.1
1.17

NOEC?
0.013
	 *
1.7
0.034
0.20

LOEC*
0.066
— t
2.2
0.127 .
0.50
Most sensitive
endpoint
Dry weight
• Total biomass
Dry weight and to-
tal biomass
                 " PW = pore water; S = sediment.                                    .
                 * LC50 = concentration causing 50% lethality; NOEC = no observable effect concentration; LOEC =
                  lowest observable effect concentration. Concentrations are for dissolved silver in milligrams per liter
                  and nominal additions of Ag to bulk sediment in grams Ag per kilogram .dry sediment
                 * Not calculable due to nature of dose-response data.           '   •
                 'Not determined as the PW Ag concentrations at both the NOEC and LOEC  were below detection
                  (<0.005 mg/L).
   The displacement of other metals likely contributed to ob-
served mortalities at higher AgNO3 amendment levels. Con-
centrations of Zn, Ni, and Pb increased in the PW as AgNO3
spiking level increased. Due to the displacement of Zn, Ni,
Cu, and Pb in (he sediment by Ag*, it is likely that Ag* in
combination with these metals, as well as decreasing pH, con-
tributed to observed reductions in survival at spiking levels
of a3.0 g Ag per kilogram dry sediment. The 10-d LCSO value
for Zn with C. tentans is 1.12 mg/L [28]. Therefore, concen-
trations of Zn alone would have  contributed 0.021, 0.124,.
0.390,1.10, and 3.52 acute toxicity toxic units (TUs) at spiking
levels of 3.0, 4.0, 5.4, 7.2, and 9.6 g Ag per kilogram dry
sediment, respectively. Ten-day LCSO values  for C. tentans
were  not available for the other metals that were measured.
Nickel and Pb would also have contributed some fractional
TUs,  whereas Cu would have contributed'fractional TUs only
at the two highest spiking levels. The observed displacement
of Zn and Cu followed the order to be expected in terms of
their respective metal sulfide solubility products [29]. This was
also previously observed in laboratory sediments spiked with
several divalent  metals [15]. Acid-volatile sulfide concentra-
tions  were very low, starting  at approximately 1.1 umol/g
before the test started and decreasing to <0.1  |unol/g during
the toxicity test in the control, low, medium, and high spiking
levels. Therefore, it would appear that the dissolution of these
metals from their associations with other ligands may follow
the same general pattern as with sulfide.
   Bond Lake. In two tests with Bond Lake sediment, no ap-
preciable dissolved Ag appeared in the PW until a spiking
level  of 0.08 g Ag per kilogram dry sediment was reached.
Zinc concentrations in PW first appeared at 0.50 g Ag per
kilogram sediment and increased with increasing Ag amend-
ments up to a high mean concentration of 1.49 mg/L of Zn at
the highest exposure. Iron concentrations in PW remained be-
tween 12.1 and 17.3 mg/L in the controls and the lower spiked
sediments but decreased noticeably at an amendment level of
O.OS g Ag per kilogram dry sediment in test 1  and 0.50 g Ag
per kilogram dry sediment in test 2. Pore-water Pb concen-
trations in test 2 started to increase at the 0.20-g Ag per ki-
logram spiking level, attaining a high concentration of 0.126
mg/L  at the highest Ag amendment level (3.1 g Ag per ki-
logram dry sediment). Copper concentrations were elevated to
0.046 and 0.067  mg/L in the PW at the two highest exposures
in test 2.
   Chironomus tentans larvae were unaffected in their sur-
vival at mean .dissolved PW concentrations up through 1.044
mg Ag per liter in the two Bond Lake sediment tests, corre-
sponding to a nominal sediment amendment level of 0.50 g
Ag per kilogram.dry sediment (Table 4). Survival was sig-
nificantly reduced at the 1.2 g Ag per kilogram dry sediment
amendment level. Mean organism dry weight was reduced only
at the exposures of 2:0.50 g Ag per kilogram dry sediment.
The mean total,biomass of.survivors was significantly less (p
£ 0.05) than, controls, with a decrease of 28.3%. for the two
tests combined at the 0.50-g Ag per kilogram dry sediment
level. Total biomass was significantly reduced to a very low
level at the exposure of 1.2 g Ag per kilogram dry sediment,
due to the combined significant reductions in both survival
and organism dry weight. A sharp change occurred between
the amendment levels of 0.20 and 0.50 g Ag per kilogram dry
sediment in the capacity of the sediment to bind Ag*. Survival
was similar at these two exposures, but dry weight and biomass
were reduced,'indicating that these were more sensitive end-
points than survival. The NOEC-LOEC range based upon .dry
weight and total biomass was 0.20 to 0.50 g Ag per kilogram •
dry sediment and 0.034 to 0.127 mg/L of dissolved Ag in the
PW. At a mean PW concentration of 7.979 rog dissolved. Ag
per liter, 46.7% of the larvae were dead at day 10, and biomass
of survivors  was significantly  reduced to only 6.8% of the
controls. Mortality was 100% at the highest exposure of 3.1
g Ag per kilogram dry sediment. As in West Bearskin Lake
sediment at high Ag amendment levels, elevated levels of other
metals and reduced pH also likely contributed to the observed
mortalities at these high Ag exposures. Values of LCSO for
PW and sediment were estimated at 15.1 mg dissolved Ag per
liter and  1.2 g Ag  per kilogram dry sediment,  respectively
(Table 5). Compared with the 10-d water-only LCSO value  of
0.057 mg/L for dissolved Ag, this PW LCSO value is 275 times
greater. It would appear that a very high percentage, perhaps
more than 99%, of the Ag that passed through the  0.2-|un
membrane filter was not readily bioavailable to result in acute
lethality as in the water-only toxicity test However, the PW
LOEC based upon  total biomass is only about twice as high
as the PW LC50 value. This may indicate that the process by
which Ag becomes bioavailable to produce mortalities is slow
and that Ag  more readily affects growth than survival.
   Comparison with other tests with freshwater benthos. This
study used the nitrate salt, which has been shown to be the

-------
 Silver tbxicity in freshwater sediments
                       Environ. Toxicol. Chem. 18, 1999    37
 most toxic salt of Ag [24,26,30]. Studies with other salts of
 Ag have resulted in either no toxicity or reduced toxicity to-
 ward freshwater benthic organisms. Amendment of sediments
 with the highly insoluble silver sulfide (Ag2S) at concentrations
 as high as 0.753 g/kg dry sediment did not significantly affect
 survival of the amphipod Hyalella azteca [31]. Hyaltlla az-
 teca was found to be the most highly sensitive of nine species
.to Ag in water-only acute toxicity tests [32].  This suggests
 that either none or only minute amounts of Ag* were released
 from Ag2S-spiked sediments. A bioaccumulation study with
 the freshwater oligochaete Lumbriculus variegatus and Ag2S-
 amended sediment at approximately 0.444 g Ag per kilogram
 dry  sediment also demonstrated a lack of toxicity and a low
 accumulation factor of 0.18 [33], further indicating the relative
 unavailability of Ag+ in PW of sediments amended:with the
 Ag2S salt. Ten-day LC50 values for H.  azteca were greater
 than the highest concentrations of sediment amendment levels
 tested when Ag was introduced as AgCl (>2.56 g/kg) or Ag2
 (SzOj),, (>0.569-1.12 gflcg) f26J.
   It is not believed that NH3 contributed highly to the ob-
 served effects. Chironomus teutons can tolerate considerable
 NH3, with a 10-d LC50 value in the range of 530 to 700 mg/
 L of total ammonia  [34]. The highest single measured NH,
 concentration in this  study of 100 mg/L occurred in a control
 sediment. Ammonia concentrations did not correlate with Ag ,
 concentrations. However, pH may have added to the observed
 adverse-responses in some cases. The genus Chironomus cm
 tolerate reduced pH, as it has been found to occur in nature
 in waters with pH as low as 4.4 [35]. Laboratory studies have
 also demonstrated that other members of the Chironomus ge-
 nus  can survive  acute exposure to pH 4 [36,37];  however,
 survival was reduced nearly 50%. Exposure of animals to a
 sudden large difference in pH, as would have occurred, at the
 higher exposures of this study, may have contributed to the
 observed toxicities at these higher exposures.    • '
   The endpoint of mean organism dry weight did not show
 a consistent decline with increasing Ag concentration in either
 of the sediments in this study. However, in the West Bearskin
 Lake exposures of 3.0, 4.0, and 5.4 g Ag per kilogram dry
 sediment, the mean  organism dry weights were reduced to
 0.65, 0.76, and 0.61  mg, respectively. Previous studies have
 demonstrated correlations between reduced growth of C. ten-
 tons larvae and reduced success of both emergence to adults
 and  reproduction [38,39]. These studies also indicated that a
 dry  weight of about 0.5 to 0.6 mg was the minimal weight for
 successful emergence and reproduction. Future studies of Ag
 toxicity to C. teutons that extend over the full life-cycle will
 be useful in interpreting .the effects of reduced weight upon
 reproductive success of this species!             '  "

 Freshwater sediment binding phases for silver
                                 ,       /      •
.  Both sediments selected for study had low AYS and TOC
 levels to minimize the binding capacity for Ag. Acid-volatile
 sulfide and TOC are considered to be two major metal-binding
 phases in anoxic sediments [16,18,40,41]. With the low AVS
 levels during the toxicity tests (<0.10 tunol/g), the presence
 of approximately 0.020 mg/L of Ag* in the PW (0.20 jwnol/
 ml) would have yielded a SEM[Ag/2]  - AVS difference of
 2:0.0. This could have potentially resulted in some toxicity,
 according to equilibrium partitioning theory [42]. Due to the
 low AVS content of our test sediments, SEM[Ag/2] - AVS
 was s=0.0 in all cases where significant toxicity was observed.
 However, mortalities were less than 100% in most cases, and
 toxicity was not attributable to Ag* alone because significant
 quantities of displaced Zn (plus Ni in West Bearskin Lake)
 were also' present when sufficient AgNO3 was incorporated
 into the sediment to yield appreciable quantities of dissolved
 Ag in the PW.
    This study has shown that two sediments, both with rela-
 tively low AVS and TOC levels, varied considerably in their
 capacities to bind Ag. Relatively high spiking levels of AgNO3
 resulted in the displacement of other metals from the sediment
 into the PW, with Zn being displaced in greatest concentration
 and Cu the least. The capacity for these sediments to either
 bind Ag* or to react with Ag* to form a nontoxic dissolved
 species appears to be largely dependent upon certain physi-
 cochemical characteristics of these oxic test sediments in ad-
 dition to AVS and TOC. Acid-volatile sulfide and TOC com-
 bined likely could not have bound all of the Ag* at several of
 the exposures in this study. For example, at the West Bearskin
 .Lake amendment level of 2.2 g Ag per kilogram dry sediment,
. a total Ag concentration of approximately 20.4 junol/g of dry
 sediment was added. One mole of AVS can bind 2 molof
 Ag*, and the partition coefficient (A.J for organic matter and
 Ag is approximately 1.82 X 105 L/kg,* [39]. This means that
 West Bearskin Lake sediment with an AVS level of <0.1 jimol/
 g could bind a maximum of 0.2 junol of Ag by AVS. From
 the work of Mahony et al. [40,41], it was estimated that ap-
 proximately 10 to 20 urnol of Ag* per gram  dry sediment
 could be bound by TOC in West Bearskin Lake sediment con-
 taining 1% TOC, dependent upon pH [42]. However, if fresh-
 water sediments are anoxic and have AVS levels in excess of
 one half the molar •concentration of Ag, then binding by AVS
 should apply in rendering Ag* unavailable to biota, as in the
 cases of the divalent metals [18], The binding  of Ag by the
 sediments of this study may have been somewhat reduced
 relative to that of in-place bedded sediments from the field.
 Manipulations involved in performing the tests likely reduced
 the AVS of the test sediments. Also,  cold storage has been
 shown to result in a decrease in TOC content during the first
 6 months of storage [43].
    The observation of high rates of survival of C. teutons at
 high dissolved concentrations .of Ag within the  peeper cham-
 bers  (e.g., up to ~300 times the water-only LC50 value) in-
 dicates that most of the dissolved Ag was not freely dissolved
 ionic Ag* and was not in a form that was readily available to
 cause acute lethality. Possibilities  for association with Ag*
 include complexes of Fe or manganese oxides or colloids of
 organic matter <0.2 (xm in size. Amorphous Fe oxides are
 known to be a major sink for metals in oxic sediments [27,44] _
 and were likely, important ligands for. these particular sedi-
 ments under the oxic conditions of the laboratory tests. Binding.
 site densities may be in the range of 1 to 15-u.m sites per gram
 sediment for amorphous iron oxides [45]. As more knowledge
 is  gained on .binding of metals by Fe oxides  in freshwater
 environments,  it  may be possible to develop more accurate
 predictions of metal bioavatlability, as suggested by Tessier et
 al. [46]. Ambient sediment quality criteria might be developed
 based upon concentrations of AVS plus organic carbon for
 anoxic waters and organic carbon plus amorphous Fe oxides.
 for oxygenated waters. Based upon the recent work of Berry
 ' et al. [18], it appears that AVS can be used to accurately predict
 when sediments containing Ag plus the divalent metals will
 not be toxic. A similar approach may be desirable to account
 for the additional binding phases.

-------
38    Environ. Toxicol. Chem. 18, 1999

                      CONCLUSIONS

   In summary, West Bearskin Lake sediment had a greater
capacity for binding Ag than Bond Lake sediment, likely due
in part  to its  greater proportion of fine-grained particles and
higher TOC and Fe concentrations. The AVS concentration
was also initially higher in West Bearskin Lake sediment, but
both sediments had levels below detection during the perfor-
mance of sediment toxicity tests. The 10-d LC50 values for
C. tentans,  based upon nominal additions of Ag to the sedi-
ments, were 2.75 and 1.17 g Ag per kilogram dry sediment
for West Bearskin and Bond Lake sediments, respectively (Ta-
ble 5).  Although a PW LC50 value was not calculable for
sediment from West Bearskin Lake amended with Ag, a PW
LC50 value of 15.1 mg/L for dissolved Ag was  determined
for Bond Lake sediment. This was 275 times greater than the
10-d water-only  LC50 value of 0.057 mg/L,  indicating that
most of the dissolved fraction was not readily bioavailable to
cause lethality. Ranges  of NOEC-LOEC for  West Bearskin
and  Bond Lake sediments were 1.7 to 2.2 and 0.20 to 0.50 g
Ag per kilogram dry sediment, respectively, based upon total
biomass per replicate or mean organism dry weight. The ad-
dition of Ag to both sediments resulted in the displacement of
other metals (i.e., Zn, Ni, Cu, and Pb) from the sediment to
the PW and reduced the PW pH values. These changes likely
contributed to the observed toxicity, particularly at the higher
Ag amendment levels. The concentrations of Ag in the sedi-
ments that caused significant adverse effects in C. tentans were
well above  concentrations reported in the environment.

Acknowledgement—The authors thank Gary Ankley, Dave Mount,
Russ Erickson, Walter Berry,  Dominic Di Toro and Tom Purcell for
their thoughtful input in the planning of this study. We are grateful
to Brenda Maldonado, Heidi Saillard, Jody Springsteele, Kathryn
Roche, and Matthew TenEyck, for their work as  student research
assistants, and to our secretary, Joyce Barnes, for manuscript prep*
aration.  Support for mis study was provided by, a grant from the
Photographic and Imaging Manufacturers Association.

                      REFERENCES
  1. ScowK,etal. 1981.Exposureriskassessmentforsilver.Technical
    Report. US Environmental Protection Agency, Washington, DC.
 2. U.S. Department of Health and Human Services. 1990. Tbxico-
    logical profile for silver. TP-90-24. Atlanta, GA.
 3. Cooper CF, Jolly WC. 1970. Ecological effects of silver iodide
   • and other weather modification agents: A review. Water Resour
    Res 6:88-98.
 4. Terhaar CJ, Ewell WS, Dziuba SP, Fassett DW. 1972. Toxicity of
    photographic processing chemicals to fish. Photogr Sci Eng 16:
    370-377.
 5. Terhaar CJ, Ewell WS, Dziuba SP. White WW, Murphy PJ. 1977.
    A laboratory model for evaluating the behavior of heavy metals
    in an aquatic environment. Water Res 11:101-110.
 6. Nebeker AV. 1982. Evaluation of a Daphnia magna renewal life
    cycle test method with silver and endosulfon. Water Res 16:739-
    744.                 .
  7. Nebeker AV, McAuliffe CK, Mshar R, Stevens DC. 1983. Tox-
     icity of  silver to steelhead and  rainbow trout, fathead minnow
    and Daphnia magna. Environ Toxicol Chem 2:95-104.
  8.  Hogstrand C, Wood CM.  1998. Toward a better understanding of
     the bioavailability, physiology and toxicity of silver in fish: Im-
     plications  for water quality criteria. Environ Toxicol Chem 17:
     547-561.
  9.  Eisler R. 1996. Silver hazards to fish, wildlife and invertebrates:
     A synoptic review. Contaminant Hazard Reviews Report 32. U.S.
     Department of the Interior, Patuxent  Wildlife  Research Center,
     Laurel, MD.
 10.  Shafer MM. 1995. Sampling and analytical techniques for silver
     in natural waters. Proceedings,  3rd International Conference,
     Transport, Fate and Effects of Silver in the Environment, Wash-
     ington, DC, USA, August 6-9, pp 99-108.
                                              D.J  Call et at.

11. Andren AW, Wingert-Runge B, Sedlak D. 1995. The occurrence
    and behavior of silver in natural waters. Andren A, Sober T, eds,
    Proceedings, 3rd International Conference, Transport, Fate and
    Effects of Silver in the Environment, Washington, DC. USA,
    August 6-9, pp 7-8.
12. Purcell TW, Peters JJ. 1998. Sources of silver in the environment.
    Environ Toxicol Chem .17:539-546.   •
I3i Shafer MM, Overdicr JT, Armstrong DE. 1998. Removal, parti-
    tioning, and fate of silver and other metals in waste water treatment
    plants and effluent-receiving streams. Environ Toxicol Chem 17:
    630-641.                        '
14. Van Cappellen P, Wang Y. 1995. Metal cycling in surface sedi-
    ments: modeling the interplay of transport and reaction. In Allen
    HE, ed, Metal Contaminated Aquatic Sediments. Ann Arbor
    Press, Ann Arbor, MI, USA, pp 2-64.
IS. Berry WJ, Hansen DJ, Mahony JD, Robson DL,  Shipley PB,
    Rogers B, Corbin JM.  1996. Predicting the toxicity  of metals-
    contaminated sediments using acid volatile sulfide and interstitial
    water normalization in laboratory-spiked sediments. Environ Tax-
    icol Chem 15:2067-2079.
16. Ankley GT, Di Toro DM, Hansen DJ, Berry WJ. 1996. Technical
    basis and proposal for deriving sediment quality criteria for met-
    als. Environ Toxicol Chem 15:2056-2066.
17. Hansen DJ, Berry WJ, Mahony JD, Boothman WS, Robson DL,
    Ankley'GT. 1996. Predicting the toxicity of metals-contaminated
    field sediments using interstitial concentration of metal and acid
    volatile sulfide normalizations. Environ Toxicol Chem 15:2080-
    2094.
18. Berry WJ, Cantwell MG. Edwards PA,  Serbst JR, Hansen DJ.
    1999. Predicting toxicity of sediments spiked with silver. Environ
    Toxicol Chem 18:40-48.
19. Benoit DA,  Phipps GL, Ankley  GT. 1993. A, sediment testing
    intermittent renewal system  for automated renewal of overlying
    water in toxicity tests with contaminated sediments. Water Res
    16:457-464.
20. U.S. Environmental Protection Agency.  1994. Methods for mea-
    suring the toxicity and  bioaccumulation of sediment-associated
    contaminants with  freshwater invertebrates. EPA/600/R-94/024.
    Duluth, MN.
21. U.S. Environmental Protection Agency. 1979. Method 415.1. Or-
    ganic carbon, total (combustion or oxidation). EPA-600/4-79-020.
    Cincinnati, OH.
22. Allen HE, Gongmin  F,  Boothman W, Di Toro D, Mahony JD,
    1991. Determination of acid volatile sulfides and simultaneously
    extractable metals in sediment In Draft Analytical Method far
    the Determination of Acid  Volatile Sulfide in Sediment. U.S.
    Environmental Protection Agency, Washington, DC.
23. Lazoichak JM. 1994. Computer  program containing latest ver-
    sions of: Dunnett's procedure, probit analysis, the trimmed Spear- *
    man-Karber method, and the linear interpolation method. U.S.
    Environmental Protection Agency, Cincinnati, OH.
24. Rodgers JH Jt, Deaver E, Suedel BC, Rogers PL. 1997. Com-
    parative aqueous toxicity of silver compounds: Laboratory studies
    with freshwater species. Bull Environ Contam Toxicol 58:851-
    858.
25. Khangarot BS, Ray PK. 1989. Sensitivity of midge larvae of
    Chironomus tentans Fabricius (Diptera Chironomidae) to heavy
    metals. Bull Environ Contam Toxicol 42:325-330.
26. Rodgers JH Jr, Deaver E, Rogers PL. 1997. Partitioning and ef-
    fects of silver in amended freshwater sediments. Ecotoxicol En-
    viron Saf 37:1-9.
27. Newman  MC, Jagoe CH. 1994. Ligands and the bioavailability
    of metals in aquatic environments. In Hamelink JL, Landnim PF,
    Bergman HV, Benson WH, eds, Bioavailability: Physical, Chem-
    ical and Biological Interactions. Lewis, Boca Raton, FL, USA,
    pp 39-61.
28. Phipps GL,  Mattson VR, Ankley GT. 1995. Relative sensitivity
    of three freshwater benthic macroinvertebrates to  ten contami-
    nants. Arch Environ Contam Toxicol 28:281-286.
29. Dean JA, ed. 1985.  Lange's Handbook of Chemistry, 3rd ed.
    McGraw-Hill, New York, NY, USA.
30. Hogstrand C, Galvez F, Wood CM. 1996. Toxicity, silver accu-
    mulation and metallothione in induction in freshwater rainbow
    trout during exposure to different silver salts. Environ Toxicol
    Chem-15:1102-1108.
31. Hirsch MR  1998. Toxicity of silver sulfide-spiked sediments to

-------
Silver toxicity in freshwater sediments
                        Environ. Toxicol. Chem.  18, 1999    39
 -   the freshwater ampbipod Hyalella azteca. Environ Toxicol Chem
    17:601^604.
32. Diamond JM, Mackler DC, Collins M, Grubcr D. 1990. Deri-
    vation of a freshwater silver criteria for the New River, Virginia,
    using representative species. Environ Toxicol Chem 9:1425-.
    1434.
33. Hirsch MR 1998b. Bioaccumulation of silver from laboratory-
    spiked sediments in thcoligochaete Lumbriculus variegatus. En-
    viron Toxicol Chem 17:605-609.
34. Whiteman FW, Ankley GT, Kahl MD, Rau DM, Balcer MD. 1996.
    Evaluation of interstitial water as a route of exposure for ammonia
    in sediment tests with benthic macroinvertebrates. Environ Tox-
    icol Chem 15:794-801.
35. Krantzberg G, Stokes PM. 1988. The importance of surface ad-
    sorption and pH in metal accumulation by chironomids. Environ
 .   Toxicol Chem 7:653-670.
36. Fisher SW, Lohner TW. 1986. Studies on the environmental fate
    of carbaryl as a function of pH. Arch Environ Contain Toxicol
    15:661-667.                                      ...
37. Rousch JM, Simmons TW, Kerens BL, Smith BP. 1997. Relative
    acute effects pf low pH and high iron on the hatching and survival
    of the water mite (Arrenurusmanubriator) and the aquatic insect
    (Chironomus riparhts). Environ Toxicol Chem 16:2144-2150.
38. Liber K, Call DI, Dawson TD. Whiteman FW, Dillon TM. 1996.
    Effects of Chironomus tentans larval growth retardation on adult
    emergence and ovipositing success: Implications for interpreting
    freshwater sediment bioassays. Hydrobiologia 323:155-167.   '
39. Sibley PK, Benoit DA. Ankley GT. 1997. The significance of
   .growth in Chironomus tentans sediment toxicity tests: Relation-.
41
    ship to reproduction and demographic endpoints. Environ Toxicol
    Chem 16:336-345.
40. Mahony JD, Di Tore DM, Gonzalez AM, Curto M, Dilg M,
    DeRosa LD, Sparrow LA. 1996. Partitioning of metals to sedi-
    ment organic carbon. Environ Toxicol Chem 15:2187-2197.
    Mahony JD, Di Tbro DM, McLeod PB, Page JJ, Santos UN,
    Trivedi DK. 1999. The role of AVS and organic carbon in par-
    titioning of silver to sediments. Proceedings,  5th International
    Conference, Transport, Fate and Effects of Silver in the Envi-
    ronment, Hamilton, ON, Canada, September 28-October 1, 1998,
    (in press).    .                 •
42. Di Tbro DM, et al. 1991 . Technical basis for establishing sediment
    quality criteria for non-ionic organic chemicals using equilibrium
    partitioning. Environ Toxicol Chem 10:1541-1583.
43. Sijm RTHM, Haller M, Schrap SM. 1977. Influence of storage
    on sediment characteristics and of drying sediment on sorption
    coefficients of organic contaminants. Bull Environ Contam Tox-
    icol 5S-.961-96&.
44. Tessier A. Couillard Y, Campbell PGC, Auclair JC. 1993. Mod-
    eling Cd partitioning in oxic lake sediments and Cd concentrations
    in the freshwater bivalve Anadonta grandis. Limnol Oceanogr
45. Gagnon C, Fisher NS. 1997. The bioavailability of sediment-
    bound Cd, Co and Ag to the mussel Mytilus edulis. Can J Fish
    Aqiua Sci 54:147-156.
46. Tfcssier A, Campbell PGC, Auclair JC, Bisson M. 1984. Rela-
    tionship between the partitioning of trace metals in sediments and
    their accumulation in the tissues of the freshwater mollusc Elliptio
    complanata in a mining area..Con / Fish Aquat Set 41:1463-
    1472.

-------
                                                                 Environmental Toxicology aid Chemistry, Vol. 18, No. 1, pp. 40-48, 1999
                                                                                                         Printed in the USA
                                                                                                   0730-7268/99 $9.00 + .00
                                                  Annual Review       _    •   •   -

                PREDICTING TOXICITY OF SEDIMENTS SPIKED WITH SILVER


  WALTER J. BERRY,* MARK G. CANTWELL, PHILIP A. EDWARDS, JONATHAN R, SERBST, and DAVID J. HANSEN
                U.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory,
                          Atlantic Ecology Division, 27 Tarzwell Drive, NarraganscM, Rhode Island 02882

                                        {Received 18 May 199*; Accepted 24 July 1998)


      Abstract—Previous experiments conducted with freshwater sediments.spiked with silver nave shown that, when expressed on a
      dry weight basis, the toxicity of silver is sediment-specific and dependent on the form of silver added (e.g., AgNO,, AgjS). This
      study was conducted to  assess the usefulness of silver  interstitial water toxic units (IWTU) and acid volatile sulfide (AVS)
      concentrations in predicting the biological effects of silver species across sediments, regardless of the  species of silver present.
      Two saltwater sediments were spiked with a series of concentrations of silver. The amphipod, Ampelisca abdita, was then exposed
      to die sediments in ten-day toxicity tests.  Amphipod mortality was sediment-specific  when expressed on a dry weight basis, but
      not when based on IWTU or simultaneously extracted metal (SEM)-AVS. Sediments with an excess of AVS relative to SEM had
      IWTU <0.5. and were generally not toxic. Sediments with an excess of SEM relative to AVS had silver IWTU >0.5, but no
      measurable AVS. and were generally toxic. Sediments with measurable AVS were not toxic. Reanalysis of the previously published
      data from the freshwater sediments spiked with silver showed mortality to be correlated with nominal SEM-AVS and with silver
      IWTU. taken together, these results support the use of AVS and silver IWTUs in predicting the toxicity of silver in sediments.

      Keywords—Acid volatile sulfide    Interstitial water    Silver    Sediments    Toxicity
                     INTRODUCTION
   Silver is a highly toxic metal, with a wide distribution in
the environment. Sediments are a dominant sink for silver
released into aquatic environments [1], Given that silver is one
of the most strongly bioaccumulated of the elements [2], this
combination of factors would seem to indicate that silver in
sediments poses a large environmental risk. However, little is
known about the biological effects or bioavailability of sedi-,
ment-associated silver in the field [2], This is because silver
is generally found in low concentrations as one component of
a mixture of contaminants and because speciation is extremely
important in silver toxicity [1].
   In the absence of compelling field data, one of the best
ways to gain insight into the toxicity and bioavailability of a
metal in sediment is to perform spiked-sediment toxicity tests
[3]. Relatively few studies have been performed with fresh-
water sediments spiked with silver, and none have been per-
formed with marine sediments. Hirsch [4,5] mixed silver sul-
fide powder into freshwater sediments and found that even at
very high concentrations of silver there  was no evidence of
toxicity to the amphipod Hyalella azteca [4] or of silver uptake
in the oligochaete Lumbriculus variegatus [5]. Call et al. [6]
found increased mortality to the midge Chironomus tentans
with increasing silver nitrate addition. Rodgers et al. [7], in
experiments with four freshwater  sediments, found that the
form of silver influenced the toxicity of silver to H. azteca.
Sediments spiked with silver thiosulfatc or silver chloride were
not toxic. Those spiked with silver nitrate were toxic. Although
toxicity increased with increasing silver concentration in both
studies of sediments spiked with silver nitrate, the toxicity of
silver was sediment specific when expressed on a dry weight
basis [6,7].                                ^

   * To whom correspondence may be addressed
(berry.walter@epa.gov). Contribution 1988 National Health and En-
vironmental Effects Research Laboratory, Narragansett, RI, USA.
   Numerous studies have highlighted the utility of interstitial
water (IW) metals concentrations and metals:acid-volatile sul-
fide (AVS) relationships in explaining the bioavailability of
sediment-associated metals to benthic organisms in both the
laboratory and the field [3,8,9]. In all of these studies, the
toxicity of metals was sediment specific when expressed on a
dry weight basis but not sediment specific when expressed on
an IW or AVS-normalized basis. When the IW contained less
than 0.5 toxic units of metal, or when AVS exceeded metal in
the sediment (when expressed on a molar basis as "simulta-
neously extracted metal," or SEM, the metal extracted in the
AVS procedure), the sediments did not cause acute mortality.
Conversely, when the IW contained more than O.S toxic units
of metal, and/or when SEM exceeded AVS in the sediment,
sediments generally did cause acute mortality.
   Acid-volatile sulfide has been used primarily to predict the
bioavailability in sediments of divalent cationic metals that
form  sulfides: cadmium, copper,  lead, nickel, and zinc but
should also be useful for other metals that form sulfides, such
as silver. Importantly, silver is different from the divalent tran-
sition metals because (1) silver is (mostly) monovalent, so
each mole of sulfide binds two moles of silver. For this reason,
the concentration of silver divided by two ([Ag]/2) is compared
with AVS. (2) In saltwater sediments the formation of highly
insoluble silver chloride during spiking complicates  spiking
and chemical  analysis.  (3) Equilibration times  after  spiking
can be very long, up to 90 d or more. (4) Silver sulfide and
silver chloride are not extracted with the conventional AVS
extraction.
   In  this study two saltwater sediments were spiked with a
series of concentrations of silver nitrate, such that in  the low
treatments AVS exceeded [Ag]/2 and in the higher treatments
[Ag]/2 exceeded AVS.  One sediment was coarse grained with
low AVS, and the other was fine grained with higher AVS.
The amphipod Ampelisca abdita was. then exposed to the sed-

-------
 Predicting the toxicity of silver-spiked sediments

 iments in  10-d toxicity tests to determine the utility of AVS
 and IW silver concentrations to predict acute mortality in sil-
 ver-spiked sediments across a range of sediment types. The
 purpose of this paper is threefold: to present data on a new
, study with saltwater sediments spiked with silver,  to  sum- ,
 marize the available studies on freshwater sediments spiked
 with silver, and to combine information from all of these stud-
 ies to evaluate the overall utility of AVS and IW silver in
 predicting the toxicity and bioavailability of silver  in  sedi-
 ments.

               MATERIALS AND METHODS
 Organism and sediment collection
    Ampelisca  abdita were collected following the methods
 described in Berry et al. [3] (Ampelisca abdita is an estuarine,
 tube-building, infaunal amphipod commonly used in sediment
 toxicity testing [10]). Two sediments of differing AVS con-
 centration were used: one from Pojac Point (initial  AVS =
 23.9-26.9 jimol S/g) and one from Ninigret Pond (initial AVS
'= 2.1 funol S/g). The Pojac Point sediment was collected from
 an uncontaminated site  in  northeastern Narragansett Bay
 (41°38'54"N and 71°23'41"W) with a Van Veen grab sampler,
 returned to the laboratory, press sieved wet through a 2-mm
 mesh stainless steel screen, homogenized, and stored at 4°C.
 This sediment had 1.3% total organic carbon (TOC) and 5%
 sand and 95% silt/clay. The Ninigret Pond (Charlestown, RI,
 USA) sediment was collected with a shovel.  The upper 5 to
 10 cm of sediment was returned to the laboratory, sieved wet
 through a  1.0-mm stainless steel screen, rinsed several times
 to remove high-organic fine particles, homogenized, and stored
 at 4°C. This sediment had 0.2% TOC and was made up. of
 100% sand.

 Water-only tests
   . Ten-day static renewal tests were conducted with A abdita
 to determine water-only concentrations causing 50% lethality (
 (LCSOs) for silver in sand-filtered Narragansett Bay seawater
 (31 ± 1 ppt, 19 ± 1°C) following the methods described in
 Berry et al. [3], except that some water samples from the tests
 were filtered before they were analyzed. Unfed amphipods
 were exposed to five concentrations of silver  nitrate (Fisher*
 certified ACS reagent, Fisher Scientific,. Fairlawn, NJ) and a
 control, with ten amphipods per replicate and two replicates
 per concentration.

 Spiked sediment tests
    Sediment spiking and equilibration. Sediments were spiked
 .with a solution of silver nitrate salts as described in Berry et
 al. [3]. The spiking solution was prepared by dissolving silver
 nitrate into 18 Mfl deionized water while maintaining the pH
 at 8.1 ± 0.2. Each jar was purged  with nitrogen, capped, and
 rolled for  IS min/d at 4°C during the equilibration period.
    The slow kinetics of silver  sulfide formation necessitated
 the sampling of sediments at 10- to 14-day intervals for AVS
 concentration during the up to 106-d equilibration period to
 ascertain when equilibrium was reached. At each sampling
 period the sediment in each jar from selected (or all) treatments '
 was lightly homogenized, and 30-ml samples of sediment were
 removed for AVS analysis. Before capping, the jar was again
 purged with nitrogen and returned to the cooler. Jars from the
 other treatments were not opened until test initiation.
    Sediment  toxicity test method.  Exposure and sampling
 methods for sediment, IW, and overlying water followed those
                       Environ. Toxicol. Chem. 18, 1999    41

 used in Berry et al. [3]. Amphipods were exposed to control
 and metal-spiked sediments in 10-d tests with continuous re-
 newal of sand-filtered Narragansett Bay seawater (31 ± 1 ppt,
 19 ± 1°C). -The nominal treatments spanned the range from
 sediments that were nominally spiked with less [AgJ/2 than
 initial AVS to sediments which had an excess of {AgJ/2 over
 AVS (See Table 2 for concentrations). For each sediment there
 were six treatments plus a control. Each treatment included
 two biological replicates to assess mortality and two chemical
 replicates for interstitial water, metal, and AVS analyses of the
 sediment at test initiation and termination. Diffusion samplers
 were placed in each  biological replicate and one  chemical
 replicate when the sediments were added to the test chambers.
 Twenty amphipods were added to each of these replicates at
 the start of the test (the next day). The day 0 chemistry rep-
 licates did not get amphipods or diffusion, samplers.        '
   One sediment toxicity test was 'conducted with sediment
 from Ninigret Pond. The sediment toxicity test with Pojac
 Point sediment was repeated (tests are referred  to as Pojac 1
 and Pojac3) because there was unexplained toxicity in one of
 the intermediate concentrations in the first sediment toxicity
 test with this sediment.
  . Interstitial water silver concentrations were taken from the
 diffusion samplers on day  10. Bulk metal, AVS, and SEM
 samples were taken on day 10. The bulk metal samples from
 Pojacl and 2 were not analyzed.

 Chemical analyses
   Water analysis. Interstitial water (from diffusion samplers)
 and overlying water were analyzed for silver • using graphite
 furnace atomic absorption spectroscopy (GFAA). Detection
 limits for silver by GFAA were I (*g/L.
   Sediment analysis. Sediment samples  were analyzed for
 AVS by the cold-acid purge and trap technique [11,12] using
nitric acid in place of HCL (Nitric acid was used in place of
HC1 because of the low solubility of silver chloride. Side-by-
side testing with HC1 showed little difference in the measure-
ment  of AVS or SEM if nitric acid or HC1 was used; M.
Cantwell, unpublished data.) Simultaneously extracted metal
and bulk metals analyses were performed using inductively
coupled plasma atomic emission spectrometry (ICP-AES). For
analyses of bulk metals excluding silver, the metals were ex-
tracted from freeze-dried sediments by microwave digestion
with HF/HNOj/HCl  acids followed by filtration. Bulk silver
in the sediments was extracted by microwave digestion in 20
ml of HC1 followed by  a postdigestion addition of 5 ml of
NH^OH. Total metals analyses of sample blanks and recoveries
of known metal additions demonstrated 85 to 100%'recoveries
from sediments, 85 to.l 15% recoveries from sample extracts,
and an absence of contamination in our analytical procedures. •
The SEM concentration reported is the sum of cadmium, cop-
per, lead, nickel, and zinc concentrations and one half of the
silver concentration  on a micromole per gram dry sediment
basis. Concentrations of all metals in sediments exceeded an-
alytical detection limits.
  • Ammonia. Samples were taken to quantify the ammonia
concentrations in the IW and test the hypothesis that ammonia
was the cause of the unexplained mortality in an intermediate
treatment in the experiments using Pojac Point sediments. Dif-
fusion samplers were placed in sediment left over from the
third Pojac Point experiment, 1 month after the experiment
was run. Ammonia was measured using an Orion ammonia
probe (Boston, MA), and pH was measured using  a Ross com-

-------
42    Environ. Toxicol. Chem. 18, 1999

                    Table 1. Mortality of Ampelisca abdita exposed to silver nitrate in two 10-d water-only tests*
                             Test 1
          Test 2
Nominal
silver (pg/L)
Control
6.5
11
18
30
50'
84
140
Nominal
LC50 = 44 u,g/L
% Mortality
15
10
20
0
50
80
60
80


Nominal
silver (pg/L)
Control
6.5
11
18
30
50
84
140
Nominal
LC50 = 27 j»g/L
Mean measured
silver (|Lg/L) <
Control
3.57
7.05
12.8
22.6
39.9
76.8
128
Measured
LC50 = 20 (ig/L
In Mortality
0
35
15
50
45
80
60
55
c

                 • LC50 = concentration causing 50% lethality.
                                           WJ. Berry et al.
bination pH electrode. The ammonia probe was calibrated us*
ing a four-point standard curve. Un-ionized ammonia concen-
trations were estimated using the Hampson model [13].

Data handling
   Ten^day LC50 values for the water-only tests were calcu-
lated by the trimmed Spearman-Karber method [14]. Detection
limits were calculated for all chemical analyses based on in-
strument detection limits and sample size. In those instances
where a mean concentration is a summation of measured data
and data below the limit of detection, one half the detection
limit was used for those values below the limit of detection.
Means for which there are no measured values above the de-
tection limit are indicated as  ND in the appropriate tables and
graphs.
   For illustrative purposes, sediments that caused >24% mor-
tality were classified as toxic. Mearos et al.  [15] found that
sediments that caused <24% mortality in tests with the am-
phipod Rhepoxynius abronius were not consistently classified
as toxic.  This criterion is similar to the  "80% of control sur-
vival"  criterion  used in the U.S. Environmental  Protection
Agency Environmental Monitoring and Assessment Program
(EMAP)  for sediment tests with A. abdita [16]. A horizontal
dashed line at 24% mortality has  been included on the appro-
priate figures for reference.
   Because silver is essentially monovalent,  2 mol of silver
are required to bind with 1 mol of sulfide. For this reason it -
is appropriate to use one half of the silver concentration, rep-
resented as silver/2 or [Ag]/2, in comparisons with AVS con-
centrations. This stoichiometry. would suggest that any sedi-
ment with an excess of sulfide relative to one half of the silver
concentration, such that ([Ag]/2) - AVS < 0.0, should not be
acutely toxic because of silver. Therefore, in this study, SEM
is defined as the sum of one half of the concentration of silver
plus the sum of the concentrations of cadmium, copper, lead,
nickel, and zinc in the SEM extract.  A dashed vertical line at
SEM - AVS - 0.0 is included on the appropriate figures for
reference.
   Many of the TW concentrations in this paper are expressed
as toxic units. A toxic unit is the measured water concentration
divided by the water-only LC50 concentration for that partic-
ular compound and test organism. For  example, a sediment
with an IW concentration equal to the water-only LCSO con-
centration for the test organism would have 1.0 IW toxic units
(IWTU). When more than one toxic  metal is  present, IWTUs
are calculated as the sum of the toxic units of the individual
 metals, e.g., IWTU^a, » (IW cone AgVLCSO^ + (IW cone
 Zn)/LC50zJ. Thus, if IW is the principal source of metal tox-
 icity, and availability of metals is the same in water-only tests
.and IW in sediment tests, 50% mortality would be expected
 with sediments having 1.0 IWTU. A dashed  vertical line al
 IWTU = 0.5 is included on the figures in this paper to indicate
 sediments unlikely to cause significant mortality. This value,
 was selected because on  the  average, water-only LCD  and
 LCSO values differ by approximately a factor of two [17] and
 because the data in our earlier experiments with other metals
 [3]  support this value as a break point  between toxic  and
 nontoxic sediments. Only silver IWTUs are reported because
 the concentrations of other metals in the IW  from these ex-
 periments were lexicologically insignificant  (Berry, unpub-
 lished data). Calculation of IWTUs was based solely on de-
 tectable metal concentrations.

               RESULTS AND DISCUSSION  .
 Water-only tests
   Two 10-day, water-only definitive tests were conducted (Ta-
 ble  1). Chemistry samples from the first definitive were lost,
 so the test was repeated. Silver concentrations from the filtered
 versus total and pre7 versus postrenewal samples from the
 second  test were nearly identical (Berry, unpublished data) so
 only the mean prerenewal, total  silver concentrations are re
 ported.  Based on nominal concentrations, the LCSOs from the
 first and second definitive tests were 44 and 27 n-g/L, respec-
 tively. The measured 10-day LCSO from the second test of 20
 jig/L was used to calculate silver IWTUs in the sediment test,*.
 with A. abdita.

 Sediment tests
   Sediment equilibration.  Acid-volatile  sulfide concentra-
 tions were measured to determine the time to equilibration. It
 took weeks or months for AVS concentrations  in the sediment
 to equilibrate after spiking, much longer than required for
 cationic metals used in other experiments [3]. This is presum-
 ably because the dissolved silver nitrate is rapidly converted
 to solid silver chloride when mixed with seawater. This silver
 chloride must first dissolve before silver sulfide can form. In
 anoxic sediments, almost all of the silver will end up as silver
 sulfide, unless the  supply of sulfide is exhausted, because the
 solubility product constant of silver sulfide is  so low, relative
 to that  of other possible complexes [18]. (See [2] for a dis-
 cussion of silver complexation in oxic sediment.) Silver sulfide

-------
 Predicting the toxicity of silver-spiked sediments

     100i
    0.01
              14
 Fig. 1. 'Acid-volatile sulfidc (AVS) as a function of time after spiking
 the sediment used in the Pojac 3 experiment. Treatments are labeled
 with nominal ([Ag J/2) - AVS concentrations in niicromoles per gram
 dry weight sediment.
is not soluble in the AVS extraction, so the measured AVS
concentration decreases as the sulfide is formed. The pattern
of change in AVS with time during the Pojac 3 equilibration
(Fig. 1) was typical of all the spikings. The AVS concentration
in the control held near the prespiking level  throughout the
equilibration period. In low treatments  there was an initial
decrease in AVS, but AVS was relatively constant thereafter.
In intermediate  concentrations the AVS  started at a reduced
level and then decreased below detectable levels, in the higher
                       Environ. Toxicol. Chem. 18, 1999    43

 concentrations, AVS was below detection at the first sampling.
 Sediment toxicity tests were initiated when AVS was no longer
 detectable in the treatments with decreasing AVS concentra-
 tion. Sediment equilibration times for Pojac  1 and 3 and Ni-
 nigret were 91, 112, and 48 d, respectively. The equilibration
 time of the Ninigret sediment may have been shorter because
 of its lower initial AVS  concentration.
.  Overlying water. Silver was not detected in the overlying
 water in all but the two highest treatments of each of the
 sediment experiments (Table 2). Even in  those treatments in
 which silver was detected in the overlying water, concentra-
 tions were generally a factor of five to ten lower than in the
 pore water. For this reason, and because there was 100% mor-
 tality in the [Ag]/2 - AVS - 52.6 treatment in the Pojac 3
 sediment, which did not have detectable silver in the overlying
 water, (detection limits were well below the  10-d LC50 con-
 centration), it seems unlikely that the overlying  water had a
 controlling effect on mortality.
   Simultaneously extracted metal. Silver concentrations in
 the SEM solution were very low in all sediment tests, even at
 extremely high silver concentrations (Table 2). There was some
.evidence of release of other metals into the  SEM as a result
 of silver addition (Berry, unpublished data), but none of these
 releases appeared to be of any toxicological significance.  '
   Amphipod mortality in relation to sediment and pore-wa-
 ter silver. Mortality and available chemistry  data for the
 spiked-sediment experiments are summarized in Table 2. Mor-
 talityxoncentration relationships were sediment-specific when
Table 2. Results of 10-d sediment toxicity tests with Ampelisca abdita exposed to silver-spiked sediments from Pojac Point and Ninigret Pond*
Treatment
(nominal
[AgJ/2 - AVS)
(nmol/g)
Mean
measured
AVS
(|URO]/g)
Adean
measured
SEM
(limol/g)
Mean,
measured
SEM - AVS
(junol/g)

Mean
pore-water
WTU
Mean
overlying
water
IWTU

Mean bulk
measured
(«5/g)
j


% Mortality
                                                         Pojac 1
Control
-21.5
-16.1
-5.38
16.1
59.2
, 242
242
27.2
17.1
7.4 .
ND
ND
ND
ND
— .
1.28
0.84
1.24
1.8
2.3
: 2^4
2
—
-25.9
-16.2
-6.16
1.8
2.3
2.4
1.98
—
0.1
0.15
0.15
0.45
7.45
67.6
81.8
— • •
ND . — "
ND —
ND —
ND * —
ND —
5.67 —
11.7 — .
- . — —
5
7.5
2.5
71
2.4
93
- 100
too
                                                         Pojac 3
Control
-14.3
-4.78
14.3
52.6
129
454
18
4.9
ND
ND
ND
ND
ND
1.01
0.87
0.83
• 1.26
1.6
1.64
29.2
-17
-4.03
0.83
1.26
1.6
1.64
29.2
0.05
0.27
1.28
4.84
. 87.6
43.4
34.7'
ND
ND
ND
ND ,
ND
8.34
7.45
2.6
2,060
3,210
6,550
10,900
23,300
. 69,800
10
-7.5
100
12.5
100
100
100
                                                         Ninigret
Control
-1.68
-1.26
-0.42
1.26
4.62
18.9
0.7
0.6
0.9
0.2
ND
ND
ND
0.04
0.04
0.04
0.04
0.04
0.12
0.05
-0.66
-0.56
-0.86
-0.16
0.04
. 0.12
0.05
0.1
0.15
0.58
0.52
4.04
' 61.9
89.6 .
ND
ND
Nb
ND
ND
2.02
6.48
2.6
' 82.7
146
302
621
1,320
4.260
15
30 •
15
27.5
17.5
100
100
 * AVS = acid-volatile sulfide; SEM = simultaneously extracted metal; FWTU
 * Data not available.
      interstitial water toxic unit.

-------
       Environ. Toxicol. Chem. 18, 1999
                 10      100      1000    10000    100000
                      Silver (ug/g Dry Wt)
B 100,
80-
£
o 40-
# 20-
0.

' + Pojacl
h XPOJ8C3

• *
• A 	
f
A.
X*
	 x 	 1


— i 	
        -30    -20
                     0     •     20
                    [Ag/2]-AVS
                                 40
C 100
80.
g
0 40.
# 20.
0.

ANin
" 4 Pojacl
• xPojac3
A
***
x>
i
1
I
__jk 	 .
X f . '


	 *J_ 	 	
0.01
0.1
                                          10
100
  t
  I
IWW
80-
60-
40-
20-
n.

ANin
4 Pojaci
- XPojacS


w 	 	 _____
T ND
I
'1 m
! , '*.,*.
       0.01
0.1
                        1
                 AVS(uMotes/g)
100
 Fig. 2, Percentage monality of the amphipod Ampelisca abdita as a
 function of dry weight silver (Ag) concentration (A), ([Ag]/2) - acid-
 volatile sulfide (AVS, B), interstitial water toxic units (IWTU, C),
 and measured AVS (D) in two saltwater sediments spiked with Ag.
 Nin =» Ninigret Pond sediment; Pojac '= Pojac Point sediment. Sed-
 iments below the dashed line at 24% mortality are not considered
 toxic. "Vertical dashed lines at simultaneously extracted metal (SEM)
 - AVS = 0 (B) and IWTU = 0.5 (C) indicate predicted break points
 in toxicity. Data points believed to be the result of interstitial water
 ammonia are included but highlighted and not connected by lines in
 A. AVS detection limit is'indicated as ND in D.
 amphipod monality was plotted against dry weight silver (Fig.
 2A). Although mortality increased with added silver in both
. Ninigret and Pojac Point sediments, more silver was required
 on a dry weight basis in Pojac Point sediment for the sediment
 to be toxic. In both Ninigret Pond and Pojac Point sediments,
 those with an excess of AVS relative to SEM (i.e., SEM -
 AVS < 0) had IWTU <0.5 and were generally not toxic (Fig.
                                           W.J. Berry et.al.

 2B and C). Sediments with an excess of SEM relative to AVS
 (i.e., SEM - AVS > 0) had measurable silver present but no
 measurable AVS. In these sediments silver IWTUs > 0.5 and
 were generally toxic (Fig. 2B and C). No sediments in which
 measurable AVS was present were toxic (Fig. 2D).
   The results of the sediment toxicity tests with Pojac Point
 sediment indicate a source of toxicity other than silver may
 be present. This  is primarily because the observed toxicity at
 the intermediate treatments does not correlate with nominal or
 measured silver  concentrations. We  suspected ammonia tox-
 icity based on a  preliminary test with Pojac sediment (Pojac
 2), which suggested that this toxicity was removed if clean
 water was run over the sediment for 10 d (Table 2). Purging
 with clean water is routinely used in  dredging programs when
 sediments have pore-water ammonia levels above  a specific
 concentration [19]. Measurement of ammonia in sediment left
 over from the Pojac 3 experiment supported this suspicion;
 un-ionized  ammonia was highest in the treatment with the
 anomalous  toxicity (Berry, unpublished data). The calculated
 concentration of un-ionized ammonia was almost one half of
 the LC50 value of 0.82 reported for this species at a similar
 pH [20]. It is also possible that the  presence of silver in the
 IW may have exacerbated the toxicity of ammonia  in the IW
 [21]. The data points showing this anomalous mortality  in
 Pojac 1 and 3 have been included in the figures that follow
 for completeness. However, to better illustrate the patterns of
• mortality attributable  to silver in the absence of ammonia,
 these data points are not connected by lines in Figure 2A and
 have been highlighted in Figure 2B  to D.
   Analysis of data from freshwater silver-spiked sediment
 tests. Rodgers et al [7] and Call et al [6] measured AVS con-
 centrations before spiking but did not present plots of mortality
 in relation  to AVS. Rodgers et al.  [7] (E. Deaver, personal
 communication)  demonstrated that the concentration response
 was not sediment specific when normalized to ([Ag]/2) - AVS
 (Fig. 3A). Silver nitrate-spiked sediments in which there was
 an excess of AVS relative to added silver, i.e., ([Ag]/2) - AVS
 <0, were not toxic to H. azteca; sediments in which there was
 an excess of silver, i.e., ([Ag]/2) - AVS >0, were generally
 toxic (Fig.  3A),  (For the freshwater studies, for which SEM
 is hot available,  a dashed line at nominal (JAg]/2) — AVS  =
 0 is  included on the appropriate figures for reference.) Mor-
 tality was sediment specific when expressed as dry weight (Fig.
 3B).
    As was  shown with A. abdita in the saltwater experiments
 and H. azteca in the experiments of Rodgers et al. [7],  sedi-
 ments with an excess of AVS over silver were not toxic to C.
 tentans [6] (D. Call, personal communication). However, most
 of the sediments with an excess of silver were toxic  (Fig. 4A).
 Call et al [6] also measured IW metal concentrations and found
 very little silver in the IW, even in sediments in which silver
 greatly exceeded AVS. Although mortality was generally low
 when IWTUs were less than 0.5, and increased when IWTU
 were above 0.5, there appears to be an almost log-linear re-
 lationship between mortality and FWTU (Fig. 4B).  This is in
 contrast to the pattern seen in the saltwater data (Fig. 3C) and
 that  seen in most other studies [3].  For several.of the treat-
 ments, the contribution of other metals to the combined IWTUs
 was significant, with zinc and copper adding up to 4.9 IWTUs.
 This was presumably because the added silver caused disso-
 lution of zinc and copper sulfide in these sediments, releasing
 them into the pore water [3,11]. This underscores the impor-
 tance of using the sum of the IWTUs of all 'metals together.

-------
 Predicting die toxicity of silver-spiked sediments
  £
  I
                                Environ. Toxicol. Chem. 18, 1999
1UU-
80
60
40
20
n.


• Sed1
• Sed3
ASede
mStOT
\
I
I
'
I 	 IP ^ • 	 *
I
|
t
	 -M 	 B 	 i 	 1 	 1 	 1 	 : 	
                     -1     0      1.2

                       [Ag/2J-AVS(nMoles/g)
       0.01
    1         10        100
Dry wt silver (ug/g)
1000
 Fig. 3.  Percentage mortality of the amphipod Hyalella azteca as a
 function of nominal ([AgJ/2) - acid-volatile sulfide (AVS, A) and
 dry weight silver (Ag) concentration .(B) in four freshwater sediments
, spiked with Ag. Sediments below the dashed line at 24% mortality
, are not considered toxic. A vertical dashed line at simultaneously
 extracted metal (SEM) - AVS - 0 (A) indicates the predicted break
 point in toxicity.
 Mortality of C. tentans from Call et al. [6] (D. Call, personal
 communication) is plotted against dry weight concentration in
 Figure 4C. Significant mortality was seen at concentrations
 greater than approximately 1,000 u,g/g. That the dry weight
 data (Fig. 4C) seems to be as predictive of toxicity as AVS
 and IWTU normalized data (Fig. 4A and B, respectively) is
 due, at least in part, to die fact that only one sediment was
, tested by Call et al. [6].
    Laboratory results in relation to sediment quality guide-
 line concentrations and concentrations found in the field. One
 way to put the spiked-sediment mortality data reported in this
 paper in context  is to compare them  with commonly used
 empirically derived sediment quality guideline concentrations.
 The threshold effect level (TEL) is used as a concentration
 below which toxicity would not be expected in a marine field
 sediment; the probable effect level (PEL) is used as a con-
 centration above which there is a higher probability of adverse
 biological effect in a marine field sediment. The available dry
 weight results of the saltwater experiments (Fig. 2A) can be
 compared, for example, with the TEL for silver (0.733 jig/g)
 and the PEL for silver (1.77 jig/g) [22]. There were no acute
 effects in many of the exposures at concentrations which, on
 a dry weight basis, might be predicted to be toxic in the field,
 even when the PEL was exceeded by several orders of mag:
 nitude. The apparent effects threshold (AET, 4.5 j*g/g [23]) is
 a value that, if exceeded, should always be toxic to Hyalella.
 The dry weight data from the freshwater experiments (Figs.
 3B and 4C) can be compared with  a draft Hyalella AET for
 silver. Compared  with the freshwater mortality data in Figures
100



80

60
40-

20

1
0-

I
I •
i
1
. •
•
•
•
I •
I
	 •$— » 	 	 	 	
1
j .

                                                                                               45
                                                                             -10    0    10    20    30   40    50
                                                                                   [Ag/2]-AVS(nMoles/g)
                                                                      B


                                                                      £
                                                                      |
\
80
60
40
20
0^
j .
I
> , j •
' '. • 1
"
, , i , '
                      0.001    0.01      0.1    .1       10
                        Interstitial Water Toxic Untts (IWTU)
               C  100
                    80
               r-
                    20
                      0.01   0.1    1     10   100   1000 10000

                                Silver (Ug/g Dry Wt)

          Fig. 4. Percentage mortality of the midge Chironomus tentans as i
          function of nominal ([AgJ/2) - acid-volatile sulfide (AVS, A), inter-
          stitial water toxicity unit (IWTU, B), 'and dry weight Ag concentration
          (C) in a freshwater sediment spiked with Ag (Call et al., unpublished
          data). Sediments below the dashed line at 24% mortality are not con-j
          sidered toxic. Vertical dashed lines at SEM -  AVS  =  0 (A) and
          IWTU = 0.5 (B) indicate predicted break points in toxicity.
          3B and 4C, this level appears to be lower than all but one of
          the sediments in the study of Rodgers et al. [7], which had
          very low AVS concentrations. These data indicate that these]
          sediment quality guidelines may be useful as screening level
          concentrations for the effects of silver in sediments, although
          they may be highly overprotective in some sediments. There-
          fore, dry weight measures should  not be used to indicate that
          a toxic effect is due specifically  to  silver  in a sediment in
          which they are exceeded. If a chemical concentration exceeds
          an empirically derived value, it does not necessarily mean the

-------
46    Environ. Toxicat. Cheat. 18, 1999
                                                            WJ. Berry et al.
                  Table 3. Contingency table for predicting toxicity due to metals from measurements of silver (Ag),
                               other SEM metals, and AVS in laboratory-spiked and field sediments
                  Nominal metal
                  and AVS in sediment
  Measured metal and
   AVS in sediment
      Prediction,of acute toxicity
                                                         Ag only
                  [Ag]/2 < AVS


                  [Ag]/2 > AVS
 AVS > detection limit
   and (SEM - AVS)
   <0.0
 AVS < detection limit
.   and (SEM - AVS)
   >0.0
 Sediment not acutely toxic due to Ag;
   no metals detectable in interstitial
   water
 Sediment may be acutely toxic due to
   Ag (but not if IWTU < 0.5)
                                                     Metals mixtures
                 «Ag]/2 + fCd] + [Cu] + [Ni]
                    + [PbJ + [Zn]) < AVS

                 ([Agl/2 + JCd] + [Cu] + [Ni]
                    + [PbJ + [Zn)) > AVS but
                    [Ag]/2 < AVS
                 ([Ag]/2 + [Cd] + [Cu] + [Ni]
                    + [Pb] + IZnJ) < AVS and
                    [Ag]/2 > AVS
 AVS > detection limit
   and (SEM - AVS)
   <0.0
 •AVS > detection limit
   and (SEM - AVS)
   > 0.0
 AVS < detection limit
   and (SEM - AVS)
   >0.0
 Sediment not acutely toxic due to these
   metals; no metals detectable in inter-
   stitial water
 Sediment may be acutely toxic due to
   these metals but not silver directly
   (but not if sum of IWTU < 0.5)
 Sediment may be acutely toxic due to
•   Ag and/or the other metals (but not if
   sum of IWTU < 0.5)
                 • SEM = simultaneously extracted metal; AVS = acid-volatile sulfide; IWTU = interstitial water toxic
                  unit; Cd = cadmium; Cu = copper; Ni - nickel; Pb - lead; Zn = zinc.
chemical caused the effect Rather, these field-derived values
are concentrations that have been associated with an effect.
   Another way to put the dry weight toxicity data in context
is to compare it with silver concentrations found in field sed-
iments. The National Sediment Inventory (NSI) [24] includes
data from both fresh and saltwater sediments of the United
States. In the NSI database, approximately 90% of the sedi-
ments collected in the field had less than 1.0 jig/g dry weight
silver, whereas approximately 99% of these sediments had less,
than 10.0 pg/g dry weight silver (J. Keating, personal com-
munication). Bearing in mind that several of the sediments
used in the studies shown in Figures 2A, 3B, and 4C had low
AVS concentrations, the results of the spiked-sediment tests
would indicate that most field sediments do not contain enough
silver to cause an acute effect if silver was acting alone. How-
ever, silver may also contribute to metals toxicity by binding
with available sulfide, freeing up other metals in the way that
spiking with silver increased the bioavailability of zinc and
copper in the study of Call et al. [6].
 .  Bioaccumulation and chronic toxicity. Hirsch [5] did not
find any accumulation of silver in oligochaetes when they were
exposed to sediments spiked with high concentrations of silver
sulfide. This is consistent with  the lack of biological avail-
ability of other metal  sulfides [8]. However, several studies
have shown divalent metals are sometimes accumulated from
sediments with an excess of AVS over metal [see 25]. Also,
some benthic organisms can accumulate metals from the over-
lying water [26], and silver can also be accumulated from food
[27].  Further, it should be  noted that, with  the exception of
the [5] bioaccumulation study, all of the studies discussed in •
this paper have examined acute  mortality. It has been shown
that the same methods used to predict acute toxicity can be
used to predict biological effects in chronic sediment tests with
some metals [28-30], but this has yet to be shown with silver.
   Predicting the toxicity of silver in field sediments. Acute
toxicity predictions in sediments with various concentrations
of silver and the other SEM metals in relation to AVS are
                 presented in Table 3. Consider first the unlikely sediment in
                 which silver is the only toxic metal present. If AVS exceeds
                 silver in this sediment,.the AVS should be above  detection
                 limits and will exceed silver in the SEM. Silver should not be
                 present in the IW, and this sediment should not be toxic because
                 of silver. If silver exceeds the sulfide in the sediments, there
                 will be no detectable AVS in the sediment, and even a small
                 amount  of silver in the SEM will exceed the measured AVS.
                 This sediment may be acutely toxic due to silver. (It may, of
                 course, be toxic due to something else.) It should  be  noted
                 that the lack of detectable AVS in a sediment does not imply
                 that all of the available sulfide is bound by silver. The sediment
                 could be completely  oxic or the AVS could be bound by an-
                 other metal that is not entirely acid soluble, such as copper
                 [3]. If IW silver measurements are available, further refinement
                 of the prediction of toxicity is possible. Sediments that do not
                 have lexicologically  significant amounts of silver in the IW
                 should not be toxic due to silver, even if silver exceeds AVS
                 in the sediment. However, any sediment in which silver ex-
                 ceeds AVS should be looked at carefully. Sediments that do
                 have lexicologically  significant amounts of silver in the IW
                 are certainly potentially toxic due to silver (Fig. 2G).
                    Next consider the  much more likely case of a sediment that
                 is contaminated with silver and other sulfide-forming metals.
                 A mole of sulfide will be bound for every 2 mol of silver (the
                 1:2 ratio is due to the fact that silver is monovalent) present
                 because the sulfide solubility product of silver is lower than
                 that of other metals  [18]. If sulfide  exceeds the sum of the
                 [Ag]/2 plus the other SEM metals (cadmium, copper, lead,
                 nickel, and zinc), measurable AVS  will exceed measurable
                 SEM, these metals should  not be present in lexicologically
                 significant concentrations in the IW, and the sediment should
                 not be acutely  toxic due to these metals. If the sum of the
                 [Ag]/2 plus the other SEM metals exceeds the sulfide in the
                 sediment, measurable SEM will exceed measurable AVS, and
                 the sediment may be  acutely toxic due to these metals. As was
                 true with the silver-only case described previously, sediments

-------
 Predicting the toxicity of silver-spiked sediments           t

 that do not have lexicologically significant amounts of these
 metals in the IW should not be toxic due to these metals, even
 if SEM exceeds AVS in the sediment. However, any sediment
 in  which SEM exceeds  AVS should be  looked at carefully.
 Sediments that do have lexicologically significant amounts of
 metals in the IW are certainly potentially toxic due to these
 metals [3].
    Theoretically, almost all of the silver  in the sediment will'
 be bound to sulfide in any sediment in which there is an excess
 of sulfide over [Ag]/2. This is because of the extremely low
 solubility of silver sulfide {18]. However,  if the sum of the
 [Ag]/2 plus the  other SEM metals exceeds the sulfide in the
 sediment, some of the other SEM metals may be present in
 the IW, and the sediment may be toxic due to these metals. In
 these sediments, the silver may be contributing to the overall
 toxicity of metals in the sediment by tying up sediment sulfide
 that might otherwise bind the other SEM metals. This is ap-
 parently what happened  in some of the sediments of Gall et
 al. [6], in which zinc and copper  were released into IW due
 to  the addition of silver to the sediment..
    Hie release of metals  into the IW in relation to sulfide
 solubility is not peculiar to silver. Berry  et al. [3] found that
 metals appeared in the IW in the order of their Kv, with copper
 appearing Jast in their experiments. What is unusual about
 silver is that the solubility of silver sulfide in the  AVS ex-
 traction is so low that any sediment with an excess of [Ag]/2
 over sulfide will have no measurable AVS  present (Table 2).
 Thus, any sediment with  measurable AVS should  not have
.silver in the IW and should not be acutely toxic because of
 silver.                          "                .

                       CONCLUSION

    It is important to remember that the data presented in this
 paper apply primarily to acute mortality and may not address .
 all effects due to chronic exposure or bioaccumulation [2]. For
 example, Hook and Fisher [31] demonstrated in preliminary
 experiments that silver may effect copepod  reproduction at
 environmental concentrations. However, taken  together,  the
 freshwater and saltwater sediment results indicate that silver: •
 AVS relationships and IWTU can provide insight into the role
 of silver in the possible toxicity of sediments. From the point
 of view of AVS and SEM measurements, these results would
 indicate that silver can be included along with cadmium, cop-
 per, lead, nickel, and zinc in sediment assessment. If the sum
 of the SEM for  these metals is less than AVS in a sediment,
 the sediment should not be acutely toxic due to these metals.
 Furthermore, even in sediments that have an excess of metal
 over sulfide, as long as there is measurable AVS, any observed
 acute mortality should not be due directly to silver in the pore
 water.
- Acknowledgement—We thank R. Burgess, C. Ingersoll, R. Pruell, W.
 Boothman, and two anonymous reviewers for comments that im-
 proved the manuscript. The manuscript also profited from conversa-
 tions about experimental design and data interpretation with D. Di
 Toro and J. Mahony. D. Call and E. Deaver provided raw data from
 their experiments and graciously consented to having them replotted.
 This document has been'reviewed in accordance with .U.S. Environ-
 mental Protection Agency, National Health and Environmental Re-
 search Laboratory, Atlantic Ecology Division policies and approved
 for publication. The contents of this publication do not necessarily
. reflect the views of the U.S. Environmental Protection Agency. Men-
 tion of trade names or commercial products does not constitute en-
 dorsement or recommendation for use.
                       Environ. Toxicol. Chem. 18, 1999    4

                      REFERENCES
 1.  Purcell TW, Peters J. 1998. Sources of stiver in the environmeni
    Environ Toxicol Chem 17:539-546.                       ||
 2.  Luoma SN, Ho YB, Bryan GW. 1995. Fate, bioavailability ami
    toxicity of silver in estuarine environments. Mar Pollut Bull 31
    44-54.
 3.  Berry WJ, Hansen DJ, Mahony JD,  Robson DL, Shipley BF
    Rogers B, Corbin JM. 1996. Predicting the toxicity of metals-
    contaminated sediments using acid volatile sulfide and interstitial
    water normalization in laboratory-spiked sediments. Environ Top-
    ical Chem 15:2067-2079.                               I
 4.  Hirsch MP. 1998. Toxicity of silver sulnde-spiked sediments tej
    the freshwater  amphipod (Hyalella azteca).  Environ Toxico,
    Chem 17:601-604.
 5.  Hirsch MP.  1998. Bioaccumulation of silver from laboratory
    spiked sediments in the oligochaete (Lumbriculus variegatus)
    Environ Toxicol Chem I7:605-;6d9.                       I
 6.  Call DJ, Polkinghorne CN, Markee TP, Brooke LT, Geiger DL, Gor'
    such JW, Robillard KA. 1999. Silver toxicity to Chironomus tentanl
    in two freshwater sediments. Environ Toxicol Chem 18:30-39.
 7.  Rodgers JA  Jr,  Deaver E, Rodgers PL. 1997. Partitioning anc
    effects of silver in amended freshwater sediments. Ecotoxicoi
    Environ Saf 37:1-9.
 8.  Ankley GT,  Di  Toro DM, Hansen DJ, Berry  WJ, Mahony JD,
    1996. Technical basis and proposal for deriving sediment qua!
    ity criteria for metals. Environ Toxicol Chem 15:2056-2066!
 9.  Hansen DJ, Berry WJ, Mahony JD, Boothman WS, Robson DL1
    Ankley GT. 1996. Predicting the toxicity of metals-contaminateci
    field sediments using interstitial concentration of metal and ack
    volatile sulfide normalizations. Environ Toxicol Chem 15:2080-!
    2094.                 -                           '
10.  Scott KJ,  Redmond MS. 1989. The effects of a contaminated
    dredged material on laboratory populations of the tubiculous am-J
    phipod, Ampelisca  abdita. In Cowgill UM, Williams LR, eds'i
    Aquatic Toxicology and Hazard Assessment, Vol 12. STP  1027.1
    American Society for Testing and Materials, Philadelphia, PA, pp
    289-303.                                              I
11.  Di Ibro DM, Mahony JD, Hansen DJ, Scott KJ, Hicks MB, Mays!
    SM, Redmond MS. 1990. Toxicity of cadmium in sediments: Role'
    of acid volatile sulfide. Environ Toxicol Chem 9:1489-1504.
12.  Di Toro DM, Mahony JD, Hansen DJ, Scott  KJ, Carlson AR
    Ankley GT. 1992. Acid volatile sulfide predicts the acute toxicity
    of cadmium  and nickel  in sediments.  Environ Sci Technol 26
    96-101.                                               I
13.' Hampson BL. 1977. Relationship between total ammonia and free
    ammonia in  terrestrial and ocean water.' Cons  Inter Explor Me
    37:117-122.
14.  Hamilton MA, Russo RC, Thurston RV. 1977. Trimmed Spear
    man-Karber method for estimating median lethal concentration
    in toxicity bioassays. Environ Sci Technol 11:714-719. Correc
    tion 12:414 (1978).
15.  Mearns AJ, Swartz RC, Cummins JM, Dinnel PA, Plesha P, Chap
    man PM. 1986^ Inter-laboratory comparison of a sediment toxicit;
    test using the marine amphipod, Rhepoxynius abronius. Mar En
    vifon Res  19:13-37.
16.  U.S. Environmental Protection Agency.  1994. Statistical sum
    mary, EMAP-estuaries, Virginian Province—1991. EPA/620/R
    94/005. Technical Report. Washington, DC.
17.  Stephan CE, Mount DI, Hansen DJ, Gentile JH, Chapman GA|
    Brungs WA. 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, USA.
IS.  Lide DR, ed. 1995. CRC Handbook of Chemistry and Physics,
    76th ed. CRC, Boca Raton, FL, USA, pp 8-85.
19.  U.S. Environmental Protection Agency. 1994. Methods for as-
    sessing the toxicity of sediment-associated contaminants with esll
    tuarine and marine amphipods. EPA/600/R-94/025. Technical Rel
    port. Washington, DC.                                   |
20.  Sceintiftc Applications International Corporation. 1993. Data re-
    port on comparative sensitivity of four amphipod species to aque-
    ous ammonia. Report 1071. Narragansett, RI,  USA.          [
21.  Shaw JR,  Wood CM, Birge WJ, Hogstrand C. 1998. Toxicity o(f
    silver to the marine teieost (Oligocottus maculosus): Effects of
    salinity and  ammonia. Environ Toxicol Chem  17:594-600. •  |
22.  State of Florida. 1994. Approach to the Assessment ofSedimen

-------
48     Environ. Toxicol. Chem. 18, 1999
                                              W.J. Berry et al.
    Quality in Florida Coastal Waters, Vol  1. Florida Department
    of Environmental Regulation, Tallahassee, FL, USA.
23. Cubbage I, Bans D, Briedenbach S. 1997. Creation and analysis
    of freshwater sediment quality values in Washington State. Pub-
    lication 97-323a. Washington  State  Department of  Ecology,
    Olympia, WA, USA..
24. U.S. Environmental Protection Agency. ,1997. The incidence
  ' and severity of sediment contamination in the surface waters
    of the United States. EPA 823/R-97/006. Washington, DC.
25. Ankley GT. 1996. Evaluation of metal/acid-volatile sulfide rela-
    tionships in the prediction of metal bioaccumulation by benthic
    macroinvertebrates. Environ Toxicol Cheat 15:2138-2146.
26. Warren L, Tessier A, Hare L.  1998. Modeling  sediment accu-
    mulation by benthic invertebrates in situ:  Relative contributions
    of sediment and overlying water compartments to organism cad-
    mium concentrations. Umnol Oceanogr (in press).
27. Fisher N, Wang W-X. 1998. Trophic transfer of silver to marine
    herbivores: A review of the recent studies. Environ Toxicol Chem
    17:562-571.
28. DeWitt TH, Swartz RC, Hansen DJ, McGovem D, Berry WJ.
    1996. Interstitial metal and acid volatile sulfide predict the bio-
    availability of cadmium during a full life cycle sediment toxicity
    test using the estuarine amphipod, Leptocheirus plumulosus. En-
    viron Toxicol Chem 15:2093-2101.
29. Hansen DJ, Mahony JD, Berry WJ, Benyi S,  Corbin J, Pratt
    S, Able MB. 1996. Chronic effect of cadmium in sediments
    on colonization by benthic marine organisms: An evaluation
    of the role of interstitial cadmium and acid volatile sulfide in-
    biological availability. Environ Toxicol Chem  15:2126-2137.
30. Sibley PK, Ankley GT, Cotter AM, Leonard EN.  1996. Predicting
    chronic toxicity of sediments spiked with zinc: An evaluation of the
    acid volatile sulfide (AVS) model using a life-cycle test with the
    nudge Chironomus tenlans. Environ Toxicol Chem 15:2102-2112.
31. Hook S, Fisher N.  1997. Sublethal response of zooplankton to
    silver: the importance of exposure route. Abstracts, 18th Annual
    Meeting of the Society of Environmental Toxicology and Chem-
    istry. San Francisco, CA, November 16-20, p 51. '

-------

-------
                                      Environ. Set Technoi 1992, 20, 96-101
 Acid Volatile Sulfide Predicts the Acute Toxlcity of Cadmium and Nlckei In
 Sediments
 Dominic M. Dl toro,"T John D. Mahony.* David J. Hansen.* K. John Scott,1 Anthony R. Cartoon/ and
 Gerald T. Ankley-L

 Environmental Engineering and Science Program and Chemistry Department Manhattan College. Bronx, New York 10471,
 EPA Environmental Research Laboratory, Nanragansett. Rhode Island 02692. Science Applications International Corporation,
 Narragansett, Rhode Island 02592, and EPA Environmental Research Laboratory, Duluth, Minnesota 55804
• Laboratory toxicity tests using amphipods, oligochaetes,
and snails with spiked freshwater and marine sediments
and with contaminated sediments collected from an EPA
Superfund site demonstrate that no significant mortality
occurs relative to controls if the molar concentration of acid
volatile sulfide (AVS) in the sediment ia greater than the
molar concentration of simultaneously extracted cadmium
and/or nickel Although it is well-known that these metals
can form insoluble sulfides, it apparently has not been
realized that AVS is a reactive pool of solid-phase sulfide
that is available to bind metals and render that portion
unavailable and nontoxic to biota.  Thus, the AVS con-
centration of a sediment establishes the boundary below
which these metals cease to exhibit an acute toxicity in
freshwater and marine sediments.
Introduction

  Predicting the bioavailability and toxicity of metals in
aquatic sediments is a critical component in the develop-
ment of sediment quality criteria (1).  The use of total
sediment metal concentration Oonol/g dry weight) as a
measure of the bioavailability concentration is not sup*
ported by available data (2). Different sediments exhibit
different degrees of toxicity for the same total quantity
of a metal.  These differences have been reconciled by
relating organism response to the chemical concentration
in the interstitial water of the sediments (3-5). In addition,
a substantial number of experiments using water-only
exposures point to the fact that biological effects can be
correlated to the divalent metal activity (M*1"} (6,7). This
suggests that the bioavailability of metals in sediments is
related to the chemical activity, of the metal in the sedi-
ment-interstitial water system.  Hence, the sediment
properties which determine the metal activity in the sed-
iment-interstitial water system also determine the fraction
of the metal that is bioavailable and potentially toxic.
  Unless explicitly stated, when we refer to metals in this
paper we mean cadmium and/or nickel. For sediments
and metals tested to date, metal activity  in the sedi-
ment-interstitial water system, as measured by acute
toxicity to benthic organisms, is strongly influenced by the
sulfide and metal concentrations that are extracted from
the sediment using cold hydrochloric -acid.  This sulfide
fraction is conventionally referred to as the acid volatile
sulfide or AVS (8). The metal concentration that is si-
multaneously extracted we term the simultaneously ex-
tracted metal or SEM. The significance of performing the
sulfide and metal extraction under equivalent conditions
  'Environmental Engineering and Science Program, Manhattan '
College.
  1 Chemistry Department, Manhattan College.
  I EPA Environmental Research Laboratory, Narragansett, RI.
  1 Science Applications International Corp.
  1 EPA Environmental Research Laboratory, Duluth. MN.

96  Environ. 3d. Technol.. Vol. 26, No. 1, 1992
      is discussed below. For [SEM]/[AVS] < 1, no acute tox-
      icity (mortality >50%) has been observed in any sediment
      for any benthic test organism. For [SEM]/[AVS] > 1, the
      mortality of sensitive species (e.g., amphipods) increases
      in the range of 1.5-2.5 jimol of SEM/junol of AVS.
        This observation is important because acid volatile
      sulfide is found in most freshwater and marine sediments.
      It is found in sediments with sandy and gravelly textures
      that do not resemble the anoxic sulfidic sediments that are
      more commonly associated with the presence of sulfide.
      Concentrations range  from <0.1 to >50 pinol of AVS/g
      (9-13).

      Experimental Procedures
        The results of four separate experiments are presented
      in this paper. The detailed experimental procedures are
      described elsewhere (14-17). The sediment toxicity tests
      generally followed ASTM recommendations (IS). Flowing
      water (~10 volume replacements/day) and aeration en-
      sured acceptable dissolved oxygen concentration. The tests
      were conducted using 200-900-mL exposure vessels with
      200 mL of sediment (3.6-cm depth)  and 600 mL of over-
      lying water.  The animals were exposed for 10 days to
     ~ control and test sediments.
        Sediments were spiked by adding 1.0 L of wet sediment
      to 2.0 L of dilution water into which a weighted amount
      of cadmium or nickel chloride had been dissolved. The
      tests were initiated by adding the sediments to the expo-
      sure containers, waiting from 1 to 6 days, and adding the
      animals.  After termination, the contents of each exposure
      container were sieved and counted.  Missing individuals
      were counted as mortalities. Parallel exposures were used
      for chemical measurements.
        The acid volatile sulfide (AVS) concentration is the
      solid-phase sulfide that is soluble in room-temperature 05
      M HC1 in 1 h. The measurement technique is to convert
      the sulfides to HgSlaq), purge it with oxygen-free nitrogen
      gas (four bubbles/s), and trap it in a gas-tight assembly
      (14, IS). The reaction vessel is followed by a pH 4 chloride
      trap (0.05 M potassium hydrogen  phthalate) and two
      sulfide traps (0.1 M silver nitrate) for trapping HjS as AgjS
      precipitate. For 10-15 g of wet sediment the detection
      limit is ~0.5 pmol/g. The simultaneously extracted metal
      (SEM) was measured in a filtered aliquot by conventional
      atomic absorption methods.  The AVS and SEM were
      measured at the start of the experiment when animals were
      added and at the termination in the parallel exposure
       The initial experiment (14) exposed two marine am-
     phipods (Ampelisca abdita and Rhepoxynius hudsoni) to
     three uncontaminated marine sediments: a fine-grained
     sediment with a relatively high AVS (sediment (15 ftmbr
     of AVS/g) from central Long Island Sound, NY; a sandy
     sediment with a relatively low AVS (iJ-jwnol of AVS/g)
     from a salt water pond in Ninigret, RI;jSpfon equivolume
     mixture of the two sediments (4.3 n&qpf AVS/g). For

0013-936X/92/0926-0096M3.00/0   © 1991 American Chemical Society

-------
 this experiment, the total acid extractable cadmium was
 measured separately.  As shown below, this is equivalent
 to the SEM measurements for cadmium. For the re-
 maining experiments, the  SEM concentrations  were
 measured directly.
   The second experiment (IS) simultaneously exposed two
 freshwater organisms, a snail, Helisoma sp., and an oli-
 gochaete, Lumbriculus variegatus, to cadmium added to
 three uncontaminated freshwater sediments from Pe-
 quaywan Lake, MN (42-ftmol of AVS/g), East River, WI
 (8.8/imol of AVS/g), and West Bearskin Lake, MN (3.6
 jonol of AVS/g). The third experiment (16) exposed  A.
 abdita to nickel added to central Long Island Sound and
 Ninigret Pond sediments.
   The final experiment (17) exposed the freshwater am-
 phipod, Hyalella ozteca, to 17 sediment samples taken
 from Foundry Cove, a small (213  ha) predominantly
 freshwater cove in the upper reach of the tidal portion of
 the Hudson River, NY. These sediments were contami-
 nated with cadmium and nickel from a battery manufac-
 turing facility (20,21). The sediments spanned the range
 from fine-grained sediments, highly enriched in organic
 carbon, to gravelly composites with low organic carbon
 concentrations. The cadmium and nickel concentrations
 are approximately  equimolar throughout the range of
 sediment concentrations present, 0.3-1000 /imol of SEM/g.
 AVS ranged from 0.1 to 47 ,anol/g. and [SEM]/[AVS]
 ranged from 0.1 to >100 with several in the critical range
 of.
            Cd2* + FeSfe) — CdS(s) + Fe2+
(1)
     r  WK»»«'™'
  using an analysis of the M(II)-Fe(II)-S(II) system with
  both MS(s) and FeS(s) present  MOD represents any
  metal that forms a sulfide that is more insoluble than FeS.
  If the added metal, [M] A, is less than the AVS present in
  the sediment then, as shown in the Appendix, the ratio of
  metal activity to total metal in the sediment-interstitial
Table I. Metel Sulfide Solubility Products*
' metal
aulfide
FeS
NiS
ZnS
CdS
PbS
CuS
HgS
log
Kfj
-3.64
-9.23
-9.64
-14.10
-14.67
-22.19 .
-38.50
•log
Kv
-22.39
-27.98
-28.39
-32.85
-33.42
-40.94
-57.25
log
(Km/KTa)

-5.59
-6.00
-10.46
-11.03
-18.55
-34.86
  'Solubility products, K^. for the reaction M!+ + HS~ ~ MS(»)
+ H+ for CdS (greenockite), FeS (mackinawite), and NiS (miller-
ite) from ref 23. Solubility products for CuS (covellite), HgS (me-
tacinnabar), PbS (galena), and ZnS (wurtzite), and pK, « 18.57 for
the reaction HS" —H* + S*- from ref 26.  Kv for the reaction M*+
+ S3' •"• MS(g) is computed from log Kvf and pK,,	

water system is less then the ratio of the MS to FeS sol-
ubility products:                      .


This is a general result that is independent of the details
of the interstitial water chemistry.  In particular it is in-
dependent of the Fe** activity. Of course the actual value
of the ratio {M^I/IMfo depends on aqueous speciation, as
indicated by eq 1. However, the ratio is still less than the
ratio of the sulfide solubility products.
  The sulfide solubility products and the ratios are listed
in Table I.  The ratio of cadmium activity .to total cad-
mium is less than 10~las. For nickel the ratio is less than
10**. By inference this reduction  in metal activity will
occur for any other metal that forms a sulfide that is sig-
nificantly more insoluble than iron monosulfide. The
ratios for the other metals in Table I, Zn, Pb, Cu, and Hg,
indicate that metal activity for these metals win be very
small in the presence of excess AVS.
• If, on the other hand, reactive metal is present in excess
of the AVS then, as shown  in the  Appendix, the metal
activity is


If no other strong complexing ligand is present,  1
the FeS  is converted to a metal sulfide via the displace-
ment reaction, eq 1. For cadmium, the sulfide extracted
is essentially constant throughout the range of cadmium
 additions, indicating that CdS is completely soluble in the
 AVS extraction. By contrast, the decrease in extracted
 sulfide for [Ni]/ [AVS] > 1 indicates that nickel sulfide is
 not completely soluble in the extraction.
   If a more efficient procedure were used to increase the
 fraction of metal extracted which did not also capture the

               Environ. Sd. Techno).. Vol. 28. No. 1, 1992 97

-------
                                 Cadmium
                  Nickel
                  100.00
(6./
»-
otUTl
                    1.00
                          East River
                              F«S
                                        4 T- 1C


                                           CdS
                                      1
                                                        100.00
                                                         10.00
                      O.O1   0.1O   1.00   10.00  1OO.OO
                           Cd/AVS (iimol/umol)
                                                          1.00
         LI Sound     • T - o d
                       * T - 10d
                                                                  F«S
                                                                               HIS
          Ni/AVS (umbl/itmpt)
Figure 1. Add votatte suffide extracted versus the cadmium and nickel AVS ratios. Data from the start (t = 0 day) and em) (I = 10 day) of
the exposure period.  The dotted few defcrtts the regions where the AVS b eflher predominantly FeS or either CdS or MS.
additional sulfide extracted, then the sulfide associated
with the additional metal release would not be quantified.
This would result in an erroneously large metal to AVS
ratio. Hence, we use the AVS as the measure of reactive
sulfide in the system and the simultaneously extracted
metal, SEM, as the measure of reactive metal Since these
quantities are operationally defined, it remains to dem-
onstrate their utility experimentally.

Results
  The four experiments described above have been per-
formed to test the utility of SEM to AVS ratios for pre-
dicting bioavailability and toxicity of metals to benthic
organisms. Three were performed with spiked sediments;
the fourth employed contaminated sediments from an
EPA Superfund site. The results are presented in Figure
2. The data for each experiment in which only the sed-
iment type is varied are superimposed with the exception
of the first experiment, where two amphipods, A. abdita
and R. hudsoni, with similar sensitivity (14) were used.
The lines connect the results for the same sediment The
lack of a unique relationship between extracted metal
concentration on a dry weight basis and organism mortality
is reflected by the different mortality-concentration re-
lationship for each sediment in the spiked metal experi-
ments.
  The scatter in observed mortality for the Foundry Cove
sediments is striking.  Metal concentrations from 0.1 to
28 jimol of SfiM/g  were not toxic in some sediments,
whereas 0.2-1000 pmol of SEM/g were lethal in other
sediments. These results reaffirm the observation that the
bioavailable fraction of metals in sediments varies from
sediment to sediment, and in this case, it varies dramat-
ically.                        .   •
  By contrast, a clearly discernable mortality-concentra-
tion relationship is observed in Figure 3, where mortality
is related to the SEM to AVS molar ratio. The chemical
theory predicts that mortality should begin to occur at
[SEM]/[AVS] = 1 (eqs 2 and 3). No mortality in excess
of 50% is observed for sediments with [SEM]/[AVS] <
1. For sediments.with [SEM]/[AVS] > 1-3, the mortality"
increases dramatically.  For sediments with [SEM]/[AVS]
> 10,80-100% of individuals from all the tested species
died.

Conclusions
  These data suggest the following conclusions. If the ratio
tSEM]/IAVS} =  1  is used to discriminate toxic from
nontoxic sediments (greater or less than 60% mortality,
respectively), then for the 117 experiments performed,
51%  are correctly classified  as nontoxic (bottom ieft
quadrant in Figure 3) and 42% are correctly classified as
toxic (top right quadrant). That 7% that are misclassified
as toxic (bottom right quadrant) follows from the as-
sumption that metal activity will invariably be high enough
to cause toxicity if [SEM]/[AVS] > 1, eq 3. It is possible
that other Uganda, associated with sediment abrption, for
example, are reducing the metal activity below that which
is lethal to the text organisms. .Also, less sensitive organ-
isms can tolerate the increased metal activity even if
(SEM]/[AVS] > 1. For organisms (hat are present when
[SEM]/[AVS] > 1, preliminary data suggest that the ex-
tent to which metals bioaccumulate is strongly influenced
by the AVS concentration (17,28).
  If a more restrictive interpretation is adopted and the
criterion [SEM]/[AVS] < 1 is used only to predict when
a sediment is not acutely toxic, then all experiments are
correctly classified. We assume that the toxic Foundry
Cove sediment with [SEM]/[AVS]  - 0.98  is indistin-
guishable from a unity ratio. Hence our data indicate that
sediments with [SEM]/[AVS] < 1, perhaps [SEM]/[AVS]
< 0.9 as a safety factor, do not cause greater than 50%
mortality in all the sediment toxicity tests performed to
date. This is directly attributable to the excess AVS in
the sediment, which assures that the metal activity in the
. sediment-interstitial water system is below the lethal metal
activity for the-organisms tested.
  It is possible that these results are due to a covariation
of AVS with other sediment properties, for example, or-
ganic carbon or iron content, which are actually controlling
the metal  bioavailability.  However, the fact that the
boundary occurs at [SEM]/[AVS] = 1, as predicted by eq
2, which is based on the supposition that  AVS is the
controlling sediment property, strongly argues that the
relationship is casual, rather than correlative.
.  It should be noted that if the AVS concentration is
effectively zero, as it would be in fully aerobic sediments,
then other sediment properties would control the metal
activity. This does not contradict the assertion that the
 [SEM]/{AVS] molar ratio of less than 1 predicts an ab-
sence of toxicity since in this case the molar ratio would
be very large.  The prediction is not much help but it is
still correct. However, even a small AVS concentration,
[AVS] ~ 0.1 pmol/g, can sequester a significant quantity
of metal and should be taken into account in determining
the potential for metal toxicity for these sediments.
98  Environ. Sci. Techno!., Vol. 26, No. 1. 1992

-------
                             AMPHIPOD - Cd - SW
                                               AMPHIPOD - Ni - SW
                     100

                      8O

                      60

                      40

                      20

                       0
                       0.1
                              1.O
                                    10.0   10O.O 1000.0
                                                                    1.0
                                                                          10.0   100.0  1000.0
                               SNAIL - Cd - FW
                DC
                O
                5

                3
100

 8O

 BO

 4O

 20

  0
                           OrO"
                                                                  OLIGOCHAETE - Cd - FW
100

 80

 60

 40

 20

  O
                                                                 D-«D
                       O.1
                              1.O
                                    10.0   tOO.O 10OO.O
                                                             0.1
                                                                    t.O
                                                                          1O.O  1OO.O  1000.0
                               FOUNDRY COVE
                            AMPHIPOD - Cd+Ni - FW
100
80
«O
40
20
0
*k it **** •*
* * .
*
*
*
• * *
                                                  LEGEND
                                            —SNAIL, OUGOCHAETE—
                                            • • KQUAYWAN UK - Cd
                                            • • EAST RIVER - Cd
                                            O O WEST BEARSKIN LK - Cd


                                            if ' FOUNDRY COVE - Cd • M
                                            •  U SOUND - Cd
                                            4>  MIXTURE-Cd
                                            O  WNMRET Mt - C4
                                            d>  U SOUND - M
                                            O  NWHQHET PP - NI
                       O.1
                              1.O
                                     10.0   10O.O  1000.0
                                       SEDIMENT (p.mol SEM/gm dry wt)

Flgura2. Organism mortaBty versus metol concantratkxi. Aa*nab and metals M Intficated for seawater (SW) and freshwater (FW) axposires.
The Ines connect resute from the same sedknenL Rytte Foundry (frveaadhittil^ihe metal corxsermall^
extracted cadmium and nickel.
ORQANI

•o

BO
4O
30
O
I vfffm!L^y
' - ''* „**'
1 •
1 *
• T**~
. * s 1 *
w ^^Dw^frfS Via *



*
•
. '
•
       O.O1     .  O.1O       t.OO       1O.OO

               SEDIMENT (iimol SEM/itmol AVS)
                                              tOOAO
Figure 9. Organism mortally versus V» mrtar rate of SEM to AVS
of the segment The seoTment motel and AVS are the averages of
tt» initial and final measured concentrations. For the Foundry Cove
sediments, the metal concentration b the motor sum of the simulta-
neously extracted cadmium and nickel. Symbols defined ki Figure 2.

  Additionally, it is important to realize that "anaerobic"
and  "aerobic" sediments are not precise classifications.
Sediments are characterized by an aerobic layer underlaid
by an anaerobic layer. The sediments employed in these
experiments did not inhibit the survival of the obligate
aerobic organisms used in the control exposures—in par-
ticular the tube-bunding amphipods. Most benthic aerobic
organisms survive in sediments that are underlaid with
                                    completely anaerobic sediments, which are characterized
                                    by significant AVS concentrations.  Our experiments
                                    suggest that the presence of AVS in the anaerobic layer
                                    is sufficient to reduce the metal activity to which the an-
                                    imals are exposed.
                                      However, we have not examined the extent to which an
                                    aerobic layer depth is sufficient to mitigate the influence
                                    of the lower layer AVS. Thus it is possible to imagine a
                                    situation where AVS at depth (>10 cm) might not control
                                    metal activity in a completely aerobic overlying layer—the
                                    top 10 cm for example—where the animals are exposed.
                                    It seems likely that even in this situation the presence of
                                    a sink of metals at depth would reduce the activity in the
                                    entire sediment to below toxic levels.  The reasoning is that
                                    the diffusions! transport of metal in the interstitial water
                                    would bring metals from a presumably higher concentra-
                                    tion in the aerobic layer interstitial water to the lower
                                    concentration in the anaerobic layer. This would even-
                                    tually deplete the aerobic layer of metals and establish a
                                    uniformly low metal activity in the  interstitial water-
                                    sediment system. Thus, even in this case, we would expect
                                    that excess AVS would predict the absence of acute tox-
                                    icity.  However, this is yet to be demonstrated experi-
                                    mentally.
                                      We believe that the data presented in this paper dem-
                                    onstrate that, for the first time, it is possible to predict
                                                                      Environ. Sd. Techno!., Vol. 26. No. 1.1992  M

-------
 when a sediment will not be acutely toxic due to cadmium
 and/or nickel contamination and, by implication, to all
 toxic metals that form sulfides that are significantly less
 soluble than FeS.  The criterion is that the molar sum of
 simultaneously extracted Cd, Cu, Hg, Ni, Pb, and Zn is
 less than the molar acid volatile sulfide concentration:
 [SEMJcj -f [SEM]C(I + [SEM]H, + [SEM]Ni +
                  [SEMJPb + [SEMJjfc/tAVS] < 1  (4)

  It should be noted that hi order to apply this relation-
 ship it is necessary to measure all the toxic simultaneously
 extracted metals that are present in amounts that con-
 tribute significantly to the molar SEM sum, typically Cd,
 Cu, Ni, Pb, and Zn. Failing to do this could lead to'an
 incorrect prediction of the lack of acute toxicity, i.e.t the
 sum of the measured SEMs to AVS ratio is less than 1,
 when in fact the unmeasured metal could increase the
 SEM concentration so that the ratio exceeds Land toxicity
 would be possible.  Thus, the acute toxicities of these
 metals are interrelated, with each contributing to the AVS
 that is bound by toxic metal. If excess AVS remains, no
 acute toxicity is expected. If excess metal remains, then
 the most soluble, Ni and Zn, will appear as free metal and
 toxicity is possible.

 Acknowledgments

  The assistance  and encouragement of the following
 people is gratefully acknowledged:  Christopher Zarba,
 Criteria and Standards Division, U.S. EPA; Richard
 Swartz, EPA Research Laboratory, Newport, OR; Mark
 Springer and Mark Huston, Region n, US. EPA; Herbert
 Allen, University of Delaware; Robert Thomann, Man-
 hattan College; and, most of all, our research assistants at
 Manhattan College, M. Hicks, S. Mayr, I. Sweeney, P.
 Morgan, C. Sydlik, L. Milevoj, and C. Begley; at the EPA
 Narragansett Laboratory, W. Berry, M. Redmond,  D.
 Robson, and K. McKenna (SAIC); and at the EPA Duluth
 Laboratory, G. Phipps, V. Mattson, E. Leonard, P. Kosian
 (AScI), and A. Cotter (AScI).
     \             •
 Appendix
  Solubility Relationships for Metal Sulttdes.  Con-
sider the following situation: a quantity of FeS is titrated
 with a metal that forms a more insoluble sulfide. We
analyze the result using an equilibrium model of the M-
(IQ-Fe(II)-S(II) system.  The mass action laws for the
metal and iron sulfides are

                                               (5)

                                               (6)

where {M2*], [Fe2+J, and [S2"] are the molar concentrations;
TM**> YP«»> and 7s*- are the activity coefficients; and KMS
and KftS are the sulfide solubility products.  The mass
balance equations for total M(II), Fe(II), and S(II) are

                                               (7)

                                               (8)

                                               (9)
where
                         (MS(9)] - [M]A

                       [FeS(s)J = [FeS(s)]j

                 IMS(s)] + [FeS(s)J = [FeS(s)]i


                   • lM2+]/[£MFtS + gT^y8^S) IMJ* <15)
                                                                                             /
                                                       which results from ignoring the leading term in eq 14. This
                                                       is legitimate because the term in parentheses in eq 14 is
                                                       small relative to [M]A due to the presence of the sulfide
                                                       solubility products. As a result, [S2'] is also small since
                                                       it is in the denominator. Hence, the leading term in eq
                                                       14 must be small relative to [M] A and can safely be ignored.
                                                         The metal activity can now be found from the i  '
                                                       equilibrium eq 5:
                                                       so that




                                                       where



                                                       and
                                                          •-.


                                                        , Equation 17 can be expressed as
                                                                 (M2*
                   Mas I
                                                                         tFeSl
                                                                                        K
(17)



(18)



(19)




(20)
                                                                                          MS
                                                                                        K,
                                                                                          p«s
The magnitude of the term in parentheses can be esti-
mated as follows.  The first term in the denominator is
always greater than or equal to 1,0Fe»* > 1, because it is
the reciprocal of two terms both of which are less than or
equal to 1, eq 18. They are afj* < 1, which is the ratio
of the divalent to total aqueous concentration, and »,»
< 1, which is an activity coefficient. The second term in
the denominator cannot be negative, ftu^Kus/Kf^ > 0,
since all of its terms are positive. Thus, the denominator
of the expression in parentheses is always greater than 1,
/W« + PHP+KuafKfjs >  1.  Therefore, the expression in
parentheses is always less than 1. Hence, the magnitude

-------
 of the ratio of metal activity to total added metal is
 bounded from above by the ratio of the sulfide solubility
 products:



 This result applies if [FeSji > [M]A so that excess [FeS(s)J
• is present.
   If sufficient metal is added to exhaust the initial quan-
 tity of iron sulfide, then JFeS(s)] =  0.  Hence, the iron
 sulfide mass action equation (6) is invalid and the above
 equation no longer applies. Instead, the only solid-phase
 sulfide is metal sulfide and
                    [MS] = [FeSL
                                                  (22)
 so that, from the metal mass balance equation

           |M*+} = TM"«M»<[M]A - [FeSWD       (23)

 This completes the derivation of eqs 2 and 3.
Glossary
[AVS]
{Fewj
IFe2*]
(FeS(a)]
[FeS(s)]|
[M2*]
[M]A
[MS(s)J

IS2*}
[S2-]
[SEM]
JSEMJcu
[SEM]ifi

[SEM]pb
            acid volatile sulfide concentration Gtmol/g)
            activity of Fe2+ (mol/L)
            concentration of Fe2+ (mol/L)
            concentration of iron sulfide (mol/L)
            initial iron sulfide concentration in the sedi-
               ment (mol/L)
            solubility product for FeS(s) [(mol/L)2]
            solubility product for MS(s) [(mol/L)2]
            divalent metal activity (mol/L)
            concentration of M2* (mol/L)
            concentration of added metal (mol/L)
            concentration of solid-phase metal sulfide
               (mol/L)
            activity of S2- (mol/L)
            concentration of S2- (mol/L)
            simultaneously extracted metal concentration
               fcmol/g)
            simultaneously extracted Cd concentration
               Ounol/g)                '
            simultaneously extracted Cu concentration
               Gunol/g)
            simultaneously extracted Hg concentration
               (pmol/g)
            simultaneously extracted Ni concentration
               (pmol/g)   •
            simultaneously extracted Pb concentration
               Gunol/g)
            simultaneously extracted Zn concentration
               Gimol/g)
A***-
[ZFe(aq)]

[£M(aq)J

[£S(aq)J
             activity coefficient of Fe"
             activity coefficient of M2*
            . activity coefficient of S*"
             concentration  of total dissolved  Fe(II)
               (mol/L)
             concentration of total dissolved M(n) (mol/
               L)
             concentration of total dissolved SOD (mol/L)
   Registry No. Cd, 7440-43-9; Ni, 744042-0; S2-, 18496-25-8?
 ZnS, 1814-98-% PbS, 1314-87-0; CuS, 131740-4; HgS, 1344-48-5.
 literature Cited
  (1)  DiToro, D. M.; Zarba, C. S.; Hansen, D. J.; Swarte, R. C.;
      Cowan, C. E.; Pavlou, S. P.; Allen, H. E.; Thomas, N. A.;
      Paquin, P. R.; Berry, W. Environ. Taxieol. Chem., in press.
  (2)  Luoma, S. N. Set. Total'Environ. 1983,28,1.
  (3)  Adapts, W. J.; Kimerle, R. A.; Mosher, R. G. ID Aquatic
      Toxicology and Hazard Assessment: Seventh Symposium;
      Cardwell, R. D., Puidy, R, Bahner, R. C., Eds.; American
      Society for Testing and Materials, Philadelphia,'PA, 1985;
      p429.
  (4)  Swartz, R. C.; Ditsworth, G. R; Schults, D. W4 Lamberaon,
      J. 0. Afar. Environ. Res. 1985,18, 133.
  (5)  Kemp. P. P.; Swartz, R C. Afar. Environ. Ret. 1988,26,135.
  (6)  Sunda, W.; Guillard, R R L. J. Mar. Res. 1976,34, 511.
  (7)  Borgmann, U. In Aquatic Toxicology, Nriagu, J. 0., Ed.;
      J. Wiley: New York, 1983; p 47.
  (8)  Berner, R. A. Am. J. Sci. 1967,265,773.
  (9)  Alter, R C. In Estuarine Physics and Chemistry. Studies
      in Long Island Sound; Saltzman, B., Ed.; Academic Press;
      New York, 1980; p 238.
 (10)  Reaves, C. PhJX Thesis, Yale University, New Haven, CT,
      1984.
 (11)  Nriagu, J. O. LimnoL Oceanogr. 1963,13, 430.
 (12)  Nriagu, J. (X; Colter, R D. LimnoL Oceanogr. 1976,21,485.
 (13)  Matisoff, G.; Fisher. J. B.; McCaU, P. L. Geochim. Cot-
      mochim. Ada 1981, 45,2333.
 (14)  DiToro,D.M.;Mahony,J.D.;Hansen,D.J.;Scott,K.J.;
      Hicks, M. B.; Mayr, S. M.; Redmond, M. S. Environ.
      Toxicol. Chem. 1990,9,1487.
 (15)  Carlson, A. R; Phipps, G. L.; Mattson, V. R; Kosian, P.;
      Cotter, A. ERL—Duluth Report No. 2471, EPA Environ-
      mental Research Laboratory, Duluth, MN, 1990.
 (16)  Hausen, D.J.; Scott, K.3.ERL—Narragansett Report EPA
      Environmental Research Laboratory, Narragansett, RI,
      1990.
 (17)  AttkkV.G.T4PhiPPB,G.L;Koeiflii,P^ Cotter, A^Mattaon,
      V. R; Mahony. J. D. Environ. Toxicol. Chem., submitted.'
 (18)  American Society for Testmg and Materials. Proposed New
      Standard Guide for Conducting Solid Phase 10-Day Static
      Sediment Toiicity Tests with Marine and Estuarine Am-
      phipods. Draft No. 5,1989, American Society for Testing
      and Materials, Philadelphia, PA.
 (19)  Morse, J. W.; Millero, F. J,; Cornwell, J. C.; Rickard, D.
      Earth Sci. Reu. 1987,24,1.
 (20)  Hazen,RE4Knap,T.J.uiCWmiumin{fce£nuironme7»t,
      Parti: Ecological cycling; Nriagu, J.O., Ed; J.Wiley: New
      York, 1980; p 400.
 (21)  Knutson, A. B.; Klerks, P. L.; Levinton, J. S. Environ.
      Poliut. 1987,45, 291.
 (22)  Boutegue, J. In Trace Metals in Sea Water. Wong, C. S.,
  .    Boyle, R,  Bruland, K. W., Burton, J. D., Eds^ Plenum
      Press:. New York, 1983; p 663.
 (23)  Emerson, S.; Jacobs, L; Tebo, B. In Trace Metal* in Sea
  ••   Water, Wong, C. S., Boyle, E., Bruland, K. W^ Burton, J.
      D., Eds^ Plenum Press: New York, 1983; p 579.
 (24)  Cornwell, J. C.; Morse, J. W. Mar. Chem. 1987,22,193.
 (25)  Landere,D.R; David, M.B^ Mitchell, M-J./nt-J-Emriron.
      Anal. Chem. 1983,14,245.
 (26)  Schoonen,M. A.A.;Bame8. H.L.G«ocWm.Cosm«Jum.
      Acts 1988,52, 649.
 (27)  Byrne, R R; Kump, L. R; Cantrell, K. J. Afar. Chem. 1988,
      25,163.
 (28)  Striplin,B.D.Shagway Harbor Field Investigation. Tetra
      Tech, Inc., Bellevue WA.  1990.

 Received for review December 4, 1990.  Revised manuscript
 received March 12,1991. Accepted July 19,1991. This research
 wot supported by an EPA Cooperative Agreement between
 Manhattan College and EPA Environmental Research Labora-
 tory, Narragansett, RI. The Manhattan College participation
 in the Foundry Cove investigation was supported by the National
.Institutes of Environmental Health Sciences, Superfund Haz-
 ardous Substances Basic Research Program, Environmental
 Medicine, New York University Medical Center.
                                                                        Environ. Sd. Techno!., Vol. 26, No. 1, 1992   101

-------
APPENDIX B

-------

-------
     Research Papers on the Unavailability and Toxicity of Metals in Surface Waters2

Key Papers .-..'*

Allen, H. E. and D. J. Hansen. 1996. The importance of trace metal speciation to water quality
criteria. Water Environment Research. 68(l):42-54.

Campbell, P.G.C., 1995.  "Interactions Between Trace Metals and Aquatic Organisms: A Critique
of the Free-ion Activity Model," Metal Speciation andBioavailability in Aquatic Systems, A. Tessier
and D.R. Turner, eds., IUPAC, John Wiley and Sons.

Janes, N. and R.C., Playle. 1995. Modeling silver binding to gills of rainbow trout (Oncorhynchus
mykiss). Environ.  Toxicol Chem.  14:1847-1858.

Meyer, J.S,, R.C. Santore, J.P. Bobbitt, L.D. DeBrey, CJ. Boese, P.R. Paquin, H.E. Allen, H.L.
Bergman and D.M. DiToro. 1999.  Binding of nickel and copper to fish gills predicts toxicity when
water hardness varies, but free-ion activity does not. Environ. Sci. Technol. (in Press).
                                  i                 •        -     •
Meyer, J.S. 1999. A mechanistic explanation for the ln(LC50) vs ln(Hardness) adjustment equation
for metals.  Envripn. Sci,  Technol.  (in Press).

Pagenkopf, G.K.  1983.  Gill surface interaction model for trace-metal toxicity to fishes: Role of
complexation, pH, and water hardness. Envrion.Sci. Technol. 17:342-347.

Playle, R.C., D.G. Dixon and K. Burnison. 1993a.  Copper and cadmium binding to fish gills:
modification by dissolved organic carbon and synthetic ligands. Can. J. Fish. Aquat. Sci. 50:2667-
2677. •'.'.-..

Piayle, R.C., D.G. Dixon and K. Burnison. 1993b.  Copper and cadmium binding to fish gills:
estimates of metal-gill stability constants and modeling of metal accumulation. Can. J. Fish. Aquat.
Sci. 50:2678-2687.

Tippingj E. 1994.  WHAM—A chemical  equilibrium model  and  computer code for waters,
sediments, and soils incorporating a discrete site/electrostratic model of ion-binding by humic
Substances.  Computers and Geosciences. 20:973-1023.

      f             .
Other Relevant Papers

Allison, J.D.,  D.S.  Brown, K.J.  Novo-Gradac.  March 1991.  MINTEQA2/PRODEFA2, A
Geochemical Assessment  Model  for Environmental Systems:  Version 3.0, Users  Manual.
EPA/600/3-91/021, USEPA ERL ORD, Athens, GA.
       2Articles listed as "Key Papers" are reprinted in this volume. Additional relevant papers
have been listed, but not reprinted.

-------
Bury, N.R., F.  Galvez and C.M. Wood.  1998. Effects of chloride, calcium and dissolved organic
carbon on silver toxicity: Comparison between rainbow trout and fathead minnows.  Environ
Toxicol Chem. 18:56-62.
Felmy, A.R., D.C. Girvin and E.A. Jenne, 1984. MINTEQ: A Computer Program for Calculating
Aqueous Geochemical Equilibria. USEPA Environmental Research Laboratory, Office of Research
and Development, Athens, Georgia.

Gorsuch, J. ans S. Klaine, Editors. 1999. Annual Review Issue: Silver Toxicity. Environ. Toxicol.
Chem. .18:1-108.

MacRae, R.K., D.E. Smith, N. Swoboda-Colberg, J.S. Meyer and H.L. Bergman. 1999.  Copper
binding affinity of rainbow trout (Oncorhynchus mykiss) and brook trout (Salvelinusfontinalis) gills.
Environ. Toxicol. Chem. (in Press).

Playle, R.C., R.W. Gensemer and D.G. Dixon, 1992.  Copper accumulation on gills of fothead
minnows:  Influence of water hardness, complexation and pH on the gill micro-environment.
Environ. Toxicol. Chem. 11:381-391.

Playle, R.C. 1999.  Physiological and toxicological effects of metals at gills of freshwater fish.
Environ. Toxicol. Chem. (in Press).                                         -

Santore, R.C. and C.t. Driscoll. 1995.  The CHESS Model for calculating chemical equilibria in
soils and solutions. Chemical Equilibrium and Reaction Models, SSSA Special Publication 42, The
Soil Society of America, American Society of Agronomy.

Schecher, W.D. and D.C. McAvoy. 1992. MINEQL+: A software environment for chemical
equilibrium modeling.  Computer Environ. Urban Systems. 16:65-76.

Wood, C.M., R.C. Playle and C. Hogstrand. 1999. Physiology and modeling of the mechanisms
of silver uptake and toxicity in fish. Environ. Toxicol. Chem. 16:71-83.

-------
     The  importance  of  trace  metal  speciation
                             to  water quality  criteria
                                        Herbert E. Allen, David J. Hansen
ABSTRACT:  Became the bioavailability of a trace metal, and conse-
quently its toxicity, is dependent on the physical and chemical form of
the metal, we have presented a detailed assessment of how speciation of
copper would be expected to affect its toxicity. Principles of chemical
speciation ate applied to demonstrate that inorganic fonns will be in
constant proportion to each other and to free copper ion during the
course of the titration of a sample of natural water with copper or in the
various treatments in a toxicity test conducted at constant pH and al-
kalinity. Binding of copper to dissolved organic matter or to suspended
paniculate matter may render the copper nonbioavailable. We have
considered a simple complexation model to describe the complexation
of copper to soluble ligands. Naturally occurring dissolved organic matter
is present at concentrations only slightly greater than that of copper.
Consequently, titration of water with copper results in a nonlinear re-
lationship between the concentration of copper present as free copper
ion plus inorganic copper species. The effects of stability constant of the
complex, concentration of tigand, and the total copper concentration
are evaluated. We have related bioavailable copper to the concentration
of free copper ion plus inorganic copper complexes, which is valid if the
pH and alkalinity of the waters used to develop a criteria are not different
On the basis of limited field data for the complexation of copper in
Narngansett Bay water, we do not expect that, significant differences in
water quality criteria (WQQ would result if the criteria were to be based
on free copper ion plus inorganic copper complexes rather than total
copper concentrations. We examined the effect of speciation of copper
in different waters as related to empirical or theoretically calculated water
effect ratios (WER). We show that, on the basis of sound chemical prin-
ciples, it. would be expected that the. most sensitive organisms would
have the greatest WER. This prediction is confirmed by the empirical
observations available. For insensitive organisms, knowledge of the con-
centration of Ugand is sufficient to reasonably predict the WER. However.
for the more sensitive organisms that give higher WERs, it is necessary
to measure or calculate the speciation of copper to predict the WER.
Use of predicted WERs may replace use of empirically derived WERs
as is now part of regulatory guidance for derivation of site-specific WQQ
if correspondence has been demonstrated. Water Environ. Res., 68,42
(1996).
KEYWORDS:   copper, criteria, metals, model, speciation, water quality.

   Bioavailability of a trace "metal to aquatic organisms, and the
toxicity of the metal is dependent on the physical and chemical
forms of the metal (Luoma, 1983; O'Donnel eta/., 1985). Benson
et al. (1994) have stressed the importance of understanding the
speciation of a metal in an aquatic system to the prediction and
interpretation of toxicity. Important factors to be considered in
the speciation of a metal include oxidation state, precipitation
and sorption, complexation, and the formation of organometallic
compounds. Kelly (1988) has written ".  . . the speciation of a
metal, rather than its total concentration, is the key to under-
standing its effect on the biota." Allen (1993a,b) has reviewed
the  principles of speciation with particular  emphasis toward
complexation and has shown  how speciation may  be used in
water quality criteria development  '                 .
  A predominance of studies have demonstrated that organism
response is correlated with free metal ion. The effect of speciation
on the bioavailability of copper has been studied most extensively
(Hodson et ai,  1979). Studies of copper toxicity to fish have
demonstrated that toxicity is not related to the total copper con-
centration, but rather to the concentration of the free copper
ion (Pagenkopf et al.. 1974). The observed toxicity of copper to
rainbow trout was strongly correlated to the concentration of
free copper ton (Brown et al., 1974). Howarth and Sprague(1978)
and Chakoumakos et al. (1979) concluded that the toxicity
of copper was dependent on the concentration of the free cop-
per ion.             •.                            .
  Although it might be tempting to conclude that the free metal
ion alone is responsible for toxicities observed, this appears to
be not supportable. Although organically bound copper appears
to be nontoxic, there is some debate over the toxicity of inorganic
hydroxy and carbonate  complexes. The most toxic inorganic
forms are Cu1* and CuOH*; however, Magnuson el al. (1979)
reported that [Cu^OHJj]5* was also toxic in some cases. Cowan
et al. (1986) performed a statistical analysis of the copper toxicity
literature and concluded that hydroxide species, but not car-
bonate species,  contribute to the aquatic toxicity of copper.
Meador (1991) studied the toxicity of ionic copper to Daphnia
magna in ah experiment in which the total copper, pH, and
naturally derived dissolved organic carbon were allowed to vary.
The same toxicity was achieved with less ionic copper as the pH
increased, which he explained by a competition of H* for Cu1*
at a receptor site on the cell surface.
  Many studies have related toxicity to the concentration of
free metal ion.  Judgment must be used in the evaluation of
studies conducted in constant or varying pH media. In constant
pH systems, the concentration of the free metal ion and the
monohydroxy complex (e.g-, Cu1* and CuOH*) covary. In mul-
tiple tests, when the pH of the system is changed, species other
than the hydroxy and carbonate complexes are affected. Com-
plexation by naturally occurring organic matter will change in
a nonlinear fashion;
  The concentration of organic ligands and the strength of their
binding of copper has been evaluated by a number of analytical
techniques (Neubecker and Allen, 1983). These chemical anal-
yses, together with computation of inorganic speciation, allow
computation of the free metal ton concentration.
  Water quality criteria and standards have been based on total
recoverable metal or on soluble metal concentrations (U.S. EPA,
1992). Recently, the US. EPA (Protnro, 1993) has recommended
the use of dissolved metal to set and measure compliance with
water quality standards, because "dissolved metal more closely
approximates the bioavailable fraction of metal in the water col-
umn than does total recoverable metal." Water quality criteria
{WQC) do  not  take soluble speciation of metal into account.
 42
            Water Environment Research, Volume 68, Number 1

-------
                                                                                                      Muen ana ransen
 For example, for cadmium, copper, lead, nickel, and zinc, hard-
 ness alone has been used to adjust fresh water WQC; no ad-
justment is used for saltwater WQC (U.S. EPA, 1986).
 <  Acute sensitivities of saltwater animals to copper in laboratory
 exposures in which an average of 83% of the copper was dissolved
 are reported to range from 5.8 ng/L (9.1 X 10~* moI/L) for the
 blue mussel to 600 jig/L (9.4 X 1(T* mol/L) for the green crab
 (ILS. EPA, 1985). A large fraction of the copper in natural water
 exists in complexed and biologically unavailable forms. In these
 natural waters! WQC derived from laboratory tests with lesser
 complexation capacities may be overprotective.
   The purpose of this paper is to demonstrate how changes in
 bioavailable metal, within the soluble metal pool, will result from
 changes in water quality, increasing concentration of metal added
 to a water, and the dilution of an effluent. The water effect ratio
 (WER), in which the ratio of toxicity test results for a site water
 compared with that for a reference water are used to correct the
 criteria, has been advocated as a means to account for the effects
 of metal speciation on a site specific basis (Stephan et at.. 1985;
 US. EPA, 1992; Prothro, 1993). We discuss  how the addition
 of metals to laboratory water versus natural surface waters will
 affect speciation; thus the WER determined for different organ-
 isms at a site. This discussion will emphasize why selection of
 appropriately sensitive species or tests is of particular importance
 if the WER is to actually account for metal's availability in site
water. From these considerations of chemical speciation, it is
clear that development of WQC based on bioavailable metal
would obviate the need for determining a WER.
   For clarity of presentation, we have chosen not  to consider
paniculate metal. However, a sorption model to account for
partitioning of metals to solids could have been included. The
WQC are based on soluble metals (Prothro, 1993). The sorption
of metal to solids removes the metals from the soluble pool and
reduces their toxicity.
  Although we specifically discuss the effects of speciation on
the bioavailabilhy of copper, the same principles apply to a
number of other cationic metals, including cadmium, chromium,
cobalt, lead, mercury, nickel, silver, and zinc. We have chosen
copper because the national water quality criteria in marine wa-
ters is very low, 2.9 jig/L, and approaches  background. Copper
tends to form strong complexes with naturally occurring organic
matter, and there have been a number of measurements of the
effects of water chemistry on the speciation and toxicity of
copper.

Principles of Chemical Speciation
   Metal ions, such as Cu2*, can form complexes with a number
of inorganic ligands, such as OH~, HCOj", NHj, and CT, and
with organic ligands, such as glycine (Stumm and Morgan, 1981;
Pankow, 1991). Because the stability constants for these reactions
are well known, the distribution of species can be easily com-
puted. The total concentration of soluble metal in the system is
the. sum of the free metal ion and of the metal contained in the
complexes. Stability constants for the  reaction of metals with
naturally occurring organic matter, such as humic acid, are be-
ginning to be developed, allowing the distribution of species in
a system containing naturally occurring organic matter to be
predicted. •
  Typical concentrations of inorganic  ligands such as chloride
are in the millimolar range. Thus, they tend to be in large stoi-
chiometric excess  relative to the micromolar  or lower concen-

January/February 1996
 tration of trace metals such as copper. The concentrations of
 hydroxide and carbonate are small, but they remain approxi-
 mately constant because of the buffering of the system. Conse
 quentfy, the addition of copper to a surface water sample for il
 titration or in a toxicity test decreases the inorganic lipnd con'j
 centration by a small and insignificant amount This results it!
 a constant ratio between free copper ion and the total concen-
 tration of inorganic copper species. The principal soluble inorj
 ganic complexes of copper in surface water are expected to be
 CuOH*. Cu(OHJj°, CuCO,°, and CutCOjfc2- The reactions fo!
 the formation of these complexes and their stability constant!
 (Stumm and Morgan, 1981) are as follows:                II
            Cu2* + OH' *r CuOH*;  KQH.I - 10"
         Cu2* + 2OH" ** Cu(OH)j0;

          Cu2* + CO,'" ** CuCO,0;
                                      10I2J
Cu2* + 2CO,2- *r Cu{CO,)j
                               2-
                                            10
                                              IMI
                                               Oil
 <3|
 <4;
  The fraction of free copper ion, Cu2*, in equilibrium witf
free copper ion plus all inorganic copper species, is «e»»* whictl
can be calculated from the concentrations of hydroxide ion'!
IOH~], and carbonate, [CO,'"], using the expression (Stumrr!!
and Morgan, 1981):  •    •                       ,        "
                             I
                                                     (5;
  The concentration of carbonate ions can be calculated from
the total alkalinity, [tot-alk], and the pH or from the concen-
tration of dissolved inorganic carbon, CT, and pH. The rela-
tionship between total alkalinity and pH is (Stumm and Morgan!
1981)
where «i and aj are the fractions of CT present as bicarbonate
and carbonate, respectively, where a( and ctj are given by the
expressions
                  [H*f
                                                    (7]
and
               *   JH*]2 + K.,[H*] + K.,K^             j

K.I and K^ in Equations 7 and 8 are the first and second acid
dissociation constants for carbonic add (Stumm and Morgan!
1981) as represented by the following expressions
          HjCO, ** H* + HCOr;  K., «=
and
    HCOr *r H*
                                        10
                                          -10 J
                                              (9]
(10]
  For a typical set of conditions, pH •= 7.8 and total alkalinity
• 2 X 10'J equiv/L (100 mg/L as CaCO,), a, = 0.966, «•
= 0.00306, and CT = 2.057 X 10~3 mol/L. The concentration
of carbonate is              <

[CO,2-] - o2CT - (0.00306X2.057 X 10~J)

                            » 6.29 X 10-* mol/liter  (11

-------
 AflenandHansen
 Equation 5 then becomes
or 2.4% of the free copper ion plus inorganic copper complexes
is free copper ion and the predominant inorganic complex is
CuCO/. It should be noted that for a fixed pH and total alka-
linity, the fraction of the free copper ion plus inorganic copper
species that will be present as the free copper ion (oc«») is un-
affected by the total copper concentration or the concentration
of copper present in organic complexes or adsorbed onto par-
ticulate matter. Likewise, the fraction of the copper that will be
present in any of the inorganic complexes, e.g., CuCO3*, will
also remain constant as the total concentration of copper varies.
  The concentration of copper binding sites present in organic
ligands extends from submicrocquivalents per liter concentra-
tions in sea water to tens of microequivalents per liter in rivers.
The organic ligands in natural waters are composed of humic
substances and other compounds that usually are not specifically
determined. The ligand concentration is commonly .considered
to be related to organic matter or organic carbon by an unspec-
ified and variable factor. Typically, the concentration of organic
ligands is in excess of the concentration of copper naturally pres-
ent in a sample.  However, addition of metal for a titration or
to establish certain test concentrations to be used in tqxicity
testing, will result in a stoichiometric excess of metal relative to
the concentration of the organic ligand. As will be shown, as1''
opposed to the inorganic complexation discussed above, the
fraction of the copper that will be complexed with the organic
matter will vary  as a function of the concentration of copper
present in the system.
  Naturally occurring organic matter contains a targe number
of ligands differing in concentration and stability constants.  A
number of different approaches have been taken to describe their
binding of protons and metal ions (Perdue and Lytle, 1983; Fish
ei al. 1986; Ephraim and Marinsky, 1986; Cabaniss and Shu-
man, 1988; Tipping era/., 1990). These approaches include dis-
crete site models, models with a continuum of binding sites of
varying pK, and models that incorporate electrostatic interac-
tions.                                         '
  Multisite binding models provide a better description of copper
titration data than does a angle site model. Because of its sim-
plicity, a single site model was used for the simulations presented
in this paper. A  single ligand model adequately illustrates the
effect of changing the ratio of copper to ligand on the inorganic
copper concentration, although it will not adequately charac-
terize metal binding to natural organic matter.
•  The complexation of copper by an organic ligand, L, can be
represented by the expression
                                                   (13)
where K.' is the thermodynamic formation constant for the for-
mation of the copper complex. Studies of copper binding in
natural waters often use electrochemical methods, such as anodic
stripping voltammetry or fixed-potential amperometry, for the
analysis of copper. These techniques respond to inorganic, but
not organic complexes, of copper in natural waters. Conse-
quently, the reaction in Equation 13 can be rewritten in terms
. of free copper ion plus inorganic copper complexes, Qijam,. Be-
 cause the charge on the organic ligand is generally not known,
 but is constant in a constant pH system, Equation  14 will be
 written without charges on the chemical species:
                                                                                                 tCuLl
                                                                                                               (14)
                                                            where K is the conditional formation constant for the formation
                                                            of CuL. The value of K is dependent on pH, ionic strength, and
                                                            on the concentrations of ions that can influence the above re-
                                                            action through participating in side reactions with either copper
                                                            or the organic ligand, for example, Zn2* and COja~..The sim-
                                                            ulations presented in this paper are based on Equation 14 in
                                                            which complexation is described by free copper ion plus inor-
                                                            ganic copper complexes.
                                                              The total copper concentration, [Cub, is given by the mass
                                                            balance on copper
                                                                            [Cu]T - ICuL] +
                                                    (15)
                                                            Likewise, the total ligand concentration, [L]T, is given by the
                                                            mass balance on ligand:

                                                                               [LJr «= [CuL] + IL]               (16)

                                                            where the total ligand concentration is frequently expressed as
                                                            the analytically determined complexation capacity. Equations
                                                            IS and 16 can be solved for [CuinwJ and [L], respectively. These
                                                            values substituted into the equilibrium constant expression in
                                                            Equation 14 yield the expression
                                                                                       tCiiL]
                                                                            ([CulT-[CuLMLlT-CCuL])
                                                                                                               (17).
                                                            for the average stability constant over the range of copper ion
                                                            added during the titration. The value of (Cu]r is the sum of the
                                                            concentration of copper in the initial water sample plus the con-
                                                            centration of copper added during the titration. The values of
                                                            K and [L]T are determined by evaluation of the titration curve
                                                            using the method of Ruzic (1982).

                                                            Effect of Complexatlon Parameters on Copper
                                                            Speclation  -
                                                              The concentration of inorganic copper in a sample is a func
                                                            tion of the concentration of total copper, total ligand, and the
                                                            stability constant for the reaction. This is shown by substitution
                                                            of Equations IS and 16 into Equation 14;
                                                                                                               (IS)
                                                            The concentration of free copper ion plus inorganic copper
                                                            complexes is given solving this quadratic equation. The fraction
                                                            of the copper present as free copper ion plus inorganic copper
                                                            complexes, a^-,,-,- is
                                                                                                               (19)
   Literature values for ligand concentrations and conditional
stability constants, as denned by Equation 14, for natural waters
are given in Table 1. Most of these values are for relatively un-
contaminated seawater. On the basis of these values, we have
used stability constants ranging from 1.0 X 107 to 1.0 X 10* and
ligand concentrations ranging from LOO X 10"* to 3.50 X 10~*
44
                                                                       Water Environment Research, Volume 68, Number 1

-------
                                                                                                       Allen and Hanssn
Table 1—Ligand concentrations and conditional stability constants for the complexation of copper in natural waters!
                                                    Ugand concentration
   Study area
  logK
     mol/U
jig Cu binding
 capaclty/L
         Reference
Adriatic Sea
North Atlantic
North Atlantic
Southern California
Delaware Bay
Nanagansett Bay
New York Harbor
North East Pacific
North Sea
Lake Ontario
Scheldt Estuary
Vancouver Island
7.5
8.3 to 10.0
7.8 to 8.6
7.8 to 8.6
7.6 to 8.2
7.3
6.8 to 8.4
8.5
7.3 to 8.4
8.6
7.2 to 7.7
6.6 to 7.8
1.3101.5X10-'
0.23 to 4.9X10-'
2.0 to 75X10-*
0.3 to 9.3 X 10'7
1.56to1.62X10-*
1.38 XUT'
3.46 to 5.49X10-'
7.6 X10-*
0.7102.0X10"'
3.4 X 10-' .
1.08 to 2.99X10-'
1.0 to 3.0X10-'
 826 to 9.53
 1.46 to 31.1
 1.27 to 4.58
 1.9 to 59.1
 0.99 to 1.03
 8.77
22.0 to 34.9
 0.48
 4.44 to 12.7
21.6
 6.86 to 19.0
 6.35 to 19.1
Plavsic ef a/. (1982)
Buckley and van den Berg (1985)
Kramer (1986)
Smaefa/. (1980)
Skrabalefa/.(1992)
Skrabal and Anen (1993)
Skrabal and Alton (1993)
Coate and Brutend (1988)
Kramer and Dufoker (1984)
Florence (1986)
Kramer and Duhker (1984)
Robinson and Brown (1991)
mol/L (6.35 to 222.4 fig copper binding capacity per liter) to
conduct simulations of copper complexation. Total copper con-
centrations were 2.90,29.0, and 145.0 jig/L (4.56 X 10-', 4.56
X tO~7, and 2.28 X 10"* mol/L), which correspond to the na-
tional water quality criteria for copper in marine water, and 10
and 50 times that value.
  The effect of ligand concentration on the binding of copper
for a stability constant of 2.0 X 10* is shown in Figure 1. Each
of the simulations corresponds to a dilution of a receiving water
at constant pH and alkalinity and then addition of copper to
give the indicated total copper concentration. We have not con-
sidered precipitation of Cu(OH)j (s) or CuO (s) in these or the
other simulations presented  in this paper; at concentrations
greater than that leading to formation of a solid phase at a par-
ticular pH, the system will be metastabte with respect to the
formation of precipitated Cu(OH)j (s) or CuO (s). As the ligand
concentration  decreases from 3.50  X  KT* (222.4 pg copper
binding capacity per liter), the concentration of free copper ion
plus inorganic copper complexes increases (Figure la). The con-
centration of free copper plus inorganic complexes increases as
the concentration of ligand is decreased. As the concentration
of ligand is decreased to a concentration similar to that of the
total copper present, the concentration  of free plus inorganic
copper dramatically increases. Consequently, the curves for the
different total copper concentrations do not parallel each other.
The fraction of copper  that is free copper ion plus inorganic
copper species depends not only on the ligand concentration,
but also on the concentration of the total copper present in the
sample (Figure 1 b). The remaining copper is present as the com-
plex with the organic ligand. For example, at a total ligand con-
centration of 5.00 X ID"7 mol/L (31.8 /tg copper binding capacity
per liter), the fraction of organically complexed copper decreases
from approximately 99% for a total copper concentration of
4.56 X 10-* mol/L (2.90 /ig/L) to approximately 94% for a total
copper concentration of 4.56 X 10~7 mol/L (29.0 /ig/L), to just
over 22% for a total copper concentration of 2:28 X 10~* mo!/
L (145.0 Mg/L).
  The dependency  of copper speciation on the stability constant
of a ligand present at a concentration of 5.00 X  10~T mol/L (31.8
pg copper binding capacity per liter) is shown in Figure 2. The
highest concentration of copper, 2.28 X  10"* mol/L (145.0 fig/

January/February 1996
                                                   S    S    2    S    3    S    3
                                                           Ugand (moVL)
                                                   •>.    «     t   '.«:    «8    «J    ^
                                                   S    8     S    £iS    8    §
                                                                   ^  *  *~    ^    o|
                                                    Ugand frig Cu binding capacity/L)
                                                1.000
                                                0.100
                                                0.010
                                                0401

t
»
t
V
\
\


""
«
t
C»'

'••.

-"-

••.

~"r':


'•;
».,w

M
(1>
**•*



* • »
                                   0    2
                                                        t    *   *   v-t    *   *
                                                        s    9J-A*a    s   3
                                                                 i    i   i    I   i
                                                        Ugand (fig Cu binding capadty/L)
                                      Rgure 1—Free plus Inorganic copper complexes as a
                                      function of ligand concentration for a stability constan
                                      of 2.0 X10* and total copper concentrations of (1) 145.0
                                      Mg/L (2.28 X 10-« mot/L), (2) 29.0 |tg/L (4.56 X 10~T mol/
                                      L), end (3) 2.90 jtg/L (4^6 X 10"* mol/L). Free plus Inor-
                                      ganic copper: (a) concentration and (b) fraction of total
                                      copper.        ,
                                                                                           45

-------
 Alton and Hansen
1O4 j

in-*
10-'-
to*-

*
i




"•"•
\
N.
"
3




	

'"•--
> '
r- . 1
i
i i i I i i i
d) I
!
|M
	 	 |<»)_

. •
- _ (3)
. .-. ~TT-*-
fe i i







3
~
1-835

-fi3 S
•6.35
•0635
•00635
•000635
!'
S      S
   K

0.100-


-*
* S
I -.^
V
\
.-x
X


«•
* * K




'--..

(1)
	
-p.


(2)
-.-'~!
                     S
     3
Figure 2—Free plus inorganic copper complexes as a
function of stability constant for a ligand concentration
of 5.00 X 10*' mol/L and for total copper concentrations
of (1) 145.0 jig/L (228 X 10~* mol/L), (2) 29.0 jig/L (4.56
X 10~r mol/L), and (3) 2.90 /ig/L (4.56 X10^* mol/L). Free
plus inorganic copper: (a) concentration and (b) fraction
of total copper.
L), is a little more than 4 times that of the ligand concentration
and the concentration of free copper ion plus inorganic copper
complexes is almost invariant with respect to changes in the
stability constant from 1.0 X 10T to 1.0 X 10'. At a total copper
concentration equal to or. less than the concentration of the li-
gand, the concentration of free copper ion plus inorganic copper
species (Figure 2a) and the fraction of the copper that is free
copper ion plus inorganic copper species (Figure 2b) decrease
with increasing values of the stability constant
  An environmentally important question regards the dilution
of a sample containing copper and a ligand. This is the situation
when a waste is discharged to the environment and dilution
occurs within a mixing zone, where the dilution water contains
no copper or ligand. We .have considered a waste effluent with
a total copper concentration of 2.50 X  10~T mol/L (1S.9 jig/L)
and a ligand having a stability constant of 2.0 X 10* (Figure 3).
In one case, the ligand concentration in the effluent is equal to
that of the copper. In the second case, the concentration of the
ligand in the effluent is S times that of the copper. The latter
case would be that expected for a waste effluent from a biological
wastewater treatment facility.-In the case of the effluent with
equtmolar ligand and copper, the concentration of free copper
                                                            ion plus inorganic copper complexes is greater than that which
                                                            would be expected  from dilution alone (Figure 3a). This is a
                                                            result of the effect of dilution on the dissociation of the copper
                                                            complex. When the effluent containing the equimolar metal and
                                                            ligand is diluted to one-half its initial concentration, the percent
                                                            of free copper ion plus inorganic copper complexes in the effluent
                                                            will increase from 13% to 18% (Figure 3b). For the sample that
                                                            contained a fivefold excess of ligand relative to copper, the cor-
                                                            responding increase in inorganic copper would be from 0.5% to
                                                            1.0%, resulting in the free copper ion plus inorganic copper
                                                            complexes doubling during the dilution. However, it should be
                                                            noted that the  concentration office copper ion plus inorganic
                                                            copper complexes in this case would be very low.
                                                              Typical titration  curves for copper additions to a ligand are
                                                            shown in Figure 4a. This figure considers the titration of the
                                                            same ligand at  four different concentrations ranging from 1,00
                                                            x 10-' to 2.00  X 10-* mol/L (6.35 to 127.1 pg copper binding
                                                            capacity per liter). It would appear that these curves differ only
                                                            in the position of the inflection point The inorganic copper
                                                            concentration has been plotted on a logarithmic scale in Figure
                                                            4b. This clearly shows the large differences in the concentration
                                                            of free copper ion plus inorganic copper complexes at a constant
                                                            small total copper concentration. It is obvious that titrations will
40 Iff*-)
3jj 10*.
3010*-
25 10*-
20 10*-
IJSIff*-
1 0 Iff* -





!*"" *«.










" -N. '
,



(2)





*%


,




(a)

d)
X
X






'



N i
- \
\

i_ Mt
~2£4 g
~ZSZ ' 'H.
-1*1 Q
Of


£
«», 1
-0.00 f
100
70 H
60-
50-
40.

20-


80





	

60 40
Effluent (%)



'

	


(b)



^- *^

20




^"

(2)
100 60 60 40 2
Effluent (%)
0

»
- 1
/
:
* ;
..-•
0 fl
                                     Figure 3—Free plus Inorganic copper complexes as a
                                     function of dilution of effluent containing 2.50 X 10"T mo!/
                                     L (15.9 |tg/L) total copper. Stability constant for copper
                                     complexation Is 2.0 X10* and ligand concentration In the
                                     effluent Is (1) 2.50 X 10~7 mol/L (15.9 /ig copper binding
                                     capachy/L) and (2) 125 X 10~* mol/L (79.4 pg copper
                                     binding capacity per liter). Free plus Inorganic copper:
                                     (a) concentration and (b) percent of total copper.
46
                                                                       Water Environment Research, Volume 68. Number 1

-------
                                                                                                     Aden and Hansen






X
B 1
§ <
1





['/ X
X
'""";•
> ^
i :




X
^x
(21'
.•
> 1
> •
« *

(a)

•X
X
f
«..•

, 1
S ;
*

X
X
,-'

....
h^ .
i 1
3 ;

X
.'
..•'

(x^'
1 1 - -
D 1
s ;
X
v ,
..-';

/'


> -1
\ ;
-222.4





~ 31^

3
                    Total Copper (mol/L)
                     •>»-:•
                     3    £   S   8
                    Total Copper big/L) .
41V*.
10''-
,.
#
* »
• F ^

?

if
' 0).
X "

1

* ,
(^

— :
. —

^'


, i •.<
j..
*•.-• -i • •
x''

(b)

> i




4
> '
r-*-*





i
OW g
»-
1

'4
rOJQQ635 £
A
        p    S    S    5.   S   3
                  Total Copper (mOfL)
        g    «g    s!'<*•:«
                             ^>  -  ip.
                  Total Copper (ns/L) '
1 1Q —
i
I
o E 10-».
> «A-*.



?"
/'

, ' "»

x
X
...'-*
'

9 "i
'"" 	 r-1 •
(2)
1 '^__



^.-"
.— •
'

. '*
-'-
»*f
	
(c)

> f
\
•--
«..
«-


, t
.-•••'
- —


> ?
'09^
-0.635


9
       0   2-   S   S   3   -3  -3    S
                 Total Copper {mot/L)
     t  O    W    f*»    ^   *^    W   ^    Ift
       A    M.MOMi    ^0^
             »    ".™N    OB    *
                  Total Copper Otg/L)

Figure 4—Free plus Inorganic copper complexes, as a
function of total copper concentration for ligand with sta-
bility constant for copper complexation of 1.0 X 10* and
ligand concentrations of (1) 1.00 X 10~T mol/L (6.35 fig
Cu binding capacity per liter), (2) S.QQX  10~T mol/L (31.8
jtg Cu binding capacity per liter), (3) 1.00 X  10~* mol/L
(63.5 |tg Cu binding capacity per liter), end (4) 2.00 X 10~*
V
8
6
4
2
0
0
(


•i
,





	

.


-^
/
I ^
/
/




•
0 5.0 10'7 1.010* 1.5 10* 2.010'* 2.5 104
Total Copper (mol/L)
) 31.8 63.5 95.3 127.1 158.S
                                                                                 TotaJ Copper
                                                             Figure 5—Free plus inorganic copper complexes as a
                                                             percentage of total copper concentration as a function
                                                             of total copper concentration for a llgand concentration
                                                             of 2.00 X 10~* mol/L, (127.1 ftg Cu binding capacity per
                                                             liter) and stability constant of 1.0 X10*. v
                                                             have different free copper ion plus inorganic copper complexes
                                                             (and consequently free copper) concentrations at all total copper
                                                             concentrations if the waters being titrated differ in ligand con-
                                                             centration. The fraction of copper that is present in the free
                                                             copper ion plus inorganic copper species does vary if. the total
                                                             copper concentrations is less than the concentration of the ligand
                                                             present in the system (Figure S).  Thus, for small additions of
                                                             copper to a sample, the increase of the concentration of copper
                                                             present as free copper ion plus inorganic copper complexes is
                                                             not proportional to the increase in the concentration of total
                                                             copper present. Of course, for copper additions in excess of the
                                                             concentration of ligand present, the increase in free copper ion
                                                             plus inorganic copper complexes begins to more closely mirror
                                                             the increase in total copper.
                                                               In a like manner, two waters having the same total concen-
                                                             tration of ligands, but with different equilibrium constants, will
                                                             have different concentrations of free'copper ion plus inorganic
                                                             copper complexes at all total copper concentrations. Conse-
                                                             quently, neither free copper ion concentration nor the concen-
                                                             tration of free copper ion plus inorganic copper complexes is
                                                             directly proportional to the total copper concentration at different
                                                             concentrations of total copper during the titration of a water, at
                                                             the same concentration of total copper for samples of a water
                                                             that have been diluted to different extents, or for different water
                                                             samples containing the same total concentration of copper.
                                                               The speciation of copper is highly dependent on the pH of
                                                             the environment In addition to the effect that pH has on the
                                                             distribution of inorganic copper species, complexation with or-

                                                             mol/L (127.1 M9 Cu binding capacity per liter). Free plus
                                                             inorganic copper (a) linear axis, (b) exponential axis, and
                                                             (c) exponential axis and low concentrations of total cop-
                                                             per.
January/February 1996
                                                                                                                 47

-------
 Men and Hansen
ganic ligands is affected by changes in the pH or the system.
Protons compete with copper ions for binding sites on carboxylic
acids and other organic compounds. Titration curves for the
addition of copper to a secondary sewage effluent at pH values
tanging from 6.5 to 8.5 are presented in Figure 6a. Complexation
parameters for the- wastewater effluent are given in .Table 2. As
the pH rises, both the stability constant and the complexation
capacity increase. As the hydrogen ion concentration increases
by one order of magnitude (decrease of 1 pH unit), the concen-
tration of free plus inorganic copper complexes can increase by
more than  an order of magnitude (Figure 6b).
   Caution  must  be exercised in the use of experimentally de-
termined complexation capacities to express ligahd concentra-
tions. Perdue (1988) has shown that the complexation capacity,
evaluated by thechange in slope of the titration curve, is a func-
tion of the stability constant of the ligand in addition to the
concentration of the Hgand.
Table 2—Parameters for copper complexation by sec-
ondary wastewater effluent  (Crasser, 1980). Titration
simulations are shown in Figure 6.
B.O 10'-
7.0 10*-

5.0 10*-

J.O1U -

1.0 W-
0.0 H
i


. -
-



.«
'f**
, ;



—


f
if,'
» "
,..,
nHA
pH7
PH7
pHB
pH8


f
,'
"tuT
» i
,.„
i
|
8

s,-
t'\
.*
3 1




/
r'
f
*
if™':
» 1

(a)

s
/'
f
r
,
/

9 1


S
/
/
f

/


» 1

/
/
f
r
i j*
/


.<**"
» 1
/ .
,'

f


*~

»• 1








* \
•9UBA 5
-4448 ^
-9A1 3 O
.4477 0>
S>
.9540 

>
           ~2SS3SSK5S5
                     Total Copper (mol/L)
       .   •q«^«o«h:»«i^o>«
           ° .»   S  S  2  5  i   I  I  6  8
                     Total Copper (jig/L)
     10*

//
f '
i . ^
/
/
^
/


,'''
<
'•'


_^ -
x



-
'•'-•
<
.,.
(•>)


--

.-'•



r.-~

-•'



TI
t

-
-

rj«i
—
/


• » »

^


JH6
)H7
JH8
>H8




5
9
0
S


. i
655 1


H

         rf   5  S   S  3   S  S 5  S  S  2
                   Total Copper (mot/L)


                    Total Copper (ug/L)
Figure 6—Concentration of free plus Inorganic copper
complexes in a secondary sewage effluent titrated at dif-
ferent values of pH. Free plus inorganic copper, (a) linear
axis and (b) exponential axis. Complexation parameters
are given in Table 2. Precipitation of Cu(OH)a (s) or CuO
(s) are not considered in these  calculations.
pH
6.5
7.0
7.5
8.0
8.5
Stability constant
5.38 X 10*
1.61 X 10»
1.82 X10«
1.15X10'
6.05 X 10*
ugana
mol/L
1.26 X 10-*
258X10-*
2.31 X 10-*
3.67 X Kr*
7.30 X W
i conconffauoti
jig Cu binding
capacity/!.
80.1
1445
146.8
233.2
465.8
Effect of Copper Speclation on Biological Response
  Toxicity tests are conducted by the addition of several different
concentrations of metal to the test water to expose organisms
in the water for a fixed period of time after which the biological
response, frequently death, is evaluated. As indicated earlier, the
organic complexes  of copper are not  biologically active.. It is
clear that to relate the biological response to the concentration
of copper present in a water, it is necessary to develop a rela-
tionship between total copper and the concentration of biolog-
ically active copper. In the case of a bioassay, the concentrations
office copper ion and of each of the inorganic copper complexes
are  proportional to the concentration of free copper ion, plus
inorganic copper complexes if all test chambers have the same
pH  and alkalinity. Consequently, it is important to evaluate the
change in free copper ion plus inorganic copper species that
results from the addition of copper to a test  water. Such
relationships are the titration curves that have been shown in
Figure 4.
  We have used this basic approach to evaluate the effect tbitt
consideration of chemical speciation would have on water qualify
criteria. We based our calculations on the results of a titraiica
of a sample of water that was collected in Narragansett Bay in
the  Winter of 1993. For that sample, we determined a stability
constant of 727 X I07 and a ligand concentration of 1J8 X Hr *
(8.77 jig Cu binding capacity per liter) using anodic stripping
voftammetry (Skrabal andAllen, 1993).
  For illustration purposes, we have assumed that all bioloycit]
results reported in the U.S. EPA's criteria document for copper
(1985) were determined in Narragansett Bay water having these
complexation properties. The  four biological species and  the
genus mean acute values that were used to establish the criterion
are  soft-shell clam (39.00 pg/L), oyster (14.92 jig/L), summer
flounder (13.93 ftg/L), and blue mussel (5.80 fig/L).
  We computed the concentration of free copper ion plus in*
organic copper species that would have been present at the genus
mean acute values for the four most sensitive genera in salt
water in the water quality criteria document We assumed that
the  complexation parameters of the water used for the bioassay
tests for each of the four species were those for the Narragansett
Bay sample that was collected .in the winter of 1993. Although
the  complexation of copper is small, it is an important factor at
low copper concentrations! Using these complexation. parame-
ters, the fraction of copper that is not bound to organic matter
at a total copper concentration of 9.1 X 10'1 mol/L  (5.8 jig/U,
which is the reported acute toxicity for the blue  mussel,, is
only 17.9%.'
48
                                                                         Water Environment Research, Volume 68. Number. 1

-------
                                                                                                       Allen and Hansen
   We have followed the procedure to establish the WQC for
 copper in marine waters (Stephan et a!., 1985) and, using the
 computed concentrations of free copper ion plus inorganic cop-
 per species, we have also computed a WQC for copper based
 on bioavailable copper. The acute toxicities for each of the four
 species are plotted versus both the reported total copper con-
 centrations and the computed concentrations of free copper ion
 plus that of the inorganic copper complexes (Figure 7). Using
 the reported total copper concentrations, the final acute value
 for copper in salt water is 5.832 pg/L (9.178 X 10~* moI/L).
 Because larva! molluscs are most sensitive to copper, the criteria
 maximum concentration (CMC) and criteria continuous con-
 centration (CCQ are both derived by dividing the final acute
 value by 2.0 and thus are the same 2.916 /
-------
Allen and Hansen
of effluent discharge include pH of the receiving water less than
that of the. effluent, high concentration of soluble or of paniculate
metal in the discharge, and low concentration of organic matter
and of particles in the receiving water.
   the above discussion provides a conceptual framework in
which to develop the data necessary to include metal speciation
in waste load allocation and discharge permits.

Effect of Copper Speciation on Water-Effect Ratio  .
   Determination of a WER has been suggested as a means to
account for the effects of differences in chemical speciation on
the toxicity of copper and other metals in natural waters (U.S.
EPA, l992;Trothro,  1993). In this procedure, toxicity tests are
conducted in a site water and in a reference (laboratory) water.
Reference water tests are used as surrogates for the laboratory
tests that were used to derive national criteria. The ratio of the
toxitities is used as a multiplier to adjust the national water
quality criteria (NWQQ to account for differences in bioavail-
ability, as measured by toxicity tests, that would be applicable
to that site. For example:
Site - Specific WQC - NWQC X WER

                              site - water LC50
                  NWQCX
                            reference - water LC50
               (20)
Brungs et al. (1992) have summarized the results of a number
of determinations of water effect ratios for different toxicants.
Their reported water effect ratios for copper are presented in
Table 3. It should be noted that, for a single location, the WER
usually increases as the organism sensitivity to copper increases.
This observation is consistent with the changes in metal specia-
tion as the concentration of total metal increases that we dis-
cussed earlier, and it must be considered in designing WER
studies.   •              '      •       "
  To further illustrate the importance of species sensitivity to
the WER, we calculated WERs using acute values from five
species having different sensitivities to copper and copper spe-
ciation for hypothetical reference and site waters. The acute val-
ues used in this calculation were 159.8 pg/L for mysid, 120 jig/
L for polychaete, 39.97 pg/L for a copepod, 14.92 pg/L for the
  oyster, and 5.8 fig/L for the mussel, as obtained from the criteria
  document. Further, Nanagansett Bay water with a stability con-
  stant of 7.27 X 10' and a ligand concentration of 1.38 X 10"'
  mol/L (8.77 fig Cu binding capacity per liter) was assumed; to
  have been the water for which these acute toxicity values were
  determined. We then computed the concentrations of free copper
  ion plus inorganic copper complexes that would have been ap-
  propriate to each species-specific acute value if laboratory tests
  would have been conducted in Narragansett Bay water (Table
  4). Using these concentrations of free plus inorganic copper,
  total copper concentrations required to produce these acute .val-
  ues were calculated for each species and the hypothetical labo-
  ratory reference water and the four hypothetical site waters, each
  with its own stability constant and complexation capacity (Table
  5). The procedure used to compute the acute toxicity values in
  the hypothetical laboratory reference and the site waters is shown
  in Figure 8.  •    .      '            '
    The toxicity of copper to each of the five organisms selected
  is approximately the same in both the Narragansett Bay  water
  and in the hypothetical reference site water (Figure 9a). In the
  four sites being evaluated, less total copper is needed to produce
  50% mortality (the LCSO concentration) than is needed for the
  Reference Site water even though the same concentration of free
  copper ion plus inorganic copper complexes is needed for a given
  species. In all cases the water effect ratio increases with the sen-
  sitivity of the organism, expressed as the total copper concen-
  tration causing acute toxicity (Figure 9b). The water effect  ratios
  for Site 2 are greater than those for Site 1 and the water effect
  ratios for Site 4 are greater than are those for Site 3.
   . It should be noted that for the three lean sensitive species
  (mysid, polychaete and copepod) the copper LCSO values, for
  Sites 1 and 3 and for Sites 2 and 4 are almost identical (Figure
  9a). Likewise, the WER values for Sites 1 and 3 and for Sites 1.
  and 4 are virtually identical (Figure 9b). The LCSO total coppei
  concentration required to be present in these four waters at the
  LCSO value for these three organisms is greater than the con-
  centration of organic ligand available for complexation of. the
  added copper. In this situation the toxicity of metal is primarily
  controlled by the excess of metal relative to the ligand present
  and is only secondarily related to the stability constant .of the
Table 3—Water effect ratio study results for copper.
         Site.
  * Fran Brungs ef af.. 1992.
  * From Thursby. 1993.
Specie*
Laboratory water acute value, |tg Cu/L
Water effect ratio
Naugatuck River, Com.*

St Louis River. Minn*



Nemadji River, We* ,
Uttle Pokegama River, Wis*
New York Harbor, N.Y."




Ceriodaphniadubla
Pimephates promotes
. DapMa magna
Arnphipod
Pimephates pmmelas
. Qncriorfiyncfius mytdss
Arnphipod . .
Arnphipod
Champia parvuta
Mytitusedutts .
Arbac/a punctulate
Mulinia teteraffs
MysVopsis bahla
17
85
17
25
84
120
90
90
11
26
49
35
SOB
1.1
1.0
15
92
'- 6.3
3.3
43
3.1
32
13
1.8
1.6
1.5
SO
                                    Water Environment Research, Volume 68, Number 1

-------
                                                                                                        Aflen and Hansen
 Table 4—Toxlclty of copper to organisms considered In simulated water effects ratio study.

                                                             Genus mean acute value
                                           Total copper*
                                                 Free plus inorganic copper
                                                        complexes"
         Organism
HOfl-
mol/L
MI/L
mol/L
Mysid
Pdychaete
Copepod
Oyster
Mussel
159.8
120
39.97
, 14.92
5.8
2.51 X KT6
1.89 X10-*
6.29 X 10-'
2.37 X 10"'
9.13 X 10-*
151.3
' 110.8
31.5
7.18
1.04
2.38 X 10-*
1.74 X 10-*
4.95X10"'
1.13 X 10'?
1.63 X 1
                            2 and 4 art 2.46 and 2.47. respectively; the WERs for mussel
                            for Sites 2 and 4 are 10.9 and 22.7, respectively.
                              If WERs are to accurately adjust WQC to reflect metals avail-
                            ability at specific sites, testing must use species sensitive at or
                            just above criteria concentrations. Use of insensitive species or
                            tests may result in low WERs and conservative site-specific cri-
                            teria. Conversely, use of species or tests sensitive below the cri-
                            terion may  result in a large WER and a site-specific criterion
                            that is not sufficiently protective. This is to be avoided. For cop-
                            per, acute tests with larval mussels or other molluscs that are
                            most sensitive will result in the highest values of the WER. Be-
                            cause acute and chronic  criteria for some other metals differ
                            markedly, use of chronic tests sensitive at the criteria continuous
                            concentration should result in larger WERs more appropriately
                            protective at this chronic criteria concentration.
Table 5—Copper complexation parameters for water effect ratio simulation. Results are shown In Figure 9.
           Sample
                                                                              Ugand concentration
  Stability constant
          mol/L
         *ig Cu binding capadty/L
Narragansett Bay
Reference Site
Sitel
Site 2
Site3
Site 4
7-27X10*
9.00 X 10*
3.50 X 107
3.50 X101
2.00 X ,10"
2.00 X 10*
1.38 x to-1;
2.00 X 10-'
1.00 X10-*
4.00X10-e
, 1.00X10-*
4.00X10-*
8.77
12.71
63.55
254.18
63.55
254.18
January/February 1996
                                                                                 51

-------
                     Total Copper (mol/L)
           ctm^»ft|h.n«.   •«;
           0    »   *   I    i   s •   i   I   §
                     Total Copper (jiS/U
 Figure 8—Relationship between free plus inorganic cop-
 per complexes and total copper for the water at the site
 used to develop the water quality criteria (dashed line)
 and Site 4 water for the assessment of water effect ratio
 (solid line). The relationship is shown for the toxieity of
 copper for the three least sensitive, organisms, used In
 the evaluation as discussed in the text
.   As a consequence .of the effect of the ligand concentration
 and stability constant on metal spcciation, and consequently on
 the bioavailability of copper, it can be concluded that the organic
 matter content of a natural water (in the absence of paniculate
 matter) should be predictive of the WER for nonsensitive or-
 ganisms such as mysid or polychaete. However, for more sen-
 sitive species, such as the blue mussel, knowledge of the organic
 matter content alone will not allow the bioavailability of copper
 to be predicted. For sensitive organisms, a more complete de-
 scription of metal complexation will be required to predict bio-
 availability. This'will necessitate either measurement of specific
 chemical species or the determination of both the concentration
 of ligands and the stability constant for their complexation with
 copper so that the concentration of bioavailable copper can be
 calculated.  If the  quantity and/or  quality of  organic matter
 changes seasonally, one WER may not be appropriate for the
 entire year. Measurements or estimates of bioavailable copper
 concentrations and a copper criterion based on bioavailable
 metal, if appropriately determined, would be preferred over the
 present site-specific approach.
   Conduct of WER studies as specified in U.S. EPA (1994) and
 simultaneous measurement of specific chemical species or sol-
 uble and paniculate ligand concentrations and their stability
 constants will permit comparisons of WERs derived lexicolog-
 ically and by chemical theory. If they are similar, WERs derived
 using chemical theory may be most appropriate because they
 can be calculated for the exact criteria concentration and are
 less expensive.
 Summary
   We have provided a detailed description of the complexation
 chemistry of copper in natural waters and its importance to
 copper speciation and to bioavailability, in particular its rela-
                                                             tionship to water quality criteria and to water effect ratios. In a
                                                             water at a constant pH and total alkalinity and at thermodynamic
                                                             equilibrium, the concentrations of free copper ion and of all
                                                             inorganic  species covary as the'total copper concentration is
                                                             changed. The bioavailability and toxieity of copper in a system
                                                             at constant pH and alkalinity will be proportional to the con-
                                                             centration of free copper ion and of inorganic copper complexes.
                                                             Because dissolved organic matter is present at low concentrations
                                                             in natural waters and forms strong complexes with copper, the
                                                             concentration of copper complexes with dissolved organic matter
                                                             changes in a nonlinear manner during the course of the titratton
                                                             of a sample of natural water with copper. Because of this non-
                                                             linearity, which is most pronounced at concentrations of copper
      500
      400
      300
      200
                                                                   100-



pH
• /
• t
• f
> t
. /
, f
£

s
s
s
s
s
s
V
V
s
s
(*)


p
— t
s
s
- s
» V
» s
* s
\ V
\ s
» s
-i
0 Narragansett Bay
M Reference SHe
N Sltel
• Site 2
m sites
U Site 4


a!


PMS

|-j

J
•
,
Figure 9—Simulated results for the development of a site
specific criterion for copper. Stability constants for Sites
1 and 2 are 3.5 X 10T and 2.0 X 10* for Sites 3 and 4.
Ligand concentrations are 1.00 X10** mol/L (63.5 pg Cu
binding capacity per liter) for Sites 1 and 3 and 4.00 X
10~* mol/L  (254.2 pg Cu  binding capacity per liter) for
Sites 2 and 4. (a) Total copper LC50 concentrations and
(b) water effect ratios.
 52
                                                                         Water Environment Research. Volume 68. Number 1

-------
                                                                                                       ASen and Hansen
    less than the equivalent concentration of ligand, the concentra-
    tion office copper ion plus inorganic copper complexes will not
   . be the same for different waters having the same total copper'
 '   concentration. Furthermore, the fraction of the copper present
    as the free'metal ion plus inorganic copper complexes does not
    change proportionally to the concentration of ligand or to the
    stability constant of the complex.
       Discharged wastewater typically contains copper and an excess
    of ligand. As this is diluted in the receiving stream, an increased
    fraction of the copper is present in the form of free copper ion
    plus inorganiccopper complexes, but because of the concurrent
    dilution, the concentration of free copper ion plus inorganic
    copper complexes decreases. If the pH of the receiving water is
    not the same as that of the effluent,  the conditional stability
    constant will change as will the ligand concentration. The ability
    of an effluent to complex copper may decrease by more than
    one order of magnitude when the pH is lowered by one unit
       It appears that if water quality criteria for copper in marine
    water were to be based on free copper ion plus inorganic com-
    plexes rather .than total  dissolved copper, the criterion would
    change only slightly. Therefore, the importance of developing a
    national criterion based on bioavailable metal is not that it will
    markedly change the criterion, but that national and site-specific
    criteria should not differ if both are expressed in terms of the
    bioavailable form.
      If receiving water standards are in terms of dissolved or bio-
    available metal, it is necessary to predict the rate and extent of
    transformation of metal from total recoverable metal to the reg-
    ulated form below the point of discharge to enable appropriate
    waste load allocation and issue discharge permits that are neither
    over nor under protective.
      On the basis of chemical speciation as a function of total
    copper concentration, water effect ratios are expected to increase
    as more sensitive  organisms are used in the evaluation. This
    prediction is borne out by experimental results. For insensitive
    organisms, the concentration of ligand present is less than the
    copper added to achieve the toxic response. The toxic response
    is primarily determined by the difference in concentration be-
    tween copper and the ligand. However, when the most sensitive
<    organisms are used in the test, the toxic response is obtained for
    concentrations of metal less than the concentration.of the ligand.
    The toxicity of the metal depends on both the concentration of
    ligand in the sample and the stability constant for the complex-
    Ation reaction.
      One goal of this simulation was to demonstrate the potential
    practical importance of theoretical calculations  of WERs as
    checks against their empirically obtained values as presently re-
    quired in  EPA guidance. If appropriate data are derived, in-
    cluding measurement of specific chemical species or ligand con-
    centrations and stability constants, as part of WER studies (U.S.
    EPA, 1994) it may be possible to demonstrate the practical utility
    of WERs  calculated directly from chemical properties of site
    water. This could permit less expensive and more accurate de-
 ,   terminations of site-specific adjustments to national WQC that
    are appropriately protective. The WER approach'to adjust na-
    tional WQC should be considered temporary with the ultimate
    goal of understanding and accounting for all metal-containing
    phases in water that account for the true biologically available
    metal in waters that represent all surface water types. Through
    this basic understanding, WQC for metals and other substances
    can be expressed on a biologically available concentration basis.
                                                             Acknowledgments
                                                               Credits. The authors thank Wilson Jardim, Universidade Es-
                                                             tadual de Campinas; Bo Shi, University of Delaware; E. Michael
                                                             Perdue, Georgia Institute of Technology; Dominic Di Toro,
                                                             HydroQual; Warren Boothman and Charles E Stephan, US.
                                                             EPA; Joseph S. Meyer, University of Wyoming; and two anon-
                                                             ymous reviewers for valuable comments. This document has
                                                             been reviewed according to EPA laboratory requirements, and
                                                             its content does not necessarily reflect agency policy.
                                                               Authors. Herbert E Allen is a professor at the Department
                                                             of Civil and Environmental Engineering, University of Delaware,
                                                             Newark. David J. Hansen is a research aquatic biologist at the
                                                             US. Environmental Protection  Agency, Environmental Re-
                                                             search Laboratory, Narragansett, Rhode Island. Correspondence
                                                             should be addressed to Herbert E. Allen, Department of Civil
                                                             and Environmental Engineering, University of Delaware, New-
                                                             ark, DEI97I6.
                                                               Submitted for publication April 1.1994; revised manuscript
                                                             submitted February 14,1995; accepted for publication March 8.
                                                             1995. Deadline for discussions of this paper is May 15. 1996.
                                                             Discussions should be submitted to the Executive Editor. The
                                                             authors wilt be invited to prepare a single Closure for all discus-
                                                             sions received before that date.

                                                             References
                                                             Allen. H. E (1993a) The Significance of Trace Metal Speciation  for
                                                                Water, Sediment and Soil Quality Standards. Set.  Total Environ..
                                                                134,23.
                                                             Allen, H. E (1993b) Importance of Metal Speciation to Toxicity. Aw.
                                                                Water Environ. Fed, Prteoaferenee Seminar Aquatic Life Criteria
                                                                for Metals. Anaheim, CA, 55.
                                                             Allen, H. E, Boonlayangoor, C, and NoO. K. E (1982) Changes in
                                                                Phyacochemical Fonns of Lead and Cadmium Added to Freshwater,
                                                                Environ. Internal.. 7,337.
                                                             Benson, W. HH Alberts, J* Allen, H. E, Hunt, C O, and Newman,
                                                                M. C (1994)  Synopsis of Discussion Session on BJoavailabflity of
                                                                Inorganic Elements: la A Mechanistic Understanding of Bioavail-
                                                                abflity: Physical, Chemical, and Biological Interactions. J. L. Ha-
                                                                melfflk. W. H. Benson, H. L. Bagman,P. F. Landnun (Eds.), Lewis,
                                                                Boca Raton, Fla, 69-77.
                                                             Brown, V. M., Shaw, T.L, and Shuiben, D.G, (1974) Aspects of Water
                                                                Quality and the Toxicity of Copper to Rainbow Trout. Water Res.
                                                                (O.B.),8,797.
                                                             Brune$,W.A,HoWennan,T.S^andSoutherian,M.T.(1992) Synopsis
                                                                ofWater-Effect Ratios for Heavy Metals as Derived for Site-Specific
                                                                Water Quality Criteria. EPA Contract 68-CO0070, VS. EPA,
                                                              .  Washington, D.C  .
                                                             Buckley, P. J. MM and van den Berg, C M. G. (1986) Copper Com-
                                                                pfexation Profiles in the Atlantic Ocean. A Comparative Study Using
                                                                Efectrochemkal and Ion Exchange Techniques. Mar. Chart.,  19,
                                                                281.
                                                             C«baniss,S.E,»ndShuman.M.S.(l988)CciiperBii«Iin8byE«ssolved
                                                                Organic Matter: L Simannee River Fulvic Acid. Geochtm. Co*-
                                                                moehim.Aaa.SMtt.
                                                             Chakoumakos, O, Russo. R. G, and Thuiston, R. V.  (1979) Toxicity
                                                                of Copper to Cutthroat Trout (Salmodarldi) under Different Con-
                                                                ditions of Alkalinity, pH, and Hardness. Environ. Sci. Technol., 13,
                                                                 213.
                                                             Coale, K. H., and Bruland, K. W. (1988) Copper Compilation in the
                                                                 Northeast Pacific. Llmnol. Oceanogr., 33,1084.
                                                             Cowan. C E, Jenne, E A, and Kinnison, R. R. (1986) Methodology
                                                                 for Determining the Relationship between Toxicityand the Aqueous
                                                                 Speciation of a Metal. Aquatic Toxicity and Environmental Fate
                                                                 Ninth Volume. ASTM STP 921. T. M. Poston and R. Purdy (Eds.)
                                                                 American Society for Testing and Materials, Philadelphia, Pa., 463.
January/February 1996
                                                                                                                   53

-------
Alton and Hansen
Grosser, P. (1980) Investigation of Copper-Natural Ligand Complexes
    by MnOj Adsorption. M.S. thesis, Illinois Institute of Technology,
    Chicago, III
Ephraim, J., and Marinsky, J. A. (1986) A Unified Physicochemical
    Description of the Protonation and Metal Ion Complexation Equi-
    libria of Natural Organic Acids (Humic and Fulvic Acids). 3. Influ-
    ence of Polyelectioiyte Properties and Functional Heterogeneity on
    the Copper Ion Binding Equilibria in an Amudale Horizons Bb
    Fulvic Acid Sample. Environ. Set.  Technoi., 20,367.
Fish, W, Dzombak, D. A., and Morel. F. M. M. (1986) Metal-humate
    Interactions. II. Application and Comparison of Models. Environ.
    Sci. Techno!.. ID, 616.
Florence, T. M. (1986) Electrochemical Approaches to Trace Element
    Speriation in Waters: A Review. Analyst. 111,489.
Hodson. P. V., Borgmann, U, and Shear. H. (1979) Toricity of Copper
    to Aquatic Biota: In Copper in the Environment. Part II. Health
    Effects. J. O. Nriagu (Ed), Wiley-Interscience, New York, N.Y.
Howarth, R. S., and Sprague, J. B. {1978) Copper Lethality to Rainbow
    Trout in Waters of Various Hardness and pH. Water Res. (G.B.),
    12,455.
Kelly, M. (1988) Mining and  the Freshwater Environment Elsevier,
    Essex, UK.
Kramer, C J. M. (1986) Apparent Copper Complexation Capacity and
    Conditional Stability Constants in North Atlantic Waters. Mar.
  .  Chan.. 18,335.
Kramer, C. J. M., and Duinker. J. C. (1984) Measurement of Copper
    Complexation by Naturally Occurring Ugands. In Complexation
    of Trace Metals in Natural Waters. C. J. M. Kramer and J. C
    Duinker (Eds.)  NUhoff/Junk. The Hague, The Netherlands, 217.
Luoma, S. N. (1983) Bioavailability of Trace Metals to Aquatic Organ-
    isms—A Review. Sci. Total Environ.. 28,1.
Magnuson, W. R., Harriss, D. K., Sun, M. S., Taylor, D. K., and Glass,
    G. E (1979) Relationships of Activities of Metal-Ligand Species to
    Aquatic Toxicity: In Chemical Modeling in Aqueous Systems: Spe-.
    datum, Sorption, Solubility and Kinetics. E. A. Jenne (Ed.) Syrup.
    Ser. 93. American Chemical Society, Washington, D.C.
Meador, J. P. (1991) The Interaction of pH, Dissolved Organic Carbon,
    and Total Copper in the Determination of Ionic Copper and Tox-
    icity. Aquat. Toxicol.. 19,13.                •
Neubecker, T. A, and Allen. H. E. (1983) The Measurement ofCom-
  '  plexation Capacity and Conditional Stability Constants for Ligands
    in Natural Waters—A Review. Water Res. (G.B.), 17,1.
OTtonnel, J. R., Kaplan, B. M., and Allen, H. E. (1985) Bioavailability
    of Trace Metals in Natural Waters: In Aquatic Toxicology ano* Haz-
    ard Assessment Seventh Symposium. American Society for Testing
    and Materials, Philadelphia, Pa.
Pagenkopf, G. K., Russo, R. C., and Thurston, R. V. (1974) Effect of
    Complexation on Toxicity of Copper to Fishes. J. Fish. Res. Bd.
    Can.. 31,462.
Pankow, J. (1991) Aquatic Chemistry Concepts. Lewis, Chelsea, Mich.
Perdue, E M. (1988) Measurements of Binding Site Concentrations in
    Humic Substances: In Metal Speciation—Theory, Analysis and
    Application. J. R. Kramer and H. E Allen (Eds.) Lewis, Chelsea,
    Mich., 135.
Perdue, E M:, and Lytle, C R. (1983) Distribution Model for Binding
    of Protons and Metal Ions by Humic Substances. Environ. .Set.
    r«*«»l., 17,654.
Plavsic, M., Krznaric, D., and Branica, M. (1982) Determination of the
    Apparent Copper Complexing Capacity of Sea Water by Anodic
    Stripping Vottammetry. Mar. Chem., 11,17.
Prothro, M. G. (1993) Memorandum: Office of Water Policy and Tech-
    nical Guidance on Interpretation and Implementation of Aquatic
    Life Metals Criteria. US. EPA, Washington. D.C
Robinson, M. G.. and Brown, L N. (J991) Copper Complexation during
    a Bloom of Gymnodinium saaguiaeum Hirasaka (Dinophyceae)
    Measured by ASV. Mar. Chem., 33,105.
Ruzic, I. (1982) Theoretical Aspects of the Direct Titration of Natural
    Waters and Its Information Yield for Trace Metal Speciation. Anal
    Chim. Ada, 140,99.
Skrabal, S. A., and Alien, H. E (1993) Unpublished results.
Skrabal, S. A.. Luther, G. W., and Allen, H. E (1992) Chemistry and
    Bioavailability of Copper in Indian River Bay, Delaware. Final report
    to Electric Power Partners.                        .   '
Sma, R. F.,Garrett, K. S., Miner, S. M.,andThum, A. B. (1980) Copper
    Complexation Capacity of Marine Water Samples from Southern
    California. Environ. Set. Technoi., 14,1482.
Stephan, C E, Mount, D. I., Hansen, D. J., Gentile, J. H., Chapman,
    G. A., and Brungs, W. A. (1985) Guidelines for Deriving Numerical
    National Water Quality Criteria for the Protection of Aquatic Or-
    ganisms and Their Uses. US. EPA, Office of Research and Devel-
    opment PB85-227049. NTB, Springfield, Va.
Stumm, W, and Morgan, J. J. (1981) Aquatic Chemistry. John Wiley
    and Sons, New York, N.Y.
Thursby, G. (1993) Personal communication.
Tipping, E, Reddy. M. M., and Hurley, M. A. (1990) Modeling Elec-
    trostatic and Heterogeneity Effects on Proton Dissociation from
    Humic Substances. Environ. Sci. Techno!.. 24,1700.
US. EPA (1985) Ambient Water Quality Criteria for Copper—1984,
    EPA 440/5-84-031. VS. EPA, Washington, D.C.
US. EPA (1986) Quality Criteria for Water 1986. EPA 440/5-86401.
    US. EPA, Washington, D.C.
US. EPA (1992) Interim Guidance on Interpretation and Implemen-.
    tation of Aquatic Life Criteria for Metals. US. EPA, Washington^
    D.C.                                  .
US. EPA (1994) Interim Guidance on Determination and Use of Water-
    Effect Ratios for Metals. EPA 823*94-001. VS. EPA, Washington,
    D.C
van den Berg. C M. G., Wong, P. T. S., and Chau, Y. K. (1979) Mea-
    surement of Complexing Materials Excreted from Algae and Their,
    Ability .to Ameliorate Copper Toxicity. /. Fish. Res. Bd. Can., 36,
    901.
54
             Water Environment Research, Volume 68. Number 1

-------

-------
    Interactions between Trace
    Metals  and Aquatic Organisms:
    A Critique of the  Free-ion
    Activity  Model
    Peter G. C Campbell
    MRS-Eau. Quibee. Canada
    1  Introduction	.'.	45
       U  Scope	45
       l!l  Metal^rgarism interactions (Notions)	46
    2  Derivation of the Free-ion Activity Model (FIAM)	48
       2.1  Historical Development	48
       2.2  Formulation 6f the Free-ion Activity Mode!	 51
    3  Critical Review of the Literature	56
       3.1  Results Conforming to die Free-ion Activity Model	;	56
       3.2  Apparent Exceptions to the Free-ion Activity Model
           (Defined Media)	62
           3.2.1  Upophffic Complexes	.....62
           3.2.2  Inorganic Ligands	65
           3.2.3  Defined Organic Ugands Forming HydrophBic Metal
               .Complexes	-	71
           3.2.4  Problematic Examples	 7?
       3J  Tests of the Free-ion Activity Model in Systems Containing
           Dissolved Organic Matter	—	79
           3.3.1  Examples Conforming to the Free-ion Activity Model	79
           332  Examples of Enhanced Toricily in the Presence of
                Dissolved Organic Matter	 82
           3.3.3  Examples of Enhanced Protection b the Presence of
                Dissolved Organic Matter	 90
    4  Coodusons	 91
       Acknowledgment	,	95
     x  Glossary	 95
       References.......'.........	....;.....	 97
1  INTRODUCTION

I.I   SCOPE

In considering the interactions of trace metals with aquatic biota, one can
identify three  levels  of concern:  (I) metal  speciation  in  the external
environment; (2) metal interactions with the biological membrane separating

Uelal tecfelfan out KetnaOaNSly ft 4
-------
                                                                     ti.ll
                                                                 it»*ft§|U
               Bff;
                  i k*4-nte constants for
dissociation of the surface complex; kj,,- rate constant for 'intemalization' or transport of the metal across the biological membrane.
Charges on Bgand not shown for simplicity. (Modified from Tessier. A. el al, in ChtmlcalanJ Biological ReguIatloH of Aquatic Systems
(Boca Raton: Lewis, 1994) Chap. 6J                                .                             •

-------
    r*'
    M ** 5* » »
 mil!!!
          !£*'
                        »*•«».
                lr*||*B:[f1
                M*filiftHW
                li'VfliHt1'!-
                Illloll^ll3'!-
                g»o g H 2j ff. jo A g Q. < M
                — *< ^^ O**** 31^5 A, IB Ct
                    l3'«1H^S8'<
                5S3
                —"«. < 5" ° " ^ " JS » JR. g
                i!
osli|l:U-e.
  q £ 8 -3 3 ? 2, S.
•Ill
            - s- 3 o-
            »" *K
               a
IK ^rfi^-F
o
                               p
                               n

-------
a-™ilp
                    II (doubling***')

                    £   S
» « 5 a § a^^a **»19**
Jt^lKisirtli
•2S.fi««P.««f a-ftS^.
 S s 3 f s 15
 0.< >1 S3
S'8 3 •9'Rs.sr
ilrii^l
                                             I
                                             I
                                I
                               ceH yield (cells^
                                            p
                                            n
       trt»rj^

-------
            I*



         S* 14

         'a

          t'  12



            10
          is e


         t '
                              .--0-"
                2  4  6  9 10  12  14
Figure 4. Variation in adsorbed Zn (O) and cellular Zn (Q) in two green algae as a function of the frce-Zn concentration, [Zn1* J, for

(A) Otlamydomonas wtabilis and (B) Sctnedamta subsplcahu. Incubation for lOmin in the presence of a fixed total Za concentration

    > 15.3 jiM) and different EDTA concentrations. [Modified from Bates, S. S. tt el.. 1 fhytol. 18.521 (1982)]
        %
  IE  s


  U
  T  I
     I
   = -=.
   S +
   r r

           *:i
           tl
     Slips
     a a aS «» s.
          ?!£
    S.2:l§i1s

    \\ll\\\  ?
  -3&L &
  311 1^
  r*a x^
 br^S

ItlilHN.
  ? s.-"
  j'§j?;
  IAS
 ^ 5 »  xIs
 g.a g.  r|J


ia|

 O O.
           E   ^" 8 °
                       3.8-
                 ss
                                             3
 in  r-      *:n
!|us  Sj<       iV
I?S^  Tl      if
                            2
                            r
                                     ss

-------
                                  2 .1
rl2'
5 S"-.
 Pill fl
     ?|"*?*: o
     as S-  -a  § '"•*

     If   ,»E '
         r^B
              y  =.
         »f
         I:
•s.«  e; s.« f I
PV  Ftil 8f
(•'i
                                      1

                                      I

                                      ^
                                      9

                                      ^
                         X
                                    3
                                     3
                           I
                           jr
                           I
;i»r|
M X n 3 O.
                       's^^l
                        §.a s
                        2-5 P-i
                   S§3-o&i^o
                   £|oall^l
                                            II

-------
       o- a 2
       5'5 S 3
       a. g. S §   .
       ~ *• — e> 9*

       3-S^?
                               5 =• P « KT M-Q g 9 a. a.-*" 2
                               o S=-o o ft,* a» o » i*  .
                               ag-SS5|§sfS|o8 J

                               :!0&>?hHliiiA
                .
TtMe t. Examples of metaJ-organism interaction* that conform to the FIAM
Species
Marine Algie
1. Thalauiosira
wttssflofii
(diatom)
2. T.weissflogii
(diatom) > .
T. oceanka
(diatom) •
T.pKudonana
(diatom)
Emiliana haleyt
(coccotithophore)
3. 7". ptfudonana
(diatom)
NanocUoris otormu •
(cUorophyte)
4. Coityaulax tamartiu'u
(dinoflageOate)
S. r. wtUtfogH
(diatom)
6. T. oceanlca
(diatom)
T.pxudonana
(diatom)
7. Dilytum brigktwtttii
(diatom)
Metal (pM'*>

Zn
(12.5-11)

Zn
(12-«.5)


"



Ot
(11X.2)


Cu
(13.7-9)
Cd
(20-7)
Mo
(9.4-7)

Cd.Cu.Pb.
Zo
Medium

Artificial seawater
(AQUID + EDTA

Filtered seawater
-EDTAorNTA






Filtered seawater
+ EDTA.TRIS

,
Artificial seawater
(AQU1L)
Artificial seawater
(AQUID+EDTA
Filtered seawater
+ EDTA

Seawater

Exposure time

3d



2dor
12 doublings





2-5d



2. 24. 48 h

Variable <4h. -5d


2dor
7 doublings
6d

Response.

Growth stimulation
f/mpZn'*)

Growth stimulation
GI v$ pZn1*)
Uptake Zn
OoglZn/C-orglrfWpZa1*)
/

"

Growth inhibition
OivspCu1*)
Uptake Cu
OoglCol^vspCu1*)
MotiUty
(% motile cells wpCu'*)
Growth inhibition
Uptake Cd. Fe
Growth stimulation
OmpMn1*)
Log [MnJa, vs pMn1*
Log uptake rate « pMn2*
Growth inhibition

Ref

12


37







38



39

24,

40


41


-------
TaUe I. (continued)
Species
Marine Bacteria
1. Marine isolate (Gram
negative)
2. Open ocean microbial
community
3. Estuarine microbial
community
Marine Invertebrates
I. Paleomonftetfugla
(grass shrimp)

2. Cratuatrta ilrginka
(oyster)
}.tt«aakes
arenaetodeittata
(aoneUd)
4. C. tirginlca
(oyster)

5. Kiithropanopeut .
harruS
(crab larvae)
6. X. harrisii
(crab larvae)


7. Aritmia sp.
(brine shrimp)

8. A.franciicana
(brine shrimp)
Freshwater Algae
' 1. CUorellanlgaris
(chlorophyte}
Ootyslis maruonii
(chlorophyte)
2. C. nitgaris
(chlorophyte)
3. Seinedtfrmu
quadricauda
(chlorophyte)

4. S, quodricauda
(chlorophyte)
Ankutrodesrma
falcaius (chlorophyte)
5. Chlomydomonos
utriabUis (chlorophyte)
Seeaedesnaa
tubspieattu
(chlorophyte)
6. C nriatilit
(chlorophyte)




(chlorophyte)


8. CUamydomonas
naAardii
(chlorophyte) . ;
Metal (pM")

Cu
Cn
(12-1.5)
Co
(12-*)

. Cd
C7-2-5)

Cd
(7.4-5.9)
Cd
'
Co
(11-8.7)

Cu

(13.44.7)
Cu
(12.4-7.9)


Cu
P-7.6)

Cu
(6.5-U)

Cu
(16.7-15.5)


Cn
(12.2-10.3)
Ctt
(12.8-6.8)
Zn
(10-5.4)
Ni
(8.5-5)


Zn



Zn
Mn
(6-4.3)
Fe
(7-5.7)
Cd
(6.3-2.9)
Cu
(10.4-4.6)
Cu
(8.2-6)

Medium

UV-trcaled filtered
seawater + NTA
Seawater +NTA
Estuarine water +NTA


Diluted sea water*
NTA
Variable KTA or O-
DOuted seawater
Variable Q-
Filtered seawater

Artificial seawater. or
filtered seawater.
+NTA \
Diluted seawater
+NTA

Sea water* NTA


Artificial seawater
+ EDTA, titrate, or
acetate
Artificial seawater
Variable pH (6-8)

Inorganic growth
medium* EOT A


Inorganic growth
medium* NTA. MES
Inorganic growth
medium (FRAQUIL)


Inorganic growth
medium (AAP)


Inorganic medium
(AAP-FeJ + EDTA



• Inorganic medium '
(AAP-Uace
metals) + EDTA
Variable pH (5-7)


Inorganic medium
Variable pH (5-8)

Inorganic medium

Variable pH (4-9)
Exposure time

2.5 h
1-3 h


4d


4d
3 weeks

I4d

-—


• —


90 min

3h


4d


20d
cheraostai
4d



I4d


lOmin



10-20 min



.<
2h


10 min


Response

Incorporation |4C-g!uco$e
(uptake >*C w pCu**)
Incorporation 'H-amino acids
(% control vs pOr1*)
Incorporation JH-amino adds
.(turnover time vs pCu1*)

Mortab'ry (% survival vs
pCd1*)
W * i
Uptake Cd (log [Cdl,,vs
Uptake Cd Gog [CdL,v»
pCd1*)
F m
Uptake Cu QCtt^, vs pCu1*)

Uptake Cu (cytosoEc Cn vs
PCu1)
Growth inhibition '
Uptake Cu (cytosolic Cm vs
pCu*')
, Growth inhibition
MortaEry
Uptake Cu (aocumulatioB
ratevsJCu1*!)

Uptake Cw
*

Gro»tb stimulation

.
Growth inhibition
(CeD numbers vs pCu1*) .
Growth inhibition

. ' • * •

Growth inhibition


Uptake Zn
QZn]^ vs [ZnJ*D
/rTui yj |2n'*D


Uptake Zn, MnOD, Fe(«)
[Mb vs [M1*)



Uptake PO^NH,, NO,
Oahibition)
ECMvspH

Adsorption Cu vs [Cu}*J
*

Ref

• 42
43
44


45


46
34.47

48

49


50

v
51

35


52


53
20



54


17



55.56




57
27

58



-------
                 Table 1. (continued)
Species
Freshwiter Inrertcbrates
1. Daphnia magna
(water bea)

2. Paratya asutraHtnsls
(freshwater shrimp)
3. P. oatraKaals
(freshwater shrimp)

Fish
1. Salmo gairdnerl
(rainbow trout)

2. S. gairdntri
'(rainbow trout)

3. S.gaifdniri '
(rainbow trout)
4. S. galrdnerl
(rainbow trout)

S. S. gairdntrt
(rainbow trout)
6. S.fabdntri
(perfused gills)


Metal (pM1*)

Cu
(9-6.3)

Cn

Cu



Cu


Cu


Cd

Zn


Cd.Cu.Zn.
Cd
(8.1-5)


Medhun

Lake Superior
water* additions
(PO^HOy
Filtered, aged tap
water +NTA
Filtered, aged tap
water* HCO,. or
ci-

Oiluted groundwater.
variable pH (6-*),
variable hardness
Diluted groundwater,
variable pH(5-9X
variable hardness
Unspecified
variable hardness
Synthetic medium
CaCl,,Mgah
NaHCO,
Diluted groundwater,
variable pH (4.7-7)
Synthetic medium
+ EDTA.
variable pH (5-7), or
variable [a1*)
Exposure time

3d


6d

6d



30d


4d


2d

4-5 d


4-7 d
Ih

• < •

Response

Mortality (I/survival time vs
{Cu1*^

Mortality (LCjrfCu1*) vs Ctay
for varying (LJ)
Mortality (LC^Cu'*) vs Cut
for varying [LQ


Mortality, critical swimming
performance (multiple
regression mortality vs
Mortality (multiple regresstoa
moruKty vs Cu1*, CuOH*)

Mortality (LCj^Cu2*) vs
hardness)
Mortality (LCJ^Zn1*) vs pH)

^
Mortality (LCM (M1*) vs pH)
Uptake Cd (log flux vs
pCd1*.)
Retention Cd (log [Cd^, vs
pCd**)
Ref

59.60


61

62



63


64


65

66


67
68.69



                                      s

       ?55Sl-
3 E
 UKUK

 nmi&
 |f£^y]

 5 H m^ m A

 If^lim
 ""75ft"

 1>-S
 i^?aq


 sl55l*M
 «r c 3. 5 5"
 g " E^ « n.
' SJ Q S' O 5" .g •*
 cUItitr
       •^11?
       •T>| i
      5 —' § <» Si"-!
      I a^wS eS.l-

      iUlH^-

      ?'•«=«! IS 5-
       *3*|i?
        >3'p§-S
      a-lS^S-^S.
     i-2,^^fij|

          *15
           t§?
                            a^^08n

                            Hl£*S:
                            SS se 5>,
                            n? -sgy-|.ff»§a§

                            vff »pB^a''MR2'
                            fiSSMSi^Sl'iS
                            *^ O S **» ^ ^ •* dk _ S O 3* ***
                            vSaasi -a a* x —
                            5B?**.»*^lSia«-
i W8 g St-S 523^3 |-»|
ivSaasi  -a a* x —

HirffiflW
sli^l  llslllll

-------
    im'iismi;
       O 9-
.o ^.
         o
         o
                      Bioaccumulatlon ratfo
                      S   S   S

-------
              TaWe 2. Apparent exception! to the FIAM—ligand forming BpophiDc organic complexes
  Species
                     Metal
                                  Medium
                    Exposure
                      time
           Response
                                                                                   Ref
I. Nituchla closltrban
  (marine diatom)

2. JTujIauloslna
  wtiuflogi
  diatom)
                      Co    Filtered seawater+6xine,    3d    Growth inhibition
3. Daptatlamafna
  (freshwater dadoceran)
4. AOonheOtt eompreua    Cta
  (marine amphipod)
                              neocuproine, or
                              ethykanthate
                      Cu    Filtered UV-treated
                    (23-9 J)    sterilized seawater
                      Cd      +6xine, DDC,
                    (16-10.8)    orS-hydroxy.
                              S-sulToDatoquiaolinc
Lake Superior
  water * DDC
Seawater+orine, or
  ethylxanthale
                      Cd
                              (compared to inorganic
                              Cu)
                     0.5-5h  Metal uptake
                              (compared to
                              inorpnicM)
2d    Uptake Cd (compared to
        inorganic Cd)
4d    Mortality (LCjo)
     .   (compared to inorganic
        C«)
                               75.76


                                107
70

77
S. Safrno gotrdnerl
(rainbow trout;
perfused g3I)
& Sabno gatrdnerl,
(rainbow trout)
Pluxlmu ptoxlmu
(minnow)
Cd
'Cd
Synthetic medium +• DDC.
eihybanthate, or
bopropyUanthate
Synthetic medium +
ethytxanthaie, or
amylxanthate
Ih
3h
Uptake Cd (log flux Cd vs
pCd^foglCdl^Vf
pCd1*)
Uptake Cd (giUs. kidney)
92
71

                                                 i3fllJt't-J  -
                                                 IliiSfUsi!  •
                                                       5*tis


-------
gill Cd (10'"i
log partition coefficient

 &,&i\>.Lo-*igc
                                                     I-
                                                        % mortality (4d)

                                                         i.i*   8  8  I
                                                               h~» lt-«-i
                                                                  *   M

-------
                         TaMc 3. Apparent exceptions to the FIAM—inorganic ligandi
Species
                      Metal
Medium
                            Exposure
                             Time
Response
                                                 Ref
I. Palaemonttet puflo
  (marine grass shrimp)
2. SabnoaloT
  (Atlantic salmon)
 Ag   Diluted jeawateriNTA;
(7.2-5)    variable NTA or G~
 Al   Synthetic soft water +F
                4d

                4d
Uptake At
  pAgC*)
Mortality (% surova! vs
          v$
46

n
                                                                         accumulated Ag (mo1»g-I)

                                                                                  54     5
                                                                          -*•-     i      i  »^i
                                                                                                   p
                                                                                                   n
                                                            K2£a5
                                                           till!   1-
                                                            >ya* •    n
                                                                       $
                                                                                            o


-------
       2«-«i"°53H

       wK&
       Hl:;[i*!
                                    'I:
                                            psrcenlmortaTity
                                                          i

                                                         I
                                                         m
                                  * -i
H
5-

-------
    Table 4. Apparent exceptions to die HAM—low molecular weight ligands forming hydrophilic metal complexes
Species
1. Daptuila magna
(freshwater
dadoceran)
Poeefla rtticulata
(guppy)

2. Parotya austratitruii
(freshwater shrimp)

3. Seltnastnm
caprtcomutum
(freshwater
chlonphyte)
4. Salna falrtfari
(rainbow trout;
perfused gill)
Metal (pM")
Cu
(10.3-8 J)

Cn
(8-7.4)

Cu
(7.1-45)

Co
(11.5-7.8)


Cd
(8-5)

Medium
Synthetic medium
+g)ycine,0-alanine.or
glutamicacid



Filtered, aged lap
water tglyciae

Synthetic medium
+citrate, or
cthylenedianine

. Synthetic medium
+cilrate

Exposure
Time
24

44


64
'
14



Ih


Response Ref
Mob1ity(ECx(Cttl*)in 86
presence or absence of
Egands)
Mortality (LCM (Or*) in
presence or absence of
Egands)
Mortality (LCj,(Cii»*) in 61
presence or absence of
• glycine)
Growth inhibition 87
0»|>.invs pCu1*)


Uptake Cd (log flux C4 68
vjpCd'MoitCdl^vi
pCd1*)
          S3-3

       aSrH
       SS^-g 2,
S. » « x « -g 3  2. 3f o 2

ifill.?  fill
h?»*2t  sffh1
?»-.!
SS,e7l
rfttt
o •=.
li
                                        calculated lCu
                                                a
                                                               p
                                                               o
                              o  o o o js  p

                              O  i» M » *  W
                                                 g g e  e e e
                                §
                              O,

                              V
                                          OB

-------
           sr«   . *    .



            "    ft




Sft<3-S.
•< SR 3 2T
     Ss"f; I1
     & o •* —>
 ts s fa XS- s-=i

 t^llil |?i
s-1. s-? s asr Ri
13
                 a s » a i
     i~s^sf"|ac.  aa.   "&L
     d?|a..*|Ri'Sap8&-ios»i


     '•|!»5littiliBJ:|

                     log flux Cd (nmokh*1)
                                     2
                                     I

-------
-
                               p
                               f>

-------
                     TaMe & Apparent exceptions to the FIAM—unexplained examples
Species
                    Metal
Medium
Exposure Time
Response
Ref
I. Aiubaeaasp.
  (cyanopbyte)* \

Z Mytiha edulit
  (marine mussel^
                  Qi.Cd.Pb Synthetic medium
                             (jnchidinf
                             citrate)* OTA
                     Cd    Filtered seawater  .
                             +'excess'EDTA,
                             bumic add, atginate,
                             orpecute
                  20d      Growth inhibition       91
                  21 d       UplakeCd            90
•Questioublr npcrincnUl dcap. uUiaf I :t (!) XTA:M eoapkm 11 different coooninfjoai: lUlc Ihil NTA Mid act itdoee. bat IB woe C*MS ii
ae tpcciattoo calcutitioBt or mctsurcmcau
•Cd MCwwdn'M loemued (boot twofoU ia itw jmnee ofiO four Dpadi; no tijaift
                                                                              icul toxScil/;
«r Ct^ecutc; Gttad eMwtatntiow *it no» ipccilat; oo ipecUtioo akuktiotu w neuwcmnu.
                                                  t beiw«a UK ttpeuioa ioet for Cd-EOTA, Cd-tlfMU. Cd-h«ro»t«.
                                       M     .-
                             HI!
                                                   •it'r»9ll<  §
                                                       |!P^!i||s
                                                       ^ S-I.M,. K-Sg-8- S^
                                                     &2||9,8||.JS|
                                                     g->a 5 a§ 3 » _.»  « «
                                                    I .sr^iis|i ?|
                                                     ^|S|gH-|8
                                                     i;it§8:ir-
                                                   nils I
                                                                           w

-------
                                                         S
  Table 7. Meul KoawiTability In the presence of natural organic matter- examples of experiment! conforming to the F1AM
Species
t. Monochrysls kiherl
(euryhaiine alga) .




2. Marine bacterium
(Gnm-oegative
motflerod)



3. Daphnlamagna
(freshwater
cladoceran)
.



Metal (pM^)
Cn
ff.9-4.9)




Cu
(10.4-7.4)




Cu
(10.3-8-6)





Medium
3% filtered seawater
:.+ filtered river water
(10.30.90%).
+ distilled water (85.
65, 3%), +TRIS
(ImM)
5% filtered seawater
+ filtered river water
(0,20, 95%), or 3%
filtered seawater
^commercial humic
acid (4.2. 21 ing L-'Q
Synthetic medium
Oncluding
EDTA)+DOC(0-
6rogL-'C).
+ phosphate or
HEPES buffer
pH 7.0. 7.5. 8.0
Exposure Tune Response Ref.
3d Growth Own 96
pCu1*)




2.5 h Incorporation I4C- 42
glucose (cpm vs
pCu")



. 4d Mortality 93
(LC,,(Cu")vs
[DOCD




                         .^* _ _ *•
      l;

msha si: E?i-*8*;i*if •???!
«••§•••» 3 S o B 5-• £ =. f- 3-5 SS'S-rt^^^Ss^
*£>£ 1*1,18-8 g|i  «*p**Plriir*«
       i ^. e*. 9 = O
     iiqilll  if.
    *l!!!flH  5^»*
§ |- s a « 3

 •^l^iiSffl
5?l »-i  ?-
3 Sl a *  a _
"-•*?l ll*6

-------
        g   .irflll*tg
            1 Q § irl£l.fa
            I 3 5-» **£•£§:
                        D.

                        &
?-"fi?
  i g-a s
     »*
     !*
   5tf
   Si
                                          "c-gfucose uptake (cpm)
                  §S .H I |-."e ^.

                  e.B'5-^8 8 s-§
                                                          I
                                                         i
                                                         5:

                                                         (A
'I
  i.
3 5 A  •». 1*  rf "* 2t C  •• S V»



:§f||fsi|lp|f^


1 S.|.|a-P||gf ^l^lS
a * »s g a«, g-a. _ 15-s-§ § g

      5 8 ?• •• s

E8,l*|sr



 >^!l|r|.«iB
 'Doftsro «»-°-
 V |Vg|s-
 •5 r* a S' S «




                                          § I
f-r-
• . i
/ \
t \
T r
i /
L*-H V
t i 1


-



.
i t
                                                     i
                                                        €5

-------
       Table 8. Metal bioavailability in the presence of natural organk matter—exceptions to the HAM
Specie*
Enhanced toxSdty
1. Simoteptutus
ternilatut
(freshwater
dadoceran)
1 Daphnta magna
(freshwater
cladocenn) .
3. Stltnastrwn
eaprleonaitwn
(freshwater
chlorophyte)
Metal (pM1*) Medium

Cu (9.0-6.8); Artesian well water
ISE + 10% filtered pond
water; filtered pond
water
. Ctt Synthetic medium or lake
(10.1-7.1) water. +TRIS
ISE
Cd - Filtered lake water (Lake
(7-5.6) Bakketjcro)+DutiienU


Exposure Time

2d
(LC,o)
6h
(uptake)
3d


4h



Response

Mortality
(LC3c(Cu**)in
diflerent media);
uptake Cu
Inunobitizatioa
(ECj,,(Cu>*)iii
dirTereat media)
Uptake I4CO,
(94 inhibition vs
pcd»*)

Ref.

93 -



97


98



Enhanced protection
4. Paratya ausiroliensls
 (freshwater shrimp)
Oi    Filtered DOM-rich
      natural waters, diluted
      with aged Up water
      and dhtflkd water
6d     Mortality (LCjotCu1*
       in different media)
                               ii(Kit.eti8?iii!fiii
                               « S I *12.1 .*S-f 5 £ * a F I^ B i
                               rlSft'tl  '.ftftat.l&^ft*

    I             HI111P
                                                I?
                                                p
                                                p
                  frfoslfSf^sRs^sl |S£|
                  HMUimrtHKi tl*l

-------
     1-11 SH^
    .1
 S°
 1:8'

* f g
a* S »
iff - UlHll*h M *f!"*P*Ii
 C 3 D rt C fi •  »<£; 8 S* K ^ ll ™  • »»O^O .T a n O ^» S
3 g  1-3.3 S" °  S'B »2 2.

i:  rlfilj  rfUl-i
                        SSg'I'l 2l
                        B 3 ff1! §= 3'
                        "II" sif'-e
                        lift-1HMI:
                        •  "'•"  sgfrf

                               IPII
                               I'iC
                               IHS-i
                               riKt
                                c»
                                  per cent inhibition

                                  S S $ 8  -
                                 ^ """I"""111111- i.""" ~j~ * |~
                           •„/
                                  /
                                            .X
                                      s:
                                      I!
                                      I
                                      Cj
                                      tit

                                      I
                                      1
                                     .a
            &§?«
f              *is
             i|f*
        III jtlli!*
                    >
                    «9
                    o
                    z

-------
                                    IjTirnggf
                              ** »• 5 a  a S.-I a 3 a 5 3£• *
                               8° 3 a OS* o * if 3 & 3 * D
                               g.55-0*0 I * 3	"
                              s a •„..•» 2 &o °-
                     :l!
       ri?!ri
       IP"
        o
!9ll *.
II If.
I S! I
 3
^sSi"

:i«S.
    ?l
 <
a. a
ll
I si:


liH!"


S* " 9 fc* &• o o 2 S" 
                               lllflafglsl'l
per cent survival

     8     1
      per cent survival
              per cent survival   o
              5
              §
<
 S
. ^ 	 f*'1
( ^ » *fr
»*?" i

r 1
0
•I
• i
V,-
» %
i^*
v'
CD
§ ^
i i i i i \f*''tj**^ i
a
o
f"
*"





in

-------
         ££15513
     Sf
              II I
              o s
         ?s>« eg (iKiifti**
             If If <^ -S S ••• *
H-H- H- If If" w

r*r* p.— P —'
*.w «~ U 5
I8|ei|5fs|||£
as ° 5r is *ta SB.
     1


     i

     z

     51
     w
0-
        rf S*» i
        ffg?8 ?

        l8.?!j? 3

LI |f I ;  rflil'*

i|*is"6  SSlffff ••

                   If 11
                   5§
                §— „ * S'-jj
                 ' ft « H*


          ^iifli1  "


           lltll'Eili
           IL?s|Si3-33-"

           I  ?a«l^

           •*.ffga*?5»
           I  »iSw?&»
                       ft a IB.5 a B.

-------
 §'2,g>-.3
 :|a g-3 S
 G- n 0 ~. --
I
                                                    - Q
                                        ^IH^nH^ri
                                        r.llli£lSli.3| ;f

                                                           p
                                                           n
                                                           3
                                                       sr °  s
                                                       O -n  <••

-------
lift
ill*
V*ay
S" ~~

                                 *I*J
                                 e 5 MM

                                 Big'*
                 'l!3o
Q&.g5»I •*."&»• «•* 3 3^*5

|i8i|«i;i|8;i^if|?|l
I-5--if;^jr||s§a-|«||2,r|
•P s»?a a I * * •" i 3 5 g a § a ?&« g»
           » B ^..
          •••ff IS

       i"B§|3lf-
       M^u,,
       •l^tfl^iil1
       irftriiiii
       iHimuir
 -Kg
 2,'^E
ifl'w
'^Vgflllfl^
s-ss s.S.«5-s-i-r»
                 T*

                 P

                 n

-------
  p
  h
  9

  I
  P
•o
-4

-------
                                          I
                                          i
                             !t"  *" * J» *»*' ™ iS

                             •5i  tit* it ?H
9 f  a p  s  9






pi^HlW^51"1
rl^f |P p]
               S  ft
ft *
               «l^tHliI-*f? rlh
8i ^^^f^S. 3--?
'ti ^ti»d i^
ii i?2M! $?

-------
   ?J  i*  ? 2 S  8  3

         LZl&;in*?tfrr&e*fff'$9lg
         iff£K£?*§ffri'*Sl8if»i'il*
                   o Ot o.  s»  o>  O«
                   s> s* .»•  :-  h»  r-


                              Ol
                                 2
 wit mim         B»B

                               g
4fU&UMk rt faifffiili* it fill
f B e !"5 S vS>fi e 2"?>£ " £.!* •px>^0wH3&u73« £ SL '!& -a
ill.*.ttSr^saaS-Uri ix SrSsrllr-g^?!?- 3^ 8-** 1&>
             ir l-gal'gs
             ^'O 3J»rftS|

             r«. SP frfr
           ! j! I 111 111
[gS-G-|aS5|5| Z> l| a.| 5^?- f| 1^ p 5- >|
['I»%N •*ll-25l l| S?, |^S?o s,| SH !'H IL
III iff.s iKf1* !l 15. IKltl ?5 h 'h 11
   5j|* i

             :
,i{lljfMif4!l r


?FlIl|* -*--"
    g -• tf~7ZS'A.& SSS
                    sSiRfl

                       *l U--,"

                   ^ PII If
                   T^Ss- 55 "^ 5.'
                        * 5f -fj
               i fii ill I PI ill I
                                           8
                                           H
                                           I


                                           I
                                           w
                                           25
                                           «l
 .  i-
               ^
               pj
                 • is p          IS w






                  Hl|!!!l! riitT^r^,
                     iii illi! Ill II S8i lot
!-l
                                     rtHH
                                    .^!i 11
                                         sf
                                           P

                                           P
                        0^3 S

-------
 102 METAL SPECIATION AND B10AVAILABIUTY IN AQUATIC SYSTEMS

 92. Block. M, and Part. P. (1986). Increased availability of cadmium to perfused
     rainbow trout (Salmo galrdneri, Kick.) giOs in the presence of the complexing
     agents diethyldithjocarbamate, ethyl xanthatc aod isopropyi xaathate.
 93. Gtesy, J. P, Newell, A, tod Levenec, G. J. (1981). Copper speciation io toft. acid.
     faumic waters: effects oa copper bioaocumulation by and toxicity to Simocephaku
 94. Daly.H. R, Jones, M.J, Hart, B.T, and Carnpbcfl. I. C (1990). Copper toxidty
     Io Pantya etatroKeuit. Ilf. Influence of dissolved organic matter, Environ.
     Toxtcoi.Chtm..9,\m.
 95. Meador, J. P. (1991). The interaction of pH. dissolved organic carbon, aod total
     copper in the determ-aatioo of look copper and toxicity. Aguat. Toxicol^ 19. 13.
 96. Sunda, W. On and Lewis, J. A. M. (1978). Effect of complexalioo by natural
     oigaak ligands on die toxidy of copper to « un^
     Um»ot.Oceanogr.,13,tTQ.
 97. Borgmaoo, U. and Charlton. C C. (1984). Copper complexation and toxicity to
     Daphntola natural «iten. /. Grtal Lakes Run 10, 393.
 98. Laegreid. M, Abtad, 1^ Klaveoess. D, and Scip, H. M. (1983). Seasonal variation
     of cadmium toxicity toward the a{ga Selenasinan caprlcomutum Printz in two lakes
     with different humus conteot, Gniron. Scl. Teehnot^ 17, 357.
 99. Gi«y, J. P, Levenee, G. J, and WiUiams, D. R. (1977). Effects of naturally
     occurring aquatic  organic fractions on cadmium toxicity  to  Simoctfltalut
     temtlatut (Daphnidae) and Gamhoia affinis (Poecfliidae), Water Ret^ II, 1013.
100. Winner, R. W. (1984). The toxicity and bioaccumulation of cadmium and copper
     as affected by hiimic acid. Aqua.  Toxteot.. $, 267.
101. Winner, R. W. (I9S6). Interactive effects of water hardness and humic acid on die
     chronic toxicity of cadmium to Daphnla pulex, Aquai. Toxioot., 9, 281 .
102. Winner, R. W, and Gauss, I. D. (1986). Relationship between chronic toxicity and
     bioaccumulation of copper, cadmium and zinc as affected by water hardness and
     hunuc acid, Aquat. Ttxieol^ 1, 149.
103. Buffle. ^ Altmann. R. S, Rkfla. M, and Tessier. A. <1990). Complexadon .by
     natural heterogeneous compounds: site  occupation distribution functions, a :
     normalized description of metal complexation, Geochim. Cosmochlm. Aaa. 54.
     1535.
104. Myen, V. B.. Ivenon, R. L, and Harriss, R. C (1975). The effect of salinity and
     dissolved  organic matter on surface charge characteristic of some euryhalinc
     phytoplankion, /, Exp. Mar. Blot. En*. 17, 59.
105. Prakash, A, and Rashid, M. A. (1968).  Influence of humic substances on the
     growth of marine phyiophnkton. DinoOageOates. Umnol. Oceanogr^ 13. 598.
106. Hargeby. A, and Petetsen, R. C (1988). Effects of low pH and humus on the
     survivorship, growth aod feeding ot Gammons pulex (L.) (Amphipoda). Fresfaat.
     to.. 19,235.
107. Phinney, J.T.andBmJaDd,K. W.(I994). Uptake of Bpophilic organic Cu.Cd and
     Pb complexes in the coastal diatom Tbalautosira wtiujiogtl, Environ. Set. Ttchnot^
     28.1781.

-------
          Pcrgamon
Environmental Toxicology and Chemistry, Vol. 14, No. tl, pp. 182(7-1858. 1995
                                      Copyright © 1995 SETAC
                                 '•  .       Printed in the USA
                                       0730^7268/95 19.50 + .00
                                              0730-7268(95)00146-8
             MODELING SILVER BINDING TO GILLS OF RAINBOW TROUT
                                    (ONCORH YNCHUS MYKISS)

                                   NANCY JANES and RICHARD C. PLAYUJ*
                    Department of Biology, WUfrid Laurier University, Waterloo, Ontario, N2L 3C5 Canada

                                  (Received 9 February 1995; Accepted 27 April 1995)


      Abstract-Rainbow trout (Oncorhynchusmykiss, 1-3 g) were exposed to -0.1 jiM diver (Ag) (~11 pg-L"1 Ag) for 2 to 3 b
      in synthetic soft water (Ca, Na -300 jiM, pH 6.5-7.5) to which was added Ca, Na, H*. dissolved organic carbon (DOC), a,
      or thiosulfate (SjOj). Gills were extracted and gfll Ag concentrations were measured using graphite-furnace atomic absorption
      spectrophotometry. The concentrations of cations (Ca, Na, H*) and completing agents (DOC, Ct, SjOj) needed to keep Ag
      off the gills were used to calculate conditional equilibrium binding constants (K) at the gills. Log K for Ag-gill binding was
      10.0, with approximately 1.3 nmol Ag binding sites per fish. All experimentally determined log AT values were entered into an
      aquatic chemistry equilibrium model, MINBQL*. to predict Ag binding at trout gills. For a series of natural waters, model-
      predicted gDl Ag concentrations correlated wdl with observed gOl Ag concentrations, with one exception, very hard city of Waterloo
      tapwater. This exception may indicate a kinetic constraint on the thermodynamic basis of the model.

      Keywords-Fish    Gills    Silver    Binding    Model
                   INTRODUCTION
   Silver (Ag) is one of the most toxic metals to fish, with
96-h LCSOs of free Ag between about 0.05 and 0.30 jiM (5.3-
32 pg-L~' [1-4]). There is a dear influence of water hard-
ness and Ag speciation on Ag toxicity, with less toxicity of
Ag in harder water (e.g., high Ca concentrations [1]) and
when Ag is bound by ligands such as Cl or thiosulfate {3].
Thiosulfate is a fixer used in developing photographic film,
and silver thiosulfate is the form of Ag that has the most po-
tential to enter the environment in photoprocessing effluents.
Although the expense of Ag ensures that efforts are made to
retain it during industrial processes, it has been estimated that
about 25% of the total annual consumption of Ag has the
possibility of entering the environment during product dis-
posal, and that lOVo of this waste enters water as unrecorded
photographic Ag [5]. Although Ag is not often found hi high
concentrations in most natural water sites, heavily contam-
inated sites may have surface-water concentrations as high
as about 0.35 pM [5].
   Waterborne Ag will interact first with the gills of a fish.
Fish gills constitute about half a fish's total surface area and
are a sensitive and critical region of contact with water [6].
Fish gills function as the main location for gas transfer, ni-
trogenous waste excretion, acid-base regulation, and ion up-
take to counter passive diffusive losses of ions [7]- The gill
epithelium contains negatively charged binding sites, which
are due to phosphate, carboxyl, amino, and sulfate groups,
among others [8]. Metals and cations are attracted to these
negative binding sites oh the external surface of the gill epi-
thelium, and competition for gifl binding sites is likely to oc-
cur between Ag and other cations such as Ca. Complexation
   •To whom correspondence may be addressed.
    by ligands such as thiosulfate. dissolved organic carbon, and
    Cl will decrease free Ag available to interact at the binding
    sites. Thus, the influence of water hardness and speciation
    on Ag toxicity to fish, and metal toxicity in general, can be
    explained through competition and complexation [9,10].
       The primary objectives of our study were to determine the
    binding ability of Ag to sites on  fish gills, to determine in-
    teractions of other cations at the Ag binding sites, to de-
    termine interactions of complexing agents with Ag, and to
    model the system for predictive purposes. The method we
    used consisted of estimating conditional equilibrium bind-
    ing constants (K) from results obtained using different li-
    gands and cations, and entering these' binding constants into
    an aquatic chemical equilibrium model, MINEQL+ [11).
    Our method is similar to Ugand-exchange methods used to
    establish conditional stability constants of metal complexes
    of organic ligands in seawater [12,13] and has been used be-
    fore to model Cu and Cd interactions at fish gills [14,15].
                 MATERIALS AND METHODS
    Experimental
       Small rainbow trout (Oncorhynchus mykiss, 1-3 g) were
    purchased from Rainbow Springs Hatchery. Thamesford,
    Ontario. Fish were acclimated to synthetic soft water for at
    least 1 week before each experiment. Soft water was pro-
    duced by reverse osmosis (Culligan Series E reverse-osmosis
    system) and had a composition of 20 to 320 pM Ca. 150 to
    400 pM Na. 100 to 290 pM a, and pH 6.5 to 7.5, depend-
    ing on the efficiency of the reverse-osmosis system. Fish were-
    fed Martin's starter food during acclimation but not during
    the experiments.
       For an experiment, six fish were randomly assigned to
    each aerated polyethylene container containing 1 L of soft
    water, modified by the addition of Ca,  Na, H*, dissolved
                                                      1847

-------
1848
                                            N. JANES AND R.C. PLAYLE
organic carbon (DOC), Cl, thiosulfate, Ag, or Cd (at 12 to
17°C, see below). Water samples were taken for later analy-
sis, and water pH was measured and adjusted to between pH
6.5 and 7.5 with dilute KOH or H2S04 if necessary. Fish
were exposed to the water for 2 h, after which the fish were
randomly sampled, stunned with a blow to the head, and the
gills from both sides extracted using stainless steel forceps.
Extracted gills were rinsed in 100 ml of reverse-osmosis wa-
ter for 10 s, shaken to remove excess water, then placed in
1.5-ml polypropylene microcentrifuge tubes. In a normal ex-
periment, sampling time meant that fish were exposed to their
respective solutions for between 2 and 3 b.
   Wet weight of the extracted gills was measured with a Met-
tler AE163 balance after their transfer to new microcentri-
fuge tubes. Following gill weighing,  1 N HN03  (Ultrex II
Ultapure Baker analyzed reagent) was added to each micro-
centrifuge tube at five times the gin wet weight. Gills plus acid
were heated for 3 h at 80°C. After heating, the digested gills
were agitated, allowed to settle, and  100 /J of supernatant
was added to 900 jd of Milli Q dtionized water (Millipore)
in new microcentrifuge tubes, and left for later analysis.
   Analysis of gill and water Ag concentrations was done
using a graphite-furnace atomic absorption spectrophotom-
eter (AAS; Varian AA-1275 with  OTA-95 atomizer). Ten-
microliter samples were injected, N2 gas was used, and
operating conditions were 5 s at ~75°C, a short (30 s) dry-
ing time at -90°C, 12 s at ~120°C, and 4 s at 2,000°C, dur-
ing which Ag concentrations were read. Water and gill Cd
concentrations were measured using conditions documented
by Varian, but also with a 30-s drying time at ~90°C. Gill
and water Ag and Cd samples were run against standards pre-
pared from Fisher-certified Ag and Cd reference solutions.
   Water samples collected at the beginning and end of the
experiments were analyzed for Na and Ca using  a Perkin-
Elmer model 3100 AAS; LaClj was added to the Ca samples
to reduce Na interference. Water Cl was measured using the
mercuric thiocyanate colorimetric method (Sigma reagents).
Dissolved organic carbon was measured using a Beckman
91SB total carbon analyzer. Water pH was measured with a .
PHM82 Radiometer pH meter and OK2401C combination
electrode.                    .
  • Chemicals used were usually made as 1 mM stock solu-
tions. Silver was added as AgNO3 (AnalaR, BDH, Poole,
England), thiosulfate as NajSjOj-SHjO (AnalaR, BDH), Cl
as KC1, Na as NaOH, H+ as H,SO4, and Ca as Ca(OH)j.
The DOC was isolated from the south side of Luther Marsh
(43'54'N, 80°24; W, near Grand Valley, Ontario) by tangen-
tial flow ultrafiltration (Millipore, Bedford, MA), and was
supplied to us by Kent Buraison, National Water Research
Institute, Environment Canada, Burlington, Ontario.
   A longer-term thiosulfate and Ag experiment was run, in
which 16 to 18 fish per 16-L bucket  were held for a week
hi 13°C aerated soft water. Fish were sampled periodically
for gill Ag, and water was sampled for water ions and Ag
concentrations. Half the water volume was replaced daily,
and each treatment was run in duplicate:
   To test the Ag-gill model, which was based on the syn-
thetic softwater experiments, rainbow trout were exposed
 2 to 3 h in Ag-supptemented water collected from various
 locations in southern Ontario. Fresh water was collected in
 May 1994 from four locations: Lake Ontario at the south pier
 of the entrance to Hamilton Harbour, Hamilton (4>18'N,
 79047-W), "Dundas pond" on the east side of the main access
 road to Dundas Valley Conservation Area off Highway 99
 (43'14'N, 80°00'W), Gait (Mill) Creek above the dams
 at Shades Mills Conservation Area, Cambridge (43°23'N,
 80°15'W), and  Grand River at old highway 8, Kitchener
 (43°25'N, 80°24'W). We also used water from Salmon and
 Jack lakes (44*43'N, 78'03'W, north of Peterborough), which
 had been stored since 1991 [15]. Finally, we used Waterloo
 city tapwater entering our laboratory, passed through acti-
 vated charcoal to remove chlorine.
   In an effort to differentiate colloidal Ca from Ca in solu-
. tion, we filtered three natural waters through 0.45- or 0.2-/tm
 cellulose nitrate filters (Sartoiius). Fifteen milliliters of MiffiQ
 water were passed through each filter under vacuum as a
 rinse, then 15 ml of sample, then another 15 ml of sample,
 which was collected. A Millipore graduated funnel assembly
 was used; all water was at 16"C.
   Statistical analysis of the gill Ag data consisted of un-
 paired t tests or one-way analysis of variance (ANOVA) fol-
 lowed by the Student-Newman-Keuls method of pairwise
 multiple comparisons, as appropriate [16]. Significant dif-
 ferences are reported at the p < 0.05, p < 0.01, and p <
 0.001 levels (*, **, *»•, respectively). Error bars in the fig-
 ures represent the 95% confidence interval (C.I.) about the
 mean gill Ag concentration, usually for six fish.

 Conditional gill stability constants
 and computer modeling

   The main objectives of this study were to determine the
 binding ability of Ag to the gills of rainbow trout, to deter-
 mine the interactions of other cations at Ag binding sites and
 the complexation of Ag by ligands in water, and to model
 these interactions in a computer program. Powerful aquatic
 chemistry programs exist that use thermodynamic stability
 constants; we chose MINEQL* (version 2.0), a chemical
 equilibrium program for personal computers, mainly be-
 cause of its ease of use [11]. Some equilibrium constants from
 MINEQL* are given in the Appendix. Results from the ex-
 posures of fish to Ag were used to calculate stability constants
 for insertion into the MINEQL* program.
   To calculate conditional gill stability constants we consid-
 ered two scenarios. The first scenario is the presence of a sol-
 ute such as Na at a concentration that does not keep Ag off
 the gills (no inhibition of Ag binding; Eqn. 1).
[Na] •A
                  Nm
-------
                                     Modeling silver binding to gills of rainbow trout
                                                  1849
   The second scenario is a solute such as Na at a concen-
 tration that keeps all Ag off the gills (complete inhibition of
 Ag binding; Eqn. 2).
                                                   (2)
Details of the derivation of these two equations can be found
in Playle et al. [IS]; a similar derivation can be found in
Midorikawa and Tanoue [12].  For Equations  1 and 2,
MIN£QL+ was used to calculate free Ag from water chem-
istry and the total concentration of Ag. This calculation takes
into account side reactions involving Ag (e.g. [13]).
   Once conditional stability constants for the gills were es-
timated (see Results), they were entered into the MINEQL+
program. For each simulation the concentrations of Ca, Na,
a, DOC, Ag, and NO3 (=[Ag]) in the test water were en-
tered into the program. Water pH was fixed to the measured
value (U., pH changes at the gill surface were ignored; see
Playle et al. [15]). This approach simplifies the model because
buffering capacity of the water does not need to be deter-
mined, and is appropriate because sensitivity analysis using
MINEQL* indicated that Ag speciation is very insensitive to
changes in water pH. Measured water temperature was also
entered jnto the model, and the system was considered open
to the atmosphere. The concentration of HCOJ" in the water
was assumed to be the same as that of Ca; this simplification
was tested using the more accurate Henderson-Hasselbalch
equation. In any case, sensitivity testing of the model indi-
cated that Ag binding to gills is unaffected by HCOf.

                       RESULTS
   Initial experiments in synthetic soft water showed that Ag
accumulation on gills of rainbow trout was significant at ex-
            0.00   0.04  0.05  0.22  0.39   0.60
                   silver concentration (pM)

Fig. 1. Silver accumulation on gills of 1- to 3-g rainbow trout ex-
posed to 0.04 to 0.50 |tM Ag for 2 to 3 h in synthetic soft water. The
clear bar represents no Ag added (control fish); striped bars repre-
sent Ag added. Asterisks indicate significant differences from con-
trol mean (*p < 0.05, **p < 0.01, ***p < 0.001). Error bars represent
95% C.I. about the means, with six fish per bar.
posures of 0.05 pM and greater (Fig. I; 2- to 3-h exposures,
Ca « 27 pM, Na = 152 jM, Cl = 100 pM, 16°C, pH 6.5).
There was an indication that gill binding sites were saturated
at 0.22 ;tM Ag exposure, because increasing the Ag concen-
tration further did not result in increased gill Ag above about
12 to 15 rimol Ag/g wet tissue (Fig. I). From these results,
we decided to use Ag concentrations of between 0.05 and
0.1 ftM in subsequent exposures.

Thiosulfate experiments
   Rainbow trout exposed to 0.07 ± 0.01 pM Ag (mean ±
so, n - 6) showed increased gill Ag concentrations in the
presence of up to 1.0 jiM thiosulf ate (Fig. 2; 2- to 3-h expo-
sures, Ca = 18 fiM, Na = 174 pM. Cl = 100 pM, 17°C, pH
6.5). It took 2.0 \M of thiosulfate to keep Ag completely off
the gills (no significant difference between these fish and con-
trol fish),  about 29 times more thiosulfate than Ag. There
was an indication that 1.0 pM of thiosulfate reduced gill Ag
to some extent (Fig.  2), but this protective effect was not
strong enough to keep gfll Ag at background concentrations.
   The protective effect of thiosulfate against Ag accumula-
tion on trout gills persisted for nearly 6 d (Fig. 3). Small ram-
bow trout held in synthetic soft water with 0.07 ± 0.03 pM
Ag (mean ± SD, n = 14) plus 2.5 pM thiosulfate showed no
Ag accumulation on their gills until 147 h. In contrast, fish
exposed to 0.06 ± 0.02 iM Ag'(*i = 14) in the absence of thio-
sulfate showed elevated gfll Ag concentrations throughout
the experiment, with particularly high Ag concentrations at
2 to 3 h (Kg. 3; Ca = 100 iM, Na - 340 j.M, Cl« 200 j»M,
13'C, pH 7.3).                                        '

Initial modeling ofAg-gill interactions
using the thiosulfate data
   To model Ag binding at trout gills, we inserted estimates
of Ag-gill binding strength into the MINEQL+ chemical
          0.00  0.00  0.05  0.10  0.60 .1.00  2.00
                  thiosulfate concentration (fiM)

Fig. 2. Silver accumulation on gills of 1- to 3-g rainbow trout ex-
posed to 0.07 jiM Ag and 0.0 to 2.0 pM thiosulfate. It took about
29 times more thiosulfate than Ag to keep Ag off the gills. Bach bar
represents the mean of six fish. See caption of Figure 1 for more
details.

-------
1850
                                            N. JANES AND R.C. PLAYLE
                       14

                       12

                       10

                        8
                    r  4
                    I  2

                        0
50                100

       time (h)
                                                                                   150
Fig. 3. Silver accumutaion on gflls of 1- to 3-g rainbow trout exposed to Ag for 6 d. Fish exposed to 0.06 pM Ag in the absence of thiosulfate
<•) showed significantly elevated gill Ag concentrations (asterisks; / test) compared to fish exposed to 0.07 pM Ag plus 2.5 pM thiosulfate
(V). The cross at 147 h Indicates significant gillAg accumulation (p < 0.05, t test) in the Ag-plus-thiosulfate treatment compared to control fish.
equilibrium program [11]. This program uses equilibrium
constants; some of the relevant constants from the program
are given in log K format in the Appendix. These values are
nearly identical to values used in other chemical programs
such as MINTEQA2/PRODEFA2 [17], and they are close
to the equilibrium constants given in Morel and Hering [10].
   The silver thiosulfate complex (AgSjOf) has a log/:
value of 8.80 (see the Appendix). It took about 29 tunes more
thiosulfate than Ag to keep Ag off trout gills (Fig. 2), so the
log Ag-gill conditional equilibrium constant must be >8.80.
The range of the Ag-giD conditional equilibrium constant can
be estimated using Equations 1 and 2, substituting SjOj for
Na and ^Ag-SjOj for^N«-sfflA«- Using Equation 1 (no protec-
tive effect of 1 jtM thiosulfate), log AA,^^ must be >10.0
(SjO, = 1 nM, log A'A.^O, - 8-8, free Ag = 0.06 /«M).
Using Equation 2 (complete protective effect), log A'ArlmA(
must be < 10.3 (SfeOj = 2 iiM, log AA.-SJO, = 8.8, free Ag =
0.06 >iM). Silver binding sites at the gill were inserted into
the MINEQL* program as "guTAg" (null component), then
defined as a complex with 1:1 Ag:gfllAg binding. From our
thiosulfate data (Fig. 2), we estimated 0.61 nmol binding sites
per fish (6.1 nmol Ag/g wet tissue for 0.07 /iM Ag exposure,
no thiosulfate; average gill basket of our fish -0.1  g) and
entered 0.61E—9 for the molar concentration of the gillAg
component.
   Log A'AnniA, values of between 10.0 and 10.3 were tried,
and the MINEQL'1' program output values were compared
to the thiosulfate results of Figure 2. Best fit to the  experi-
mental data occurred with log AXg^uAt -1°>° and 0.61 nmol
binding sites per fish. Predicted values for these conditions,
using water chemistry available from the experiments (Ca =
18 J.M, Na=174 jiM, Cl «100 pM, pH 6.5,17»Q, are given
in Figure 4. The addition of thiosulfate to the system kept
Ag off the gills only at relatively high thiosulfate: silver mo-
lar ratios. The correlation coefficient between observed and
           MINEQL*-predicted gill Ag (using log A^mA, = 10.0) was
           r = 0.924 (p< 0.01).
              In contrast, free Ag+ concentrations in the water, as cal-
           culated by MINEQL*, decrease rapidly as the concentration
           of thiosulfate is increased. If gill Ag concentrations were di-
           rectly related to free Ag* then observed gill Ag and gill Ag
           predicted from free Ag* concentrations would track one an-
           other, but the fit is poor (Fig. 4) and the correlation coeffi-
           cient is not significant (r=0.525,/» 0.05). In other words,
           although free Ag* is reduced by the addition of thiosulfate
           to the water (log A^^OJ = 8.8; see Appendix), trout gills
           successfully outcompete thiosulfate for Ag (log *A»-IUIA« =
          . 10.0), so gill Ag concentrations do not directly follow free
           Ag* concentrations.

           Complication ofAg and competition
           for Ag binding sites

              Complexation of Ag by Cl and dissolved organic carbon
           (DOC) and competition by Na, Ca, and hydrogen ions (H*)
           for Ag binding sites on the gills were investigated to better
           define and model Ag binding at fish gills. The addition of
           1 i.3 mM of Q kept Ag off the gills during a 0.11 pM Ag ex-
           posure for 2 to 3 h in synthetic soft water, whereas 1.5 mM
           of Cl did not (Fig. 5; Ca = 155 jtM. Na = 396 /»M, 12°C,
           pH 7.5). Chloride forms a number of weak complexes with
           Ag, depending on the ratio of Cl:Ag (see Appendix). Using
          • Equation 1, log A^ was estimated at <5.7 (log^AMOu* =
           10.0, free Ag = 0.07 pM, Cl = 1.5 mM). Using Equation 2,
           log AX,xa was estimated at >4.8  (logA'A,.,WA, = 10.0, free
           Ag=0.07 j
-------
                                      Modeling silver binding to gills of rainbow trout
                                                                     1851
                         10
                      I
                      I
                              0.0
0.5             1.0            1.6

    thiosulfate concentration (pM)
                                                                                            2.0
Fig. 4. Observed and MINEQL+-predkted gill Ag concentrations for the thiosulfate data of Figure 2. Exposure was 0.07 pM Ag and 0 to
2 jiM thiosulfate, in synthetic soft water, for 2 to 3 h. Error bars about the observed 'data (•) are the 95% C.I. Two predictions of gill Ag
concentrations were made using MINEQL*. The first prediction used log ^Af-gUiA. = 10.0 and 0.61 nraol binding sites per fish (D); for clar-
ity, these symbols are shifted to the right in the figure. The second prediction of gill Ag (T) was based directly on free Ag* in the water, as
calculated using MINEQL*. Gill Ag predicted using log X^nuc ~I0 ° was a much better fit to the observed data. The open circle in the
bottom left corner represents background gill Ag. See text for more details.               •
Ag = 0.17 MM, Ca = 315 * 9-0(tog ^A^DIA. =
10.0, free Ag - 0.12 fiM, DOC - 24.2 rug C/L «1.21 pM).
Predicted gill Ag values will be discussed in the next section.
   As an illustration of competition for Ag binding sites at the
                 gills, 16inMNa kept Ag off trout gills, whereas 1.6 mMNa
                 did not (Fig. 7). Exposures were to 0.11 pM Ag in synthetic
                 soft water, for 2 to 3 h (Ca = 155 >iM, Cl = 265 pM, 12°C,
                 pH 7.5). From Equation 1 it was estimated that log XN*SUIAI
                 was <5.7 (log A'A.^MA* = 10-0. free Ag = 0.08 pM, Ha =
                 1.6 mM). From Equation 2, log A'Nl^fmA,> 4.7 OogAXt^nA,=
                 10.0, free Ag - 0.08 jtM, Na = 16.0 mM). Predicted gill Ag
                 will be dealt with in the next section.
                    Competition for Ag binding sites by H* was investigated
                    0.3      0.4       1.6

                   chloride concentration (mM)
   11.3
Fig. 5. Observed and predicted gill Ag concentrations in the pres-
ence of 0.3 to 11.3 mM Cl. The dear bar represents gill Ag concen-
trations from six control fish; striped bars represent fish gills from
O.it jiM Ag exposures (six fish per bar). Error bars are 95% C.I.
**• = significantly greater than control concentrations (p < 0.001).
Predicted gill Ag concentrations (hatched bars) were calculated us-
ing the MINEQL * program, with die insertion of the final param-
eters given in Table 1 (see text for details).
                                                                      2.4     3.0     6.0    7.0    10.8   14.6   24.2

                                                                                 DOC concentration (mo C/U
                 Fig. 6. Observed and predicted gib Ag concentrations in the pres-
                 ence of 3.0 to 24.2 mg C/L DOC. Striped bars represent gjll Ag con-
                 centrations from fish exposed to 0.17 pM Ag, six fish per bar. *, •*.
                 •** = significantly greater than control concentrations (*p < 0.05,
                 •*p < 0.01, ***p < 0.001, respectively). See caption of Figure 5 for
                 more details.

-------
1852
                                            N. JANES AND R.C. PIAYIE
                                                                  Table I. Input data used in the Ag-gill model
                                                          Complex
                      initial log*
Final log A"
            0.4        0.4  .     1.6       16.0
                 sodium concentration (mM)

Fig. 7. Observed and predicted gill Ag concentrations in the pres-
ence of 0.4 to 16.0 mMNa. Striped bar* represent gill Ag concen-
trations from fish exposed to 0.11 jtM Ag, six fish per bar. See
caption of Figure 5 for more details.
using a Ag concentration of 0.06 pM Ag at pH 6.8, 5.5, and
4.5 (2- to 3-h exposure, Ca « 260 /»M, Na = 388 /iM , Cl =
250 pM, 12°Q. There was no inhibition of Ag binding to the
gills (all 6.0 to 6.5 nmol Ag/g wet tissue) at even the lowest
pH used. From Equation 1, the log H+-gillAg conditional
equilibrium constant was estimated at <7.1
10.0, free Ag = 0.04 pM, pH 4.5 = 32 iM H4).
   Concentrations of Ca up to 10.6 mM did not reduce Ag
binding to trout gills (5 to 7 nmol Ag/g wet tissue, 2- to 3-h
exposure to 0.05 ^M Ag, Na = 399 pM, Cl = 265 pM, 12°C,
pH 7.1).  From these results and liquation 1, logA'Ck.f]1|A(
was estimated to be <4.5 (log XAMJIIA* - 10.0, free Ag =
0.03 iM. Ca = 10.6 mM).
Final modeling of Ag-gill interactions

   From the thiosulfate data (Fig. 2), the Cl and DOC com-
plexation data (Figs. 5 and 6), and the Na (Fig. 7), H*, and
Ca competition data, we estimated conditional equilibrium
constants, the number of Ag-gill binding sites, and the num-
ber of Ag binding sites on DOC (summarized in Table 1).
Using the constraints of these initial estimates, we introduced
the log AT Values one by  one into the MINEQL* program
and best-fit predicted gUl Ag concentrations to the observed
values by trial and error.
   The first fitting exercise was already described for thio-
sulfate and is illustrated in Figure 4, and log A^^UA* was
optimized to 10.0 using 0.61 nmol Ag binding sites per fish.
The DOC data were best predicted (Fig. 6) using 35 nmol
binding sites per rug C/L, IoglTAt.Doc of 9.0, and setting
1 .0 nmol binding sites Ag per fish. This higher value for bind-
ing sites was used to scale the predicted values to the observed
values in Figure 6, because the introduction of DOC (back-
ground = 2.4 mg C/L) into the model reduced the amount
of Ag available to bind with the gjllAg sites: The correlation •
Ag-giiJAg :
Na-gillAg
H-gillAg
Ca-gillAg
Ag-DOC
H-DOC
10.0-10.3
4.7-5.7
<7.1
<4.5
9.0-9.2
. 4.0
10.0
4.7
5.9
3.3
9.0
4.0
                                                          Binding sites
                                                            gillAg: 0.6 to 1.9 nmol binding sites per fish (varied for scaling)
                                                            DOC: 33 nmol binding sites per mg C/L (initial = 50)

                                                          Initial conditional equilibrium constants Gog K) were calculated from
                                                          the experimental data using Equations 1 and 2 (see text). Final tog K
                                                          values were chosen by best-fining the initial log K values to the ex-
                                                          perimental data.
of this fit was 0.946 (p < 0.01). Varying the number of gill
binding sites used in the model by a factor of two or three
has little effect on the correlation between observed and pre-
dicted values (varying binding constants has a much greater
effect, because they are log values), and this was done to fit
predicted values to the vertical scale of the figures. We had
no data for H* interactions with DOC, and we used the
value previously used in Playle et al. [15] of logJCH-Doc =
4.0 (Tablet).
   The best fit to the Na data (Fig. 7) was log #N.-»aiAg =
4.7, with the number of Ag-gill binding sites 1.2 nmol per fish
(r = 0.973, p < 0.05). Competition by H+ for Ag binding ,
sites was more difficult to fit, because we did not have a com-
petitive effect at pH 4.5, and to take an experiment with fish
'much below pH 4.5 is unrealistic. We assumed a full com-
petitive effect at pH 3.5, so the final log K value was set to
5.9, with 1.1 nmol Ag-gillAg binding sites per fish. In a sim-
ilar manner, full protective effect of Ca was assumed to occur
at 150 mM Ca (with 1.4 nmol Ag binding sites per fish), and
final log Jfoi^iiiAi was estimated to be 3.3 (Table 1). We
could not physically keep Ca in solution in our system at con-
centrations >11 mM.
   The Cl fitting (Fig. 5) gave an indication of the power of
our modeling approach. Unlike Ag-DOC interactions, for
which there are no equilibrium constants in MINEQL* and
the Na, H*, and Ca competition at Ag binding sites, for
which equilibrium constants were calculated for the first time
by us, the MINEQL'1' program  already contained  Ag-CI
equilibrium constants. Thus, no fitting was required, aside
from setting log XA«.«IIIAS = 10.0 and setting the number of
gill binding sites to 1.25 nmol/fish. In the previous section,
it was estimated from our gill Ag data that log A^-ci,,was
 >4.8 and <5.7. Running the MINEQL* simulation indi-
cated that with 1.5 mM Cl and 0.11 jiM Ag concentra-
tions only AgCI was formed (log K=3.3; see Appendix), and
Ag was still bound to the gills (Fig. 5). At 11.3 mM Cl and
0.11 ^M Ag, MINEQL* calculations indicated that AgClJ
was formed (bg *A*
-------
                                      Modeling silver binding to gills of rainbow trout
                                                                                  1853
                   .4.0'
  Ag—9.0—DOC'
          water
          Ag-8.8-S20,

  AB-3.3-CI

          Afl-5.3-CI,

  Ca-3.2—CO,
                                                                   2.5
H
                                  Na
 Fig. 8. Illustration of the injeractions of the most important param-
 eters io the Ag-gffl model. Complexation of Ag in the water may
 reduce Ag binding to trout gills, depending on the relative concen-
 trations of the Ugands and the relative magnitude of their conditional
 equilibrium constants (loglvalues from Table I are indicated in
 the figure). Competition with Ag at a single Ag-gill binding site
 (asterisk) is also dependent on relative binding strengths and cation
 concentrations.                                 •
 (mean = log 5.5). The Cl data were a relatively independent
 check of the fitting process, and the good fit between pre-
 dicted and observed gill Ag (r = 0.956, p < 0.05) nicely il-
 lustrates the power of the MINEQL* aquatic chemistry
 program.
    A schematic illustration of Ag binding to the gill, cation
 interference with that binding, and Ag and H+ binding to
 DOC is given in Figure 8. The log K values for these reac-
 tions are from Table 1; some of the more important equilib-
 rium constants from MINEQL+ (Appendix) are also given
 in the water portion of the diagram. In essence, complexation
 of Ag (e.g., by DOC) can occur in water, rendering Ag un-
 available to bind at the gills, and cation competition with Ag
 (e.g., by Ca) occurs at the binding sites on the gills themselves.
    Once Ag is on the gills, it can exert a toxic effect at the
 gill itself or after active uptake or passive diffusion into
 the gills, then into the blood of the fish. An indication of
 the toxic effect of Ag is ion loss from rainbow trout into the
 surrounding water (e.g.,  [7]). In two of our experiments
 (Figs. 1 and 2) the concentration of Na in the synthetic soft
 water was low enough (152 and 174 pM, respectively) that
 we were able to measure Na less from the fish to the water
 over 2 h (six 1- to 3-g trout per liter of water). There is an
 Na diffusion gradient from inside the fish into the water
. (plasma Na ~ ISO mM [7]). Sodium loss (net flux per unit -
 weight = Jna [7]) plotted against gill Ag yielded a correlation

                                                                a
                                                                m
                                                                £
                                                                   1.5
                                                                   1.0
                                                                   0.5
                                      0.0
       5          10

gill Ag (nmol Ag/g wet tissue)
                                                                               15
                               Fig. 9. Sodium losses from fish to the test water, on a fimol/g/h ba-
                               sis, plotted against mean gfll Ag concentrations. Each point was de-
                               termined from six. 1- to 3-g rainbow trout held in IL of test water
                               for 2 h. The regression line (r = 0.778, n = \2,p< 0.01) does not
                               include the outlying point indicated by the open circle. See text for
                               more details.
                               coefficient of r = 0.560 (p < 0.05) if all 13 data points are
                               considered. Omitting one outlier (presumably due to sample
                               contamination) yielded a more significant correlation of r=
                               0.778 (p < 0.01; Fig. 9). These results give some indication
                               that the acute effect of Ag at trout gills is an ionoregulatory
                               disturbance (see Discussion).

                               Model testing

                                  Once the model was constructed by inserting the final val-
                               ues given in Table 1 into the MINBQL+ program, it was
                               necessary to test how well the model predicted Ag accumu-
                               lation on fish gills. We ran a thiosulfate experiment in Lake
                               Ontario water to see how much thiosulfate was needed to
                               protect against Ag accumulation in natural, moderately hard
                               water, to  compare with published thiosulfate protection
                               against Ag toxicity [3], In addition, we ran Ag experiments
                               in a variety of waters, to test the predictive capabilities of the
                               model. All predictions of gill Ag concentrations were made
                               using water chemistry given in Table 2, before viewing the
                               gill Ag results from the experiments.
                                  In water from Lake Ontario; only 0.5 pM of thiosulfate
                               was needed to keep 0.09 pM Ag off trout  gills (Fig. 10), and
                               predicted gill Ag matched observed values well (r = 0.954,
                               p < 0.001). For scaling predicted to  observed results, the
                               number of gill binding sites was set to 1.9 nmol per fish.
                               Water chemistry for the  Lake Ontario  water used in the
                               model prediction is given in Table 2.
                               ^   Observed and predicted gill Ag concentrations of fish held
                               2 to 3 h in Ag-supplemented natural waters also correlated
                               well, with one notable exception. The exception was city of
                               Waterloo tapwater, in which high gill Ag concentrations were

-------
 1854
                                             N. JANES AND R.C. PLAYLE
                             Table 2. Input exposure water chemistry data used in model testing
                                                                 Measured concentrations (pM)
                                                                                          Ag
Water
Reverse osmosis
Gait Creek
Salmon Lake
Grand River
Jack Lake
Dundas Pond
Lake Ontario
Waterloo tapwater
pH
7.9
8.2
7.5
8.2
7.4
8.1
7.8
8.3
' JAA;
(mg C/L)
1.7
7.9
4.4
8.S
8.3
4.7
4.8
3.3
Ca
140
1,882
822
2,094
506
2,029
1,058
2,860
Na
1,090
505
46
1,045
78
489
545
980
Cl
784
664
32
1,146
32
505
602
1,249
Background
0.01
0.01
0.01
0.01
0.00
0.01
0.00
0.02
Ag added
0.06
0.10
0.09
0.10
0.08
. 0.09
0.12
0.06
AH values entered into the model are the mean of two or three measurements, accent for background A« where
          ft = 1. Exposure temperature was 15*C.
seen, whereas the model predicted nearly background levels
(Fig. 11). The model predicted low gill Ag because of the high
Ca, Na, and Cl concentrations in the water, plus the mod-
erate DOC concentration (Able 2). The relationship between
observed and predicted gill Ag is clearly seen in Figure 12,
where the straight line through observed and predicted gill
Ag, excluding Waterloo tapwater, has a significant correla-
tion coefficient  (r=0.896,p < 0.01). The correlation includ-
ing Waterloo tapwater is not significant (r=0.543, p > 0.05).
   Likely reasons for the poor prediction of the Waterloo
tapwater results will be given hi more detail in the Discussion
section, but we suspected colloidal Ca as a factor. A filter-,
ing experiment  was run to try separating colloidal from dis-
solved Ca. Filtering with 0.45- or 0.2-/im filters gave similar
results, so they are averaged together. Jack Lake water gained
4 yM Ca and 6 jtM Na  upon filtering, Gait Creek water
jgained 8 jM Ca and lost 6 pM Na. while Waterloo tapwater
lost 72 fiM Ca and lost 8 jiM Na upon filtering. Small losses
or gains of Ca and Na after filtering probably represent an-
        0.00   0.00   0.06   0.10   0.60   1.00   2.00
                   WoautW* eoncmtntton (jiM)
Fig. 10. GOlAg concentrations of rainbx
-trout i
sd to 0.09 /.M
Ag for 2 to 3 bin Lake Ontario water in the presence of 0 to 2 pM
thiosulfate (15°C). dear bar = no Ag; striped bars = 0.09 pM Ag
exposure, with 95% C.I. indicated. *•• = significantly greater than
control values atp< 0.001. Hatched bars = model-predicted gill Ag.
aJytical variability, but the relatively large loss of Ca from
filtered Waterloo tapwater may indicate the presence of col-
loidal Ca.
   In addition to testing the model using a variety of waters,
we also added Ag and cadmium (Cd) together in a series of
exposures to determine if the presence  of  another metal
would affect Ag binding to trout gills. Cadmium was chosen
because it binds to fish gills with a similar binding constant
to that of Ag dog -Kcd-gnKa = 8.6). and its accumulation on
gills at -0.1 pM exposure concentrations is as easily mea-
sured as is Ag [15].
  • The presence of Cd did not significantly affect gQlAg con-
centrations, except for a tendency for slightly lower gill Ag
concentrations (Table 3). Thus, the gill model predicts gUJ7
Ag equally well whether the exposure was to Ag alone (r =
0.896 (fi = 8),p<0.01) or was in association with low con-
centrations of Cd (r = 0,886 (it =  8). p < 0.01; including .
Waterloo tapwater, r = 0.344 (n = 9), p > 0.05). In these!
treatments, Ag and Cd concentrations were 0.09 ± 0.02 and
0.11 ±0.01 jiM, respectively (mean ± 1 so (n - 8)).  Back-
ground Cd in control exposures was 0.004 pM (n = 2). The
significant difference between the  controls (Table 3) was
likely an artifact of the choice of water used in the control
exposures: In the Ag-pius-Cd experiments the controls (no
metals added) were from Salmon Lake and Lake Ontario
(six fish from each), whereas the Ag controls were from all
waters (four fish from each) except Salmon and Jack lakes.
For  completeness, gill Cd concentrations are included in
Table 3. Of most interest  is the Waterloo tapwater result,
where Cd was kept  off trout gills.

                     DISCUSSION

   We were able to calculate the equilibrium binding constant
for Ag binding to rainbow trout gills flog AAf-ttiiAc = 10.0),
with about 1.3 nmol Ag binding sites per fish. The binding
constant was determined  from complexation experiments
with thiosulfate GogA^-SiO, -8.8; Figs. 2 and 4). Know-
ing  the Ag-gill binding constant, Ag complexation by Cl
(Fig. 5) and by dissolved organic carbon (DOC; Fig.  6),

-------
                                      Modeling silver binding to gills of rainbow trout
                                                                                  1855
Fig. 11. Observed and predicted gill Ag concentrations for a series of lake waters to which 0.06 to 0.12 pM Ag was added. To scale the pre-
dicted values to the observed values, the predicted value for the reverse osmosis (R.O.) water was set to 100% of the observed value. Means
with 95% C.I. are indicated. *, **, •'* = significantly greater than control values at*p< 0.05, **p < 0.01, and ***p < 0.001, respectively.
Controls are the mean of 24 fish held in an waters except Salmon Lake and Lake Ontario, four fish each. See Table 2 for fuU names and chemistry
of the waters. .      •                                          -
and competition for Ag binding sites by Na (Fig. 7), Ca, and
H*, we were able to calculate other equilibrium binding con-
stants. These log A" values were inserted into the MINEQL+
aquatic chemical equilibrium model  [11] to model Ag-gill in-
teractions. .This model (summarized in Fig. 8) was used to
predict Ag accumulation on gills of trout held in Lake On-
tario water supplemented with Ag and thiosulfate (Fig. 10),
and to predict gill Ag of trout held in Ag-supplemented nat-
ural waters (Fig. II). Variables needed for prediction of gill
Ag were water pH and molar concentrations of Ag, DOC,
Ca,Na,  and ClfTable 2).
   Although the model predicted gfll Ag concentrations well
most of the time, it did not predict the high gill Ag concen-
trations in the Waterloo tapwater (Figs. 11 and 12). Water-
loo tapwater is very hard and is a mixture of well water and
Grand River water. This exception to the otherwise good pre-
                              dictions by the model may indicate a kinetic constraint oh
                              the thennodynamic model. If the Ugh concentrations of Ca,
                              Na, and Cl, and moderate amounts of DOC (Table 2), are
                              not available or are only slowly available to interact at Ag
                              binding sites at the gills or with Ag in the water, then their
                              protective effects will be overestimated. For example, Ca ex-
                              isting as colloidal CaCOj would not necessarily be readily
                              available to interact at the gills. Our filtering experiment in-
                              dicated that some Ca may be in large enough colloids to be
                              retained by 0.45- and 0.2-jun filters, but most of the Ca,
                              whether colloidal or free in solution, passed through the fil-
                              ters. Colloidal particles passing through filters is a concern
                              when trying to spetiate elements (18].
                                ' The possibility of colloidal Ca being slow to interact at
                              fisH gills and thus slow to protect against Ag deposition on
                              fish gills is analogous to organic-metal colloids controlling
                         Table 3.  Statistical comparison (unpaired t test) between gill Ag concentrations
                                   from the Ag and the Ag-phis-Cd-supplemented waters
                                        Oil! Ag (ranol Ag/g wet tissue)
          Water
Ag exposure     Ag + Cd exposure
                                                                    (test
Ofll Cd (nmol Cd/g wet tissue)

Ag + CdI exposure    ANOVA
Controls
Reverse osmosis
Gait Creek
Salmon Lake
Grand River
Jack Lake
Dundas Pond
Lake Ontario
Waterloo tapwater
1.7 ± 1.1 (24)
11.2 ±3.1 (4)
6.8 ±3.6 (8)
21.6 ±9.3 (8)
7.8 ±3.6 (8)
17.8 ±8.6 (8)
9.9 ± 5.3 (8)
10.8 ±2.3 (8)
23.8 ±1.4 (4)
2.5 ± 0.8 (12)
8.5 ± 3.5 (6)
5.7 ± 1.3 (6)
16.4 ±3.0 (6)
6.8 ±1.3 (6)
14.2 ± 3.5 (5)
8.5 ±2.0 (6)
9.5 ±3.2 (6)
24.2 ±5.4 (6)
p«0.03
/» = 0.26
p = 0.49
p = 0.21
p = 0.57
p = OJ9
p = 0.55
p = 0.39
p = 0.91
*
NS
NS
NS
NS
NS
NS
NS
NS
3.3 ±0.5 (12)
7.1 ±1.6 (6)
4.8 ±0.5 (6)
6.1 ± 1.5 (6)
4.3 ± 0.1 (6)
6.6 ±1.4 (6)
5.2 ±0.7 (6)
5.5 ±0.9 (6)
4.1 ± 0.5 (6)
_
*»*
*
*•*
NS
***
*•
***
NS
           Mean ± I so (n). The correlation between gill Ag concentrations of the two exposures was r = 0.967 (9), significant
           at p < 0.001. Gill Cd concentrations for the Ag-plus-Cd exposure are also given (mean ± 1 so (it)). In this case the
           statistical test (one-way ANOVA followed by Student-Newman-Keuls test) compared gill Cd of exposed fish with
           control values.
           »p < 0.05. »*p< 0.01; •••/»< 0.001.

-------
 1856
                                              N. JANES AND R.C. PLAYLB
  t
     30
     20
  1 10
                            10
                       predicted gill Ag  '
20
Fig. 12. Comparison of observed and predicted gill Ag concentra-
tions (in nmol Ag/g wet tissue) presented in Figure 11. The open cir-
cle represents the Waterloo tapwater result, which was not included
in the linear regression. See text for details.
trace-metal speciation in seawater [18]. The "onion" model
[18] suggests die existence of colloids composed of layers, so
that only the outer layer of a colloid is available to interact
in the water. It would therefore take time for metals on in-
ner layers of the colloid to react with a ligand, for example.
Calcium in Waterloo tapwater does not necessarily need to
be in colloidal "onion" form to reduce its effectiveness at
keeping Ag off the gills; it just needs to be in some form such
as colloidal CaCO}, which may only slowly dissociate. Sim-
ilarly, Na, Cl, and DOC in the water may not be in available
forms to be protective. Thus, a kinetic constraint on the
thermpdynamic model may exist for some waters such as
hard well water.                               .     ,  ^
   Why, then, did Waterloo tapwater keep Cd off trout gills"
(Table 3), but did not keep Ag off the gills? The answer may
lie in the relative Cd:Ca binding at Cd binding sites at the'
gill, compared to Ag:Ca binding at Ag binding sites. Log
Acd-suicri = 8.6, and log ATci-goicd ^ 5.0 [15], a difference of
log 3.6 (-4,000 times). In contrast, log^At^otAc = 10.0, and
log AcktuiAt = 3.3 (Table 1; Fig. 8), a difference of log 6.7
(~5 million times). Another way of viewing this question is
that 1 mM Ca kept 0.05 pM Cd off the gills (a 20,000:1 ra-
tio; [15]), whereas 10.6 mM Ca did not keep 0.05 pM Ag off
the gills (a 212,000:1 ratio; see Results). Thus, Ca in a slowly
dissociating colloidal form would not reduce the protective
effect of Ca against Cd accumulation on gills as much as the
protective effect of Ca against Ag, because there would likely
be enough Ca in the free, fast-reacting form to interact at Cd
binding sites on the gills to keep Cd off the gills.
   Although our methods do not'identify specific binding
sites on the gOls for Ag, the strength of Ag binding at the
gill (log AAc-fUiAf = 10.0) suggests covalent binding to, for
example, sulfhydryl groups. Silver has a low ionic index (0.8)
and a high covalent index (2.9; see [8,19]), which also in-
dicates covalent rather than ionic binding (class B metal
cation; [20]). Cadmium also binds relatively strongly to gills
(log/Toj.,iiicd = 8.6; [15]). Cadmium has both a Ugh ionic
index (4.1) and high covalent index (2.9), so Cd may form
ionic interactions with, for example, carboxyl groups [8], or
may also bind covalently with sulfhydryl groups. In our ex-
periment Cd did not interfere with Ag binding to trout gills
(Table 3), which suggests that these metals bind at different
sites on the gills. Cadmium interferes with Ca uptake at fish
gills [21], whereas Ag interferes with Na and Cl balance in
fish [22], again suggesting different binding sites.
   Taken individually, unnaturally high concentrations of
DOC, Cl, Na, Ca, and hydrogen ion were required to keep
Ag off trout gills. Dissolved organic carbon at >14.6 mg C/L
kept all 0.17 pM Ag off the gills (Fig. 6), while our lakewater
values were £8.5 mg C/L (Table 2). Previously, a4.8 mg
C/L DOC was needed to keep about 0.27 ?M Cu off fish gills
(14]. With similar DOC-metal binding of about log K=9 for
each metal, these results corroborate the high affinity for Ag
at the gills  (log A"AMtuA« = 10.0) compared to Cu binding
at the gills (log*cu-rfiie» = 7-4; [15]). The logA^ooc and
log XCU-DOC values are in the weak range, as defined for Cu-
humic acid interactions in seawater [13].
   Chloride alone did not keep Ag off trout gills at rea-
sonable concentrations: >1.5 mM Cl was needed to keep
0.11 ^M Ag off the gills (Fig. 5). Average North Ameri-
can freshwater contains about 230 /iM Cl ([10]; also see our
Table 2). LcBlanc et al. [3] found no toxicity of Ag in soft
water at a Cl:Ag molar ratio of 16,000:1. In our experiments,
103,000:1 Cl:Ag kept all Ag off the gills, and 14,000:1 Cl:Ag
did not keep Ag off the gills (Fig. 5). The 16,000:1 Ci:Ag
ratio for no toxicity [3] is intermediate between our no pro- >
tection and complete protection ratios. However, LeBlanc
et al. [3] added Cl as NaCl, so would have had some addi-
tional protective effect due to Na (see below).
   Sodium, like Cl, had very little effect on Ag accumula-
tion on trout gills until very high concentrations were reached
(16 mM; Fig. 7). Average freshwater has about 390 pM Na
([10]; also see our Table 2). Calcium was expected to com-
pete with Ag for gill binding sites, but did not do so at up
to 10.6 mM. Average freshwater has about 500 pM Ca ([10];
our Table 2). Similarly. H+ did not successfully compete
with Ag even at pH 4.5. Calcium and M+ therefore interact
at Ag binding sites at gills very weakly (Fig. 8).
   Taken together, completing agents such as DOC. Cl, and
thiosulfate, and competing solutes such as Na and Ca, ex-
erted cumulative protective-effects against Ag accumulation
on trout gills. For example, in the Ag and thiosulfate expo-
sures in Lake Ontario water, only 5.6 times more thiosulfate
than Ag was needed to keep Ag off the gills (Fig. 10), as
opposed to the 29:1 S^iAg ratio in synthetic soft water
(Fig. 2).  Less thiosulfate was needed to keep Ag off trout
gills in Lake Ontario water, because of the higher DOC, Ca,
Na, and Cl concentrations in Lake Ontario water (Table 2)
compared to synthetic softwater (see Results for the first thio-
sulfate experiments). The ratio of thiosulfate to Ag in Lake
Ontario water (5.6:1) is remarkably close to that from Le-

-------
                                       Modeling silver binding to gills of rainbow trout
                                                    1857
 Blanc et al. [31, where a measured 5.7:1 Sp3: Ag molar ra-
 tio eliminated Ag toxicity to fathead minnows.
   Cumulative protective effects of water chemistry were also
 seen in our experiments designed to test the Ag-gill inter-
 action model. Waters of lowest DOC and ion content (e.g.,
 Salmon Lake) showed highest trout gill Ag accumulations
 (both measured and predicted), and waters of highest DOC
 and ion content (e.g., Grand River) showed lowest gill Ag
 concentrations (Fig. II; Table 2). The notable exception, as
 discussed earlier, was Waterloo tapwater.
   Sodium efflux from trout to the exposure water was gen-
 erally high when gill Ag concentrations were high (Fig. 9).
 Metals stimulate ion effluxes by affecting the permeability
 of tight junctions of gill epithelial cells, allowing an ion such
 as Na to passively diffuse down its electrochemical gradient
 (23]. Ion fluxes to water are a potential noninvasive tool for
 assessing metal effects on fish (7). Our results' indicate a Ag-
 induced impairment of ionoregulation in trout, which agrees
 with preliminary results of Wood et al. [22]. In their study,
 adult rainbow trout showed large losses of plasma Na and
 Cl (from about 140 to 100 mmoI/L over 6 d) in response
 to -0.09 jiM Ag exposure in Lake Ontario water,[22]. It is
 not known whether the losses were due to increased efflux
 or to decreased  active uptake of Na and CI.
   In our 6-d exposure to Ag, trout exposed to Ag in the ab-
 sence of thiosulfate showed high initial Ag accumulations on
 the gills, followed by a decline in gill Ag (Fig. 3). This accu-
 mulation pattern has been seen for other metals such as alu-
 minum [24], and likely represents initial metal accumulation
 followed by sloughing of the metal with mucus. Mucus se-
 cretion at the gills is a standard acute response of fish to a
 toxicant [25]. Note that we had  no fish mortalities in our
 6-d experiment despite the fact that the Ag concentration
 (0.06 fiM) in soft water without thiosulfate was within the
 range of LCSOs for soft water  (e.g., [1]).     .       ,
   The small but significant accumulation of Ag on gills of
 fish exposed 147 h to Ag plus thiosulfate (Fig. 3) may be a
 result of enough time for the very low-concentration of free
 Ag+ to interact at the gills, likely through a disjunctive path-
 way of ligand exchange [26]. Alternatively, this Ag accumu-
Jation may represent diffusion of the Agi^Os complex into
 the gills. Although acute accumulation of metals on the gills
 is due to free ions, not complexed metal [14], diffusion of
 complexed metal may become important over the longer term.
   In conclusion, we have been able to determine Ag-gill
 equilibrium binding constants, and we have inserted these
 values into an aquatic chemistry equilibrium model to pre-
 dict Ag interactions at trout gills. The model considers com-
 plexation of Ag in the water surrounding a fish and cation
 competition for Ag binding sites on the gills. In essence, we
 have inserted biological components into a powerful aquatic
 chemistry program. This approach will ultimately allow bet-
 ter understanding and prediction of interactions of metals
 such as Ag with sensitive biological membranes.
Acknowledgement—Vie thank Kent Burnison for supplying and
analyzing the DOC used in this study. We also thank Lydia HoUis
•and Diane Stanley-Horn for their capable laboratory assistance.
Evelyn Playle assisted with lake and pond water collection. This re-
search was supported by internal research grants from Wilfrid Lau-
rier University and by a research grant from the Natural Sciences and -
Engineering Research Council of Canada.
                     REFERENCES

 I. Davits, P.H..J.P.GoetU, Jr. and J.R.Sinley. 1978. Toxicity of
    silver to rainbow trout (Satmo gairdneri). Water Res. 12:113-
    117.
 2. Nebeker, A.V., C.K. McAnliffe, R. Msnar and D.G. Stevens.
    1983. Toxicity of silver to steelhead and rainbow trout, fau
    head minnows and Daphnia magna. Environ. Toxicol. Ghent.
    2:95-104;
 3. LeBlanc, G.A., MX Mastone, AJ. Paradke, BJF. Wilson, H.B.
    Lockbart, Ir. and 1LA. RobHlard. 1984. The influence of spe-
    ciation on the toxidty of silver to fathead minnow (Pimepha-
    les promelas). Environ. Toxicol. Chem. 3:37-46.
 4. Dlamond,J.M.,D.G.MacUa,M.Cou1iuandD.Gniber.l990.
    Derivation of a freshwater silver criteria for the New River, Vir-
    ginia, using representative species. Environ. Toxicol.  Chem.
    9:1425-1434.
 5. Taylor, M.C., A. Demayo andS. Rceder. 1980. Guidelines for
    surface water quality, Vol. 1—Inorganic chemical substances,
    silver. Environment Canada, Inland Waters Directorate Water
    Quality Branch, Ottawa, Ontario.
 6. Lauren, DJ. 1991. The fish gill: A sensitive target for water-
    borne pollutants. In M.A. Mayes arid M.G.  Barren, eds.,
    Aquatic Toxicology and Risk Assessment: Fourteenth Volume.
    STP 1124. American Society for Testing and Materials, Phila-
    delphia, PA. pp. 223-244.
 7. Wood, CM- 1992. Flux measurements as indices of H* and
    metal effects on freshwater fish. Aquat. Toxicol. 22:239-264.
 8. Reid, S.D. and D.G. McDonald. 1991. Metal binding activity
    of the gills of rainbow trout {Oncorhynchus mykiss). Can. J.
    Fish. Aquat. Scl 48:1061-1068.
 9. Pagenkopf, G.K. 1983. Gill surface interaction model for trace-
    metal toxicity to fishes: Role of complexation, pH, and water
    hardness. Environ. Sri. Technol. 17:342-347.
10. Morel, F.M.M. and J.G.Hering. 1993. Principles and Applica-
    tions of Aquatic Chemistry. John Wiley & Sow, New York, NY.
11. Schecher, W.D. and D.C. McAvoy. 1992. MINEQL+: A soft-
    ware environment for chemical equilibrium modeling. Comput-
    ers. Environment ana Urban Systems 16:65-76.
12. Mldorikawa, T. and E. Tanoue. 1994. Detection  of a strong
    ligand for copper in sea water and determination of its stabil-
    ity constant. Anal. Chtm. Acta 284:605-619.
13. Hlrose.K. 1994. Conditional stability constants of metal com-
    plexes of organic ligands in sea water: Past and present, and a
 .   simple coordination chemistry model. Anal. Chim. Acta 284:
    621-634.
14. Playle, R.C, D.G. Dbcon and K. Burnison. 1993. Copper and
    cadmium binding to fish gills: Modification by dissolved organic
    carbon and synthetic ligands. Can. J. Fish. Aquat. Set. 50:
    2667-2677.
15. Playle, R.C., D.G. Dixon and K. Burnbon. 1993. Copper and
    cadmium binding to fish glib: Estimates of metal-gill stability
    constants and modelling of metal accumulation. Can. J. Fish.
    Amiat. Sci.  50:2678-2687.
16. SigmaStat. 1992. SigmaStat Statistical Software, Version 1.02.
    Jandel Scientific, San Rafael. CA.
17. Allison, J.D., D.S. Brown and KJ. NoTo-Gradac. 1991.
    MINTEQA2/PRODEFA2, a geochemical assessment model for
    environmental systems: Version 3.0. User's manual. U.S. En-
    vironmental Protection Agency, Washington, DC.
1.8. Mickey, D J. and A. Zirino. 1994. Comments on trace metal
    speciation in seawater or do "onions" grow in the sea? Anal.
    Chim. Acta 284:635-647.
19. Nleboer, E. and D.H.S. Richardson. 1980. The replacement of
    the nondescript term 'heavy metals' by a biologically and chcm-

-------
1858
                                                N. JANES AND R.C. PLAYIX
    ically significant classification of metal ions. Environ.' Pollut."
    lB:3-26.                                       (
20. Sluram, W. and JJ. Morgan. 1970. Aquatic Chemistry. John
    Wiley & Sons, New York, NY.
21. Verbost, P.M., J. V« Rooij, G. FMk, R.A.C. Lock and S.E.
    Wendelaar Bonga. 1989. The movement of cadmium through
    freshwater trout branchial epithelium and its interference with
    calcium transport. /. Exp. Biol. 145:185-197.
22. Wood, C.M., S. Monger and C. Hogstrand. 1993. The physio-
    logical mechanisms of tcnddty of sflver and other metals to fresh-
    water fish. Proceedings, First International Conference on
    Transport, Fate, and Effects of Silver in the Environment. Mad-
    ison, Wl, August 8-10, pp. 89-92.                 ;
23. Evans, D.H. 1987. The fish gill: Site of action and model for
    tone effects of environmental pollutants. Environ. Health Per-
    spect. 71:47-58.
24. McDonald, D.G., CM. Wood, R.G. Rhem, M.E. Mueller, D.R.
    Monnl and H.L. Bergman. 1991. Nature and time course of ac-
    climation to aluminum in juvenile brook trout (Salveliniafon-
    tinalis): I. Physiology. Can. J. Fish. Aquat, Sd. 48:2006r2015.
25. Florence, T.M., G.M. Morrison and J.L. Slanber. 1992. Deter-
    mutation of trace element spedatlon and the role of speciation
    in aquatic toxicity.Sc/. Total Environ. 125:1-13.
26. Bering, J.G. and F.M.M. Morel. 1990. Kinetics of trace metal
    complexation: Ligand-exchange reactions. Environ. Set. Tech-
    nol. 24:242-252.
                       APPENDIX
   Selected equilibrium constants (K) from the MINEQL+ com-   "
puter program (version 2.0) (SjOj = thiosulfate).
          Complex
logA"
          Agdf
          AgNO,
          NaSjOT
          NaCOf
          NaHCO,
          CaSjO,
          CaCO,
          CaHCOj*
          H,CO,
 8.80
 13.60
 3.27
 5.27
 5.29
 5.51
-0.29
 0,60
 1.27
 10.08
 1.90
 3.15
 11.33
 1.70
 10.33
 16.68

-------
                       In press  in  Environmental
                       Science and  Technology
 Binding  of Nickel and Copper to
 Fish Gills  Predicts  Toxicrty When
 Water Hardness Varies, But Free-Ion
 Activity  Does  Not
 JOSEPH S.  MEYER.' '
 ROBERT C.  SANTOS'*  IQE  P.  BflRi
TARRV D.  DtBHtrtiJCONSIE  ]. BOESl
 PAUL  R. PAQUIN.* HERBERT  E. ALLEN.*
 HAROLD L.  BERGMAN,* AND
 DOMINIC  M.  DITORO'
 Department of Zoology and Physiolog. University of
 Wyoming Laramle. Wyoming 82071-3166. HydmQual. Inc..
 1 Lethbridge Plata. Mahuiah. New Jersey 07430, and
 Department of Ctvil and Eniironmental Engineering,
 University of Delaware. Neuwk, Delauvre 19716-3120
 Based on a biotic-ligand model (BLM), we hypothesized
 that the concentration of a transition metal bound to fish gills
 {[MgaJII wiB be a constant predictor of mortality, whereas
 a free-ion activity model is generally interpreted to imply that
 the chemical activity of the aquo ('free") ion of the
 metal will be a constant predictor of mortality. In laboratory
 tests, measured [NigJ and calculated [Cu^J were
 constant predictors of acute touchy of Ni and Cu to
 fathead minnows (Pimephales prometad when water
 hardness varied up to 10-fold, whereas total aqueous
 concentrations and free-ion activities of Ni and Cu were
 not Thus, the BLM, which simultaneously accounts for (a)
 metal speciation in the exposure water and (b) competitive
 binding of transition-metal tons and other cations to
 biotic ligands predicts acute toxic'ity better than does free-
 ion activity of Ni or Cu. Adopting a biotic-figand modeling
 approach could help establish a mere defensible,
 mechanistic basis for regulating  aqueous discharges of
 metals.
 lotrodictioi
 For several decades, water hardness (the sum of the nor-
 malities of die divalent canons In solution} has been known
 to affect toxiciry of transition metals to aquatic organisms.
 with mortality at a specified total dissolved metal concentra-
 tion decreasing as hardness increases V). The U.S. Envi-
 ronmental Protection Agency (USEPA) method to compen-
 sate for the hardness effect in site-specific discharge permits
 has been to calculate a hardness-adjusted LCSO (median
 lethal concentration at a specified eiposure time) using a
 regression equation of the form
ln(LC50)
                                ; -r b
U)
 where  H — water hardness and a and b are regression
 coefficients fined to toxiciry data (e.g.. refs 2,3). However.

   • Corresponding author phctie:  (307 766-20IT: fax: (307. 766-
 5625: e-mail: meyerjeuwyo.edu.
   1 University of Wyoming-
   ' HydroQual. Inc.
   • University of Delaware.
 hardness is the only water-quality parameter for which such
 a procedure is used. Other water-quality parameters that
 modify metal toxiciry (e.g-, pH, alkalinity, dissolved organic
 carbon (DOC), total suspended solids (TSS)) can be accounted
 for only by using costly acd time-consuming, site-spedfic
 toxidty testing and tortuous calculations that still have no
 underlying mechanistic baas (e.g.. .water-effects ratios (4)).
   Toprovideamoremecbanistic basis forprcdictingtoxiciry
 of metals to aquatic biota, die free-ion-activity model (FIAM)
 was proposed (5. 6). Based  mainly on research  in which
 toxidty correlated with free-ion activity as the concentrations
 of organic ligands and dissolved organic matter were varied
 in exposure solutions (Tabjes 1 and 7 in ref 6). the FIAM has
 generally (but Incorrectly) been interpreted to imply that a
 constant degree of biological effect {e.g., mortality) wul occur
 at a constant chemical activity of the aquo ('free") ion of a
 metal ({M1'} for divalent transition-metal cations such as
 Cd*-, Cu»-, XP-, W; andZh*-), independent of otherwater-
 qualiry parameters. However, free-ion  activity does not
 appear to be a good predictor of toxidty across some water-
 quality conditions (Q. including when water hardness varies.
   Building on that concept a recently resurrected surface-
 interaction model (7) of metal binding to fish giUs not only
 incorporates an equilibrium between {M*~} and me con-
 centration of meal accumulated at binding sites on me giB
 surface (0 but also incorporates competition of M*~ with
 other cations (e.g., Ca*~, H*i for binding at the receptor sites.
 Historically, Zitko and Carson (8) suggested the cation-
 competition concept prior to Pagenkopfs (7) formal pre-
 sentation of a chemical-spedation model that incorporated
 a biotic Ugand. And although Morel (5. p 303) stated in his
 explanation of the FIAM,"_ the free metal ion activity is tbe
 parameter that determines the physiological effect of the
 metal ~", he later specifically mentioned (a) the importance
 of cation competition for binding sites on the surfaces of
-biotic ligands (5, p 307) and (b) the relationship between
 measurable physiological effects and the degree of cora-
 plexation of metals with reactive ligands on (or in) organisms
 (5, p 303). According to this type of model, which we refer
 to as a biotic-ligand mode! (BLM) to distinguish it from the
• current general interpretation of the FIAM, free-ion activity
 is anecessaiy but not sufficient component to describemetal
 accumulation and, presumably, toxiciry. Thus. bioavaOabiliry
,of metals can be decreased in two ways: (1} by decreasing
 {M1-} and, thus, decreasing  the potential for M to bind to
 receptor  sites (e.g., as increasing [DOC] does through
 equilibrium partitioning aaong the dissolved ligands) or (2)
 by increasing the concentrations of competing cations and,
 thus, decreasing the amount of M bound to receptor sites
 (e.g., as increasing {Ca2-} does).
   We hypothesized that 3 the BLM is correct, the amount
 of a specified metal accumulated on fish gills CM,d. expressed
 as mol M-g tissue'1) will be a constant predictor of mortality,
 independent of other water-quality parameters (except when
 {H*} becomes high enough to cause add toxidty). Borgmann
 (/) and Pagenkopf (7) alluded to but did not directly test tnjs
 concept for fish exposed to  transition metals, although it
 later was tested with Atlantic salmon (Satmosalar) exposed
 to Al and Zn as pH varied '10). Herein we present the first
 published evidence that {M*~ } is not a constant predictor of
 metal toxidty to fish as warer hardness increases, but • M^u]
 is.

 Experimental Section
 Ni Toxidty Tests. We exposed subadult (1-6 g) fathead
 minnows (FHM; Pimepha^s promeUu) to NiSO4 in four 96-
 10.1021/MSWMSq CCCr S11.00
 FuMishxi en Mto ewoaoooo
                             Amt-can Clwmial Sod*ly
                                                  PAGE EST: U
                                                      VOL «. NC a. «mc, i%VIRON. SD. & TECHKOW. * A

-------
h, continuous-Dow toxidty tests. In each test, a control and
11 serially diluted concentrations of Ni (25% decrease in
|Ni»uil at each serial dilution] were tested in a mixture of
well water and reverse-osmosis-treated, deionized (RO-DI) •
well water at the University of Wyoming's Red Buttes
EnvironmemaJ Biology Laboratory.  Exposure waters con-
tained a different concentration of Ca and a different series
of NiSO4 concentrations in each test, but temperature, pH.
alkalinity, and [DOC) remained constant  Water-quality
parameters (average or range of values} were as foflows:
temperature, 20 *C pH. 73 (range = 12-75); alkalinity. 0.5
mX; DOC. 20%). We sampled
gills at 24 h because results of preliminary studies indicated
accumulation of Ni by FHM gills was relatively rapid and
consistent with a one-compartment uptake-depuration
model in which the amount of Ni accumulated at 24 h was
~85% of  its predicted asymptotic value. Thus, sufficient Ni
associated with (i.e., accumulated on or in) the gills before
the majority of FHM began dying in exposure concentrations
that bracketed the LC50. Median lethal accumulations (LASOs.
analogous to the term LCSO) of Ni on FHM gills were
calculated by regressing logit'mortality) on In(INignl).
   Based on (a) die measured exposure-water quality and
(b) published stability  constants (but with  log K values
adjusted  to zero tonic strength using the Davies equation
(//)) for complexes of the major anions with the major cations
and Ni (12). we calculated [NP-J and {Ni*-} at the INi^J
LCSO for each toxidty test using the geochemical spedation
program MIXTEQA2 (13).
   Cn Toxidty Calculations. To further test our hypothesis,
we analyzed published acute-toxidty data for FHM exposed
to CuSO4 at various water hardnesses. Erickson et aL (14]
conduaed 12 96-h stan'c-nonrenewal toxidty tests in which
larval FHM were exposed to CuSO,. and {Ca1-] and |Mg*-]
were adjusted to vary water hardness among the tests (tests
"none" and 'add 2 mN CaSCV in set S2 and all 10 tests in
set S3 in their Table 2, all of which were conduaed at pH
7.8). We calculated {Cu*~} and [CtifBi] at the [Cudwindl LCSO
in each test by entering [CitsmwJ. other inorganic water-
quality parameters (15), and ."DOC] 0.8 mg-L*') into the
geochemical  speciation program CHESS (16). For  these
calculations, the CHESS model was used to simulate (a) Cu-
   1!'
   Oo
   -iw
                                         P*W
                    Hardness  {mN)
ROURE1. Manured 9S-b LC50» (nediio Itttaf i
Mtraoonlfor
[NU1 INP-1 aod {II?-} la exposure water* aad USOi (medial
lethal accuBulation. expressed on a wet weight (wwj of twwe
basis} for [NI^L for Mead aiuows (FHM| *xpot*d to NiSO, at
varion water baidaeis«$.U50imffieoac«^^
actintr ({}) mragod over (he 954 oxponra; IAS* an ateocorad
time  birrioa of  Hi ascodattd wlft RON  |ills at 24 h (bat
eomsaoRdiig to 50% awrtalitr it 96 b). Error ban npmmrt 95%
aud WW Enw barsfcr [N?-]«d {H?-} at a spoeoled kardaess
aro •nportiooal la cia to tfat emr bar for IKi^J.

organic matter Interactions described in the humic-sub-
stances model WHAM (17), tssuming the dissolved organic
matter (DOM) was composed of 10% humic add and 90%
ruMc add, and (b)WndingofCu»-,Cal-andH- to FHM gills
using published conditional stabilit}* constants (S). Least*
squares Gnear regression equations wen computed for
[CtidiwM]. {Cu*-}, and [CUpaJahr vs [Ca], and two-tailed
significance of each regression slope (H^ slope = 0) was
tested at a = 0.05.

Resilts til Diseissioi
The 96-h iNlaail LC50 increased 10-fold as (Cai increased
10-fold (Figure 1), as expected from the hardness correction
equation fit the USEP.Vs Ni criteria document (2). Moreover,
the (XP-j LCSO also increased 10-fold over the same range
of water hardness, and the {Ni1-} LCSO increased 7-fold
(Figure  1). Thus.  IXi**]  and  {HP'} were sot  constant
predictors of Ni toxidty (Le. . one would have to know water
hardness in addition to (NF") ir {NP-} to beable to accurately
predict the LCSO). But INi^ was approximately a constant
predictor of toxidty. because the LA50 increased only 2-fold
over the 10-fold increase In water hardness and appeared to
approach a plateau at high hardness (Figure 1).
   The 96-h jCiidimM] LC50 increased significantly by 3.1-
fold as [Ca! increased 5-fold (Figure 2a), as expected from
the hardness correction equation in the USEPA's Cu criteria
document (3). Moreover, the {Of-} LCSO increased sig-
nificantly by 43-fold (Figure 2b). Thus, {Cu2'} was not a
constant predictor of Cu trarfdty. But [Cu^c was ap-
proxunateVaconstantprediaoroftoxicity.bec2usetheLA50
did not increase significantly over the 5-fold increase in water
hardness and appeared  to approach  a  plateau at high
hardness (Figure 2«. Tabulations of the data plotted in
Figures 1 and 2 are available from the lead author.
   As noted by Erickson et aL  (14) for their plot of Cumi
LCSO vs hardness, die slope of our plot of Cu*.** LCSO vs
(Caj (Figure 2a) appears to decrease slightly as |Ca| increases-
even if both axes are In-tranrformed. This cuniUnear trend
is counter to  predictions of BLM-type models but can be
 B • ENVIRON. SO. It TCCHNOL / VOL a. NO. «.

-------
       1
           2-
        a
       •-. o


      I"
      o  0.0
     <- 0.02 -,
         0.01 -
     ^ 0.00
                           Ca  (mN)
 FIGURE 2.
 far [CvteMl in exposure watea. (b) calculated 36-b LCSfe far
 {Co1-}, and (e) calcolated USDs (madlu lenat accmnutanoas.
 Harassed on a wet might fww) of tissM basis) few {Cu*L far
 fathead aunaowi (HIM} exposed to CaSO« at wioos  water
 barfMfses(t4.USOiiraCr,-ofCu in the presence
 of humic acid, and Penttinen et aL (19. reported a loss of the
 protective effect of DOM as water hardness increased when
 Daphnia magna were exposed to Cd. Second. Cu disrupts
. ionoregulation in fish  {20). Because gills may be  more
 permeable to body tons at low [Ca] than they are at higher
 iCa]  (20). the lower than expected CuoMhtd LCSOs at low
 {Gal  might  resuh  from diminished, direct physiological
 protection by gill-bound Ca-a process that is not accounted
 for by our BLM. This might also explain  why the LASOs of
 INifnJ (Figure 1) and [Cu^W (Figure 20 appear to increase
 slightly at low hardness and approach plateaus at high
 hardness. Third. Erickson et aL (M) only plotted results from
 their toxidry test set S2 in their Figure 3. and the nonlinearity
 of their Figure 3 was influenced strongly by the two extreme
 LC50 values. However, when we included two additional data
 points from their toxicity ten set  S2 with their set S3, the
 trend appears  to be more linear at the high IQ: end of our
 curve (Figure 2a) than it does at the high hardness end of
 their curve. Finally, combined with the uncenainty about
the LC50 estimates (see 95% confidence intervals in Table
2 and Figure 3 in (J4U. the apparent curvilinear relationship
at low [Gal in Figure 2a might even be an artifact.
   We collude that accumulations of N'i and Cu on FHM
gills are approximately constant predictors of toxidty when
the concentration of Ca (the major competitor with Mi and
Cu for bicding to the  gill) in exposure waters increases.
whereas the free-ion activitiesofNiandCu in exposure waters
are not constant predictors cf toxStity. This also appears to
be valid when pH is varied among Cu toxicity tests, although
{Cu2*} is just as good a predictor of toxicity as [Cu,a]c« is
when [DOC is varied (21). And. emphasizing the importance
of cations other than Ca1* as competitors with some metals.
K* (but net Ca1") ameliorates acute and chronic toxicity of
thallium  (Tf) to the amphipad HyaltUa ttzteat, and body
burden ofTI is an appnndmatdy constant predloorof chronic
lethality and growth effects across a range of aqueous K*
concentrations (22). Therefore, the BLM. which simulta-
neously accounts for (a) metal spedadon in the exposure
water and 3>) competitive binding of transition-metal ions
and other cations to biotic figands predicts acute toxicity
better than does free-ion activity of Cu. Ni, and Tt
   Adopting a blotie-ligand modeling approach could help
advance  the regulation of aqueous discharges of metals
beyond the current phenomenological approach and es-
tablish a more defensible, mechanistic basis. In the future,
regulatory limits might be based on accumulation of meials
on biotic figands (e.g.. fish gOb. soft tissues of invertebrates,
algal cells) measured in the field or predicted in dynamic
simulation models that would estimate the number of dairy
exceedences of a regulatory Emit at a site downstream from
a metal discharge (2J). Howem, H* (10) and Ca2" (20) bound
tofishgillsperfonnimportaDLdtrectphysiologicalnuKOons
(e.g, altering membrane permeability and ion transport)
beyond just competing with transition-metal cations for
binding sites. Such beneficul functions have not yet been
incorporaiedintoaBLNLThcSrbiotic-Bgandmodelingcould ,
be used to complement  (but not totatty replace) toxidty
testing, by providing much greater temporal coverage than
is currently-feasible within th* finantial constraints and time
limitations of standard fish, invertebrate, and algal toxidty
tests.

Acbtwteigneitt

This research was funded by a cooperative agreement
between the U.S. Environmental Protection Agency and the
Unrversiryof Delaware: a grant from the Intemadonal Copper
Association: and a grant tc J.S.M. (and others)  from the
National Science Foundac^n's  EPSCoR  Program. EQen
Axonann assisted with the geochemical spedction calcula-
tions. Comments from three anonymous reviewers improved
the manuscript.


Ihcntirt Cited

 (1}  BorEnana U. In Aquatic roxicolog): Nriagu. }. O.. Ed.; John
     WUey and Sons: New Yori. 1983; pp 47-72.
 '2)  Ambittu Vtaur Quality Crxriafor AtdW -  JM6. U.S. Envi-
     ronmental PntectionAgenrr. Washiegton.DC1986:EPA*40/
     5-86-r/X.
  •3}  Ambimt Water QuaU/y Oteria for Copper -  J9M. 'J.S.
     Em-in-amental Prelection ««ency: Washington. DC 1985: SPA
     WQ/S-W-031.
  -41  Intenm Guidance en Dcie~nination and Ust cf Water EFtca
                                              : \Vish-
     iapcr. DC 1994: EPA/823 B-94/00).
  *5) Morti P.
     New Vufc. 1983.
                                                                       VOL t*. NO. UL an* I INVflON. SC t TECMNCL •

-------
  (6) Campbell P. G. C In Metal Spxiatlon and BloauOabUiry in
     Aquatic Syitemr. Testier, A.. Tinner. D. JL, Eds.; John Wiley and
     Sons:  Ouchester, 1995; pp 45-102.
  (7) Pagenkopt G. K. Environ. SO. TechnoL 1983.17.342-347.
  (8) Playte, R. C:Diioii 0.0:Burcaon.Hi Can./, fish. Aquat. Sri'
     1993, 50,2678-2687.
  (9) Zitto. V.; Carson. W. G. Chemaphert 1976, 5, 299-303.
 (10) Roy,R.R,;CanVbelP.G.CArJfltrftrton99S.3J.155-176-
 (11) Serklz. S. M: Allison.). D.; Perdue, E. M.; AUen, R E.: Brown.
     D. S. Water Kes.l9X.30.1930-1933.
 (12) XlSTCHtiadty Selected StabiUnCorutartu of Metal Compleus
     Database.  Version 5.0. NIST Sundard Reference Database 4&
     \ationa) InstltuteofStandardsindTechnology: Ga&henburg,
    MD. 1998.
(13) Allison, I. D.: Brown. D. S.; Novo-Gradac, K. J. MPtTEQAZf
    PRODEFA2.AGtoi*f*-
    Factors Model far Site-specific Copper Water Quality Criteria:
    US. Environmental Protection Agency: Duhiih, UN, 1987
    (revised 1996).
                                                                              .,.
                                                                    Chem. 1998, 17.24J8-2S03.

                                                                ao; McDonald, O. O; Reader. J. P. Dalziet T. R. K, In Ara roiirin-
                                                                    and Aquatic Animal. Morris. S. R., Taytar. E W., Brown, O. J.
                                                                    .A.,BrowT>.J.A,EdsjCambrtijeUnvienayPress: Cambridge.
                                                                    UK 1989: pp 221-242.
                                                                (21J DiToro,D.M.;AlteaaE:Be^nan,HL;Mahony,J.D.:.Me>-«.
                                                                    f. Sj  Paqufn. P.  3L Saniors. R. A. Oiembuy of copper
                                                                    bloavalUbBiry: a model of acne undcity to fish (nanuscripi
                                                                (22) Borgmahn.U.;Che*ai.V^ Norwood W.P--t«*n«f JBrtirf»-
                                                                                        	^—^^ — «•• i*Mw,wav
-------
                      In press  in Environmental
                      Science and Technology  .
 A Mechanistic Explanation for the
 in(LCSO) vs  In(Hardness) Adjustment
 Equation  for Metals

 IOSEPH  S.  MEYER*
 Department of Zoology and Physiology. University of
 Wyoming, Laramie, Wyoming 82071-3166            (
 I demonstrate that a combination of (a) competitive
 binding of transition-metal cations, hardness cations, and
 protons to transition-metal-binding sites on fish gills
 and (b) aqueous complexation of transition-metal cations
 by HCOs" and C0f~ explains why the regression slopes of
 InJLCSO) vs In(hardness) for five divalent transition metals
 (Cd, Cu, Ni, Pb, and Zn) are ~1, where LC60 is the
 median lethal concentration. For these calculations, I
 assumed the amount of the  transition metal bound to the
 fish gill at 50% mortality is constant (i.e., independent of water
 quality). Although the slopes theoretically should vary
 between 0  and 2 (at extremely low and high hardness,
 respectively), a slope of -1 is expected at midrange hardness
 (~20-200 mg-L-' as CaCOj) if alkalinity covaries with .
 hardness—a common condition in most laboratory toxichy
 tests. But if alkalinity is held constant while hardness is
 varied, 'a slope of ~0.5  is expected at midrange  hardness.
 Although predictions of LCSOs using regressions  of In-
 {LC50) vs In(hardness) might be acceptable for regulating
 discharges of transition metals to waters in the midrange
 of hardness, extrapolations beyond this range might drastically
 overpredict metal toxicity.


 Inlndictnw
Toxldty of transition metals to aquatic biota varies as  a
 function of several water-quality parameters (e.g.. hardness,
pH, alkalinity, concentration of dissolved and paniculate
organic matter) (1, 2). The U.S. Environmental Protection
Agency (USEPA) water quality criteria for Cd.Cu, Ni, Pb, and
Zn (3-7) specifically account for the increase in LCSO (median
 lethal concentration] as water hardness increases, using
regression equations written in the following general form
            ln(LC50) = fl-lnftiardness) + b         (1)
where a = slope, b = ordinate intercept, LCSO is in units of
/ig-L"', and hardness is in units of mg-lr1 as CaCOj. The
form of this equation has never been mechanistically justified;
instead, eq 1 merely provides a convenient, empirical fit to
the data. Surprisingly, though, the slopes for all five transition
metals are -1 (range = 0.76-1.27; Table 1).
   In this paper, I propose an explanation for these apparently
coincidental relationships between LCSO and water hardness.
This argument Is based on the concept of competitive binding
of cations to fish gills (8,9).

Model
I present the model in terms of H", Ca1*, and  a generic,
divalent transition-metal ion M**  binding to  fish  gills.

  •Correspondlngauthorphone: (307)766-2017;fax: (307)766-5625;
 e-mail: meyerJ0uvwyo.edu.
                                 TABLE 1. Coefficients for the Equations Predicting Acceptable
                                 Concentrations (prL-ti of Transition Metals as • Function gf
                                 Water Hardness (am-lr* as CaCO,) for Freshwater Biota
                                 Exposed to Cd. Cn, Ni, Pb, and Tsf

                                  metal     slop*'    intercept*     depudeot variable     rcf

                                  Cd    1.128    -3.828    criterion max. concn    (31
                                  Cu    0.9422 ,  -1.464    criterion max. concn    (4
                                  Ni     0.76      4.02     final acute concn       (5)
                                  Pb    1.273    -1.460    criterion max. concn    (6)
                                  Zn    0.8473    0.8604   criterion max. conch    17)
                                  •Tl» coefficient tin eq 1. »Tht coefficient 6 In eq 1. 'The general
                                 MnnoflheprKfictorequMkinbdeperKiertviriafalaeexpltlopeHnlhanlness)
                                 + intercept}.


                                 However.it also applies to Mg1* instead of (or In conjunction
                                 with) Ca*~ as the hardness modifier of toxidty. The model
                                 also may apply to other biotic ligands such as soft tissue of
                                 invertebrates.
                                   Equilibrium equations for binding of Ca*+ and M" to
                                 sites at which M binds on a fish gill can be written as (9)
                                                          (MggillMI
                                and
                                                          ICaegillMl
                                                                                 (2)
                                                                                 (3)
                                where XM^OU  « stability constant for M1* binding to
                                M-blnding cites on the gill (L-moH), Kc*4at - stability
                                constant for competitive binding of Ca1+ to M-blnding sites
                                on die gill (L-mol-'), |M»*|» aqueous concentration of M»*
                                (mol-L-1), |Ca**1 - aqueous concentration of Ca2* (rnoH,-').
                                {tsgillMl = concentration of unoccupied M-blnding sites on
                                the gill  (mol-L-'), JMBgillMl = concentration of MngfllM
                                complexes (mol-L-'), and {OuBgOlM) = concentration of
                                CBBgillM  complexes (moHr1).
                                   ForsimpUdty.IhaveignoredactivitycoerBcientsinthese
                                equations. Moreover, I have expressed the concentrations of
                                the binding sites and the MBgfflM and CasgUiM complexes
                                as if they are uniformly distributed in the exposure water.
                                This implies that the fish gills are in equilibrium with the
                                entire volume of water in the exposure chamber.
                                   Combining eqs 2 and 3.1 express (M2*) as a function of

                                                                                 (4)
                                   Assume (MegUlMI at 50% mortality is constant (i.e.,
                                independent of hardness, as demonstrated recently for Cu
                                and Ni (JO) and first proposed as a general concept by (1)
                                and («). Equation 4 can be rewritten as   •
                                                              (Ca1*]
                                                                                 (5)
                                where LCSqgb^is the concentration of the aquo ion of M that
                                causes 50% mortality, and  K is a constant (=  (Kc^&tti
                                ^H^HMJ-iMHgillMlwtoo,^). Combining eq 5 with the equi-
                                librium spedation relationships in the Appendix, the LCSO
                                of a divalent transition metal to fish can be approximated
                                algebraically as a function of pH, water hardness, alkalinity.
10.WJV«S80714v CCC: 118.00
PuMiOnd on Web 00/00/0000
« JCMX Amiric*n Chwnic*! SI»«IY
                                                 •PAGE EST: 4Jt
                                               VOL **, NO. Kx. nou / ENVIRON. SO. & TECHNOL • A

-------
 TABLE 1 Approximation? to Iq 7 it Varions Alkalinity and Hardness Conditions
hardness
               alkalinity
                                approximation ta sq ?'
oxpiioatioo
    low    constant
                           ln(LC50)» 0.0-ln«Ca»*l) + ln(«  At low [Ca'+L [CaJ+MCaag«IM] (= i_^__	
                                                          is approximately constant because the concentration of
                                                          unoccupied M-binding sites on the gill dmgWM})
                                                          is approximately constant fi.e., little Ca binds to the gill.
                                                          causing little change in IMagiUM], (HagillM] and
                                                          IsgillM]);thus. InUCa**}) a ln((CaagillM)) + Inconstant).
          equal to hardness  In(LCSO) a 0.0»ln([Caz*l) + ln(M  Same as above, and Alk (alkalinity) approaches zero as ICa3*)
                                                          approaches zero.              .
          constant          In(LCSO) a 1.0-ln(ICaz*l) + ln{«  If (MsgillM] is constant at 50% mortality (independent
                                                          of |Caz*|). {CaagillMI will approachfrTOTn
                                                          («(BglllMnMl - (MBgillML- eq A-Blas Ca fills the
                                                          remaining binding sites on the gill (i.e., as H* te
                                                          outcompeted and no unoccupied M-binding sites remain).
          equal to hardness  ln(LC50) • 2.0*ln((Ca2+]) + ln( W  Same as above, but Alk = hardness results in LC50 «
   low

   high




   high
   • k is a eompoih* constant that, in each aquation. r«pr*wnts tha product* of th« individual re and all otto terms (or ratio* of Mrms) in *q
 7 that are -apprexirnaMly constant whan hardness a axtramaly tow or high,
, die amount of Ca2* bound to the transition-rnetal-binding
 sites on the fish gill and the amount of unoccupied metal-
. binding sites, as follows (see derivation of eq A-4 in the
 Appendix)
                 (CasgillM]

where LC5(g^ls the total dissolved metal concentration
at 50% mortality. Alk b die alkalinity of the exposure water
(eq-L'1), and k" and K" are constants. Taking the natural
logarithm of both sides results in a relationship that begins
to resemble eq 1:
                          - ln([
                                       l) +
                                                   (7)
 Because JT'ADc«14 kf/lO-f*1 at low alkalinity and JT'-ATk
 »1 -f ir/lO-f" at high alkalinity, eq 7 simplifies and attains
 the same form as eq 1 when hardness is very low (e.g., <~1
 mg-L*1  as CaCOs) or very high (e.g.,  >~1000 mg-L'1 as
 CaCO})- At low hardness (either when alkalinity b constant
 or when alkalinity equals hardness), the slope approaches 0
 (Le..lnaC50)«jiO.O-ln([Ca'+J)+ln(t). where Jtisacomposite
 constant: Table 2); at high hardness the slope approaches
 either 1 or 2, depending on whether alkalinity is held constant
 (slope approaches 1; Le., InflJCSO)  • 1^-InaCa2*]) + ln(t);
 Table 2) or alkalinity equals hardness (slope approaches 2;
 i.e., bi(LC50 • 2.0-taflCa1*!)  + ln(tt; Table 2).

 Example
 To demonstrate the effects of hardness, alkalinity and pH on
 transition-metal ftnddty, I present results of simulations for
 Cu binding to fathead minnow (FHM; Pimephalespromelas)
 gills. These calculations incorporate all of the Cu complexes
 and activity coefficients ignored for heuristic purposes In
 die previous section and in the Appendix. Although it would
 be more mechanistically appropriate to express LCSOs of Cu
 in units of/tM and hardness and alkalinity in units of meq-L'1,
 I have retained the more traditional units of/tg'L-1 for LCSOs
 and ing-Ir1 as CaCCh for hardness and alkalinity to allow
 easier comparison to the historic database on metals tenacity.
    Using toxidty data reported for FHM larvae by Erickson
 et al. (11), I first calculated the amount of Cu bound to the
 FHM gills at 50% mortality using the geochemicalspeciation
 program MINTEQA2 (J3. For this and the subsequent
                                                                10000
                                                             T-  1000
                                                             o
                                                                                 •tope
                                                             -I     10
                                                                               10
                                                                                      100
               1000    10000
                                                                       Hardness (mg-L"1 as CaCOj)

                                                         HGIfflEl. l^dictedLCSQsofCoMitofatbeadrolaaowsutunliiess
                                                         varies at pK I. assorala> that the amount of Ca bound to the gills
                                                         Is constant at 50% Mortality (U, iadeseadent of water qaslitr). Co
                                                         bound to tte glib was held constant at 365% of the 10-* M Ca-
                                                         binding sites. Tba solid earn b predicted LGSOs whan alkalinity
                                                         eqnabbvdness;lhedash«dainebpfadlctedl£SOswheDalktnnhy
                                                         b hold constant at SO ng-L'1 as CaCO>
                                                         calculations In this example. I assumed the density of Cu-
                                                         binding sites was 10-* M (le., ~1 g of fish per L of exposure
                                                         water). Although  this  choice of binding-site  density b
                                                         arbitrary, the calculations were insensitive to changes in
                                                         binding-site density at s!0-»M. Based on Figure 1 in Erickson
                                                         et al (11). I assumed the 96-h LC50 of Cu In flow-through
                                                         exposures is 0.2 ftM (lZ.7fig-L'1) at pH 7.0 and alkalinity and
                                                         hardness equal to -50 mg-L'1  as CaCQ>- Additionally, 1
                                                         replaced the default stability constants in M1NTEQA2 with
                                                         thevaluesinVersJonS-OoftheNadonallnstituteorStandards
                                                         and Technology's electronic database 03) (but with log IT
                                                         values adjusted to  zero ionic strength using the Davies
                                                         equation (14)) and used conditional stability constants
                                                         reported for binding of Cu*-, Ca2*, and Ht to FHM gills (log
                                                         if= 7.4,3.4, and 5.4, respectively (9)). Because (a) those gill-
                                                         binding constants are not precise estimates and (b) the ionic
                                                         strength  of the exposure solutions in which they  were
                                                         determined .was only --0.0001  M (i.e., using the Davies
                                                         equation (J4), <4% correction would be needed to adjust for
                                                         nonzero ionic strength), I did not correct the gill-binding
 • * ENVIRON. SO- Si TECHNOL / VOL. xx. NO. xx. uxx

-------
       10000
                     10
100     1000    10000
             Hardness (mg-L"1 as CaCO,)

FIGURE 2, Predicted LCSOs of Co** to fathead mlnnowi is pH
varin fnm S la 9. based m die assumption dial (a) the amount
of Co bound to die gills M 50% mortality b 365% of the 1(H M
Co-binding sites and 
                                                             1000
                                                         Ll
                                                          3  100
                                                         o
                                                               10-
                                  0.1
                                                                         10000
           1        10      100     1000

            Hardness  (mg-L"1 as

FIGURE 1 Data osed la <4) to generate die slope of 034 for the
regression of lo|LC50) on la(hardness, for Co. The  solid carve
represents die simulated LCSOs calculated tathb paper: the dashed
One Is a slope of 1. Species abbreviations are DM=Diphaia mtgat,
DP = DapAo/* pattetrit. CS = cfalnook sermon, CTT = cnflhroat
trout. RBT « rainbow trout, FHM = fathead eihuow. and BG =
bloegill.

is negligible and competition by Ca1* for M-blnding sites is
minimal because of  die large percentage of unoccupied
binding sites on die gUL Thus, an Incremental Increase in
hardness and alkalinity has almost no effect on metal binding
to die fish gill, and die slope of m(LC50) vs m(hardness) is
-0.
   2. But if hardness and alkalinity are extremely high (e.g.,
> 1000 mg-L'1 as CaCOj), a large percentage of die metal in
solution is complexed with HCOr  or  COj2".  and die
M-bindlng sites on die gill are alroo^cnmpletehr occunicd
byCa1;
                                    ig sites on die gill i
                                    Thus, an increnu
         THIS, an incremental increase in hardness and
         requires a large increase in aqueous metal con-
centration to overcome (a) die additional complexation of
M1* withHCOr and GCtf- and (b) the tendencyfor additional
Ca2* or Mg»* to displace die metal from die gUL Under these
conditions, die slope of ln(LC50) vs InOiardness) is -2.
   3. At intermediate hardness and alkalinity (-20-200
mg-L'1 as CaCQ^h-die range in which most laboratory toxidty
tests are performed (see tabulations of hardnesses and LCSOs
in (3-7))—these effects are intermediate  and die slope of
ln(LC50) vs InOiardness) is -1.
   LCSO data used to calculate die slope of 0.94 in die Cu
criteria document (4) demonstrate tills coincidence (Figure
3). All of the hardnesses in that dataset were between 10 and
400 mg-L'1 as CaCCfe thus, die theoretical slopes should
have been between -1 and -1.5 (die calculated slopes for
die eight spedes ranged from 0.61 to 136 {4: p 47)). The
pattern for each species approximately parallels the LCSO
curve I simulated for FHM larvae, although die data for less
sensitive species (e.g., bluegDIs) and die less sensitive adult
We stage of FHM lie well above die curve. Variability within
die datasets for most of the species obscures a tighter fit to
die theoretical curvilinear relationship, in part because (a)
a variety of strains within a spedes probably were tested, (b)
the toxicitytests were conducted ata variety of temperatures,

             . VOL IX. NO.«. nn / ENVIRON. SO. ft TECHNOL * C

>*;•••

-------
 pHs and alkallnities. and (c) the range of hardnesses is too.
 narrow to clearly demonstrate the curvilinear relationship. -
    Although hardness and alkalinity are approximately equal
 in many surface waters (Including those used in many toxic! ty ••
 tests), some transition-metal toxidty tests are conducted at;
1 constant alkalinltj^while hardness is  varied  (or vice versa)
 to demonstrate the effect of hardness (or alkalinity) on metal
 toxidty. In trie midrange of hardnesses, the slope of the In-
 (LC50) vs Infjiardness) regression for divalent transition
 metals should be ~O.S when alkalinity is maintained constant
 Supporting this prediction, I found a slope of 0.50 (r2=0.85,
 P ~ 0.0002) when I regressed ln(LC50) on ln(hardness) for
 the 10 data1 points in Figure 3 of Erickson et al. (11) (96-h
 LCSOs for FHM larvae exposed to Cu; hardness ranged from
 37 to 134 mg-L-' as CaCOJ.
    A similar effect of alkalinity on transition-metal toxicity
 should occur as hardness is held constant Thus, it is not
 surprising that Erickson et  aL  (11) (their Figure 1) only
 observed -30-40% increases in LC50 of Cu** when they
 increased alkalinity from ~45 to -150 mg-L-1 as CaCQj at
 pH between 7 and 9 and a hardness of-45 mg-L'1 as CaCOj
 (15). Such  small  effects of  alkalinity on  LC50 would be
 predicted if die  slope of the  InfLCSO)  vs bi(alkalinity)
 regression in this alkalinity range was -03-a realistic value
 because hardness was held constant At higher alkallnities,
 Erickson et al. (11) in theory would  have seen a greater
 percentage increase in LC50 for the same percentage change
 in alkalinity, whereas at low alkalinlties, they would have
 seen almost no effect of alkalinity on the LC50. Knowledge
 of alkalinity and hardness is crucial to accurately predict the
 LCSOs of transition metals using m(LC50) vs Infliardness)
    Erickson et aL (11) (their Figure 1) also showed that LC50
 of Ciimd increased as pH Increased between 6 and 9. This
 agrees, with the gill-binding theory (Figure 2), because
 competitive binding of protons to the gfll at pHfc6is minimal;
 thus, the only effect of increasing pH is to increase the COf~
 concentration (at constant alkalinity) and the amount of Cu
 completed with that ligand. However, at pH <6  proton
 binding to the gin can be significant because (HI becomes
 high enough (> 10~* M) for protons to compete with Ca**
 and Cu2- (Le., (H*)-*H^BO, approaches (Ca'+l-/^^ and
 (Cu^J-JtoMgDCo). Therefore, at low alkalinity and hardness,
 the LCSO of Cuuui at pHS theoretically should be higher than
 at pH £6 (Figure 2), but at high alkalinity and hardness, the
 reductionincoraptexationofCubyHCOj-andCOj2- caused
 by the tower pH should help offset the Increase in the LC50
 of Cum* that otherwise might occur. This prediction does
 not take into account the potential onset of add toxicity at
 pH <6 that wfllbe caused byaccumulation of protonson the
 gill, a process  that would tend to decrease the Cum*
 concentration at which 50% mortality, occurs when the
 toxicants act jointly.
    I have presented examples for Cu because of the extensive
 LC50 and  gill-binding data available. However,  similar
 hardness-related trends should occur with Cd, Ni, Pb, and
 Zn. In general, I predict the slopes of m(LC50) vs bi(hardness)
 for these four transition metals will tend to be lower than
 they are for Cu in the same hardness  range (if alkalinity
 covaries with hardness), because Cd, Ni, Pb, and Zn have
 lower affinities for COa1- (13). Butlf alkalinity is held constant,
 I predict the slopes for all five transition metals  will be
 approximately the same in the same hardness range. The
 pH-related trends for toxidty of Cd, Ni, Pb, and Zn to FHM
 are less pronounced (Pb) or are the opposite (Cd, Ni, and Zn)
 of the trend for Cu (16), in pan because the chemical
 spedation of Cd. Ni, Pb, and Zn is much less affected by
 changes in pH than is the spedation of Cu (1).
    In conclusion, the LC50 of a transition metal at a specified
 hardness can be predicted with relatively minor error In the
hardness range 20-200 mg-L'1 as CaCCb by knowing the
LC50 at a different hardness and assuming (a) a slope of -1
for the relationship between ln(LCSO) and In(hardness) and
(b) alkalinity co varies with hardness. Such predictions might
be acceptable for regulating discharges of transition metals
to surface waters. But extrapolations to hardnesses outside
this range might drastically overpredict the toxicity of the
metal (e.g., extrapolating from an LCSO determined at a
hardness of -30 mg-L-1 as CaCQj to point A or B in Figure
3).Conversery(startingwithanLC50detenninedatareIatively
low or high hardness, extrapolations to midrange hardnesses
might drastically underpredict the toxicity of the metal. Such
errors might be magnified if the alkalinity does not change
proportional to the hardness in these extrapolations. Rather
than relying  on amechanistic tn(LCSO)  vs  In(hardness)
regression equations, a better approach might be to calculate
uptake of transition metals using a biotic-ligand model and
then predict toxidty from an empirical relationship between
mortality and the amount of accumulated .metal.

Acknowledgments

This research was funded by a subcontract to the University
of Wyoming under a grant from the International Copper
Assodation to HydroQuaL Inc., a subcontract to the Uni-
versity of Wyoming undera cooperative agreement between
the U.S. Environmental Protection Agency and the University
of Delaware, and a National Science Foundation EPSCoR
grant to the University of Wyoming. Eflen Axtmann assisted
with the geochemical spedation calculations. Comments
from Williams Adams, Herbert Allen, Paul Paquln, and two
anonymous reviewers improved the manuscript

Appenfu

Mass Balance on Metal in Solution. To construct a tractable
mass balance for a divalent transition-metal cation to which
fish are exposed. In a toxidty test. I make the following three
assumptions:
   1. No particles or organic ligands are present.
   2. Complexes of divalent transition-metal cations with
more than one HCQr, CQf~, or OH-  (e.g., CuOiOW,
CutCOjh*-. Cu(OH)j°, or Cu(OH),-) are negligible. Although
this is not correct at high pH and alkalinity, the complexes
can be ignored for heuristic purposes.
   3.At acutely lethal concentrations under realistic biomass
loadings (e.g.,  <10 g fish-L*1), the amount of the transition
metal that complexes with the fish gills will be a negligible
percentage of the total amount of dissolved metal remaining
in the water. For example, MacRae et aL (1 7) estimated 0.03
/
-------
                       (HCO,-
                                 .= 101MJ
                           IMOH+}
               *MHOy
                           IMHCOal
   Additionally, assuming the aqueous system is buffered
only by die hydroxyl and carbonate/bicarbonate systems,
the alkalinity (Alk, expressed as eq-L"') is (IB):
      Alk .-  01  + IHCOj'] +

   Substituting these relationships into eq A-l produces
 By combining eqs 5 (which expresses die LCSO of M1* as a
 function of ICa2*] and (Ca«gulMl) and A-2, die LCSO of die
-tuiul dissolved  metal diC&A^ri) can be expressed as a
 function of (a) die competitivTomding of cations at die gffl
 surface and (b) the equilibrium of die free-metal ion with
 die aqueous inorganic Kgands, as follows:
                                            .•ID-*"*
                                                 (A-4)
where JTand t^ are constants.
   Mass Balance on GDI Binding Sites. Assuming H*.Ca2>,
and M1* are die only cations of importance that bind to die
M-binding sites on a fish gill {although Mg*~ could be
                                                          substituted for or act in combination with Ca*+), an ap-
                                                          proximate mass balance for die M-binding sites on die gill
                                                          is
                                                                        (MogfflM) + ICaegillM] + [HegillM] +
                                                                                                       NgUIMI
                                                          or
                                                          tCasgUlMp
                         - (IMsgillM] + [HsgillM] +
                                       JagillMJ)  CA-5)
Lheralnre Ched

 (1) Borgmann, U. In Aquatic Toxlcotogr, Nriagu, I. O., Ed; John
    Wley and Sons: New York. 1983; pp 47-72.
 C2) Campbdl f.G.C.lnMetalSptctatfonaiuiBU)avallabiUtytn
    AjuatfcS/jMmnTessIer.A, Turner, D. R.,Eds.; John WBey and
    Sons: Chf Chester, 199&pp 45- 102.
 (3) Ambient Watr Quality Criteria for Cadmium - ISM. U.S.
   ' Environmental Protection Agency: Washington, DC, 1965: EPA
    440/5-84-032.
 (4) Ambient Water Quality Criteria for Copper  - 1984. U.S.
    Environmental Protection Agency: Washington, DC, 1985; EPA
    440/5*04-031    ;/
 (5) Ambient Water Quality Criteria for Nickel - 1986. US. Envi-
    iDiimentalProtectionAgenqr. Washington, DC, 1986; EPA440/
    5-86-004.        ."»•»-»
 (61 Ambient Water Quality Criteria for Lead - 1981. US. Environ-
    mental Protection Agency:  Washington, DC, 1985; EPA 440/
    S-84427.
 (7) Ambient Water Quality Criteria for Zinc- l967.U£.Envtson-
    mental Protection Agency:  Washington, DC 1387; EPA 440/
    5-874)03.
 (8) Pagenkopt. G. K. Environ. Set. TechnoL 1983, 17, 342-347.
 (9) Piay)e,RC;Dixon,D.&:ButnlsoaK.Can./.fIsA.A;uatSci
 •  1993,50,2678-2687.
(10) Meyer. ). &; Santore, a C; Bobbltt, J. P.; DeBrey, L IX: Boese,
    C LiPaouin. P. RjAlten. RE.: Bergman. H.L:DfToro. P.M.
                                                                                                           *
(11) Erickwn. R. k Benolt, D. A,: Mattson, V. R.; Nelson, H. P., IK
    Leonard, £ N. Environ. VudeoL Chem. 1996,15. 181-193.
(12) Allison, I. Dj Brawn, 0. &.: Novo-Gradac. K, I. tONTBQAS
    PKODEFA2.A(kochemicalAaettmentModetf»Environmental
    Systems Venten 3.0 Urn's Manuat US. Environmemal Pro-
    tection Agency:  Washington. DC 1991; EPA/600/3-91/021.
(13) NJSTCriOcalfySeleaeastabaityConftantsofMetalComplexes
    Database, Vertton 5.0. MST Standard Reference Databate 46;
    NationallnsaRiteofSiandardsandTechnokigy: Gaithenburg,
    MD. 1998.
(14) SerUz, S. M.: Allison, I. D.; Perdue, E M.; ADen, a E; Brown,
    D. S. Water Res. 1996,30,1930-1933.
(IS) ErtcksoaR.^Benolt,D.A;Manson,V.R.APre>wr/p«Taiidr/
    Facton Model for Site-specific Copper Water Quality Criteria;
    VS. Environmental Protection Agency: Duluth, MM,  1987
    (revised 1996).
(16) Schubauer-Beiigan,M.KjDierkes,J.R;Monson,P.D.:Ankley,
    G. T. Entiron. ToxicoL Chem.  1993,12,1261-1266.
(17) MadUe.RfcSmith,O.E:Swoboda-Coiberg,N.;Meyer,).5^
    Bergnian,RL£nrirottrojda>lO>«m.Acceptedforpubllcation.
(18) Pankow, I. F. Aquatic Chemistry Concepts Lewis: Chelsea, ML
 '   1991; p 171.


Received for review July 15, 1998. Revised manuscript re-
ceived November 30,1998. Accepted December IS, 1998.

ES980714Y
                                                           PAGE EST: 4.8   VOL «. NO.«. ntxx / ENVIRON. SO. ft TECHNOL * E

-------

-------
Gill  Surface Interaction Model for Trace-Metal ToxIcKy to Fishes:  Role of
Complexatlon,  pH, and Water Hardness          ,

Gordon K. Pagenkopf
Department of Chemistry, Montana State University, Bozeman, Montana 69717
•  A model has been developed to account for the varia-
bility in trace-metal toxicity to fishes at different values
of alkalinity, hardness, and pH.  The model utilizes
trace-metal speciation, gill surface interaction, and com-
petitive inhibition to predict effective toxicant concen-
tration (ETC). Copper; cadmium, lead, and zinc bioassay
data have been utilized.

  A review of the many research projects that have in-
vestigated trace-metal toxicity to fishes provides at least
three general conclusions: (1) for a particular trace metal,
some chemical species appear to be more toxic than others;
(2) the presence of elevated concentrations of the hardness
cations ions, Ca!+ and Mg2+, reduces trace-metal toxicity,
(3) LC50 concentrations vary from metal to metal. These
are not the only generalities of course, but they do provide
a basis for the development of a model that can account
for changes in toxicity as a function of pH, complexation
capacity, and hardness of the test waters. Currently there
is disagreement regarding the relative importance of water
     hardness and trace-metal complexation (1-4), This paper
     presents a model that combines both factors. What follows
     is an identification of. the chemical principles that arc
     believed necessary to couple trace-metal toxicity to pH.
     hardness, and trace-metal complexation. The identified
     principles are utilized to formulate quantitative relation-
     ships, and finally, predicted variation in trace-metal tos -
     icity is compared to that observed  in laboratory tests.

     Basis for Gill Surface Interaction Model (GSIM)
       The following are set forth as basic to the development
     of GSIM:
       (l)For acute toxicity to fish, trace metals alter the gill -
     function, and the fish die as a result of respiratory iro-'
     pairment.    .  .   • .,                           .
       (2) Of the trace-metal species present in a test water,
     some are significantly more toxic than others.
       (3) The gill surfaces are capable of forming complexes
     with the metal species and hydrogen ion present in the test
     waters.
342  Environ. Sci. Technot., Vol. 17. No. 6, 1983
0013-936X/83/0917-0342$01.50/0  © 1983 American Chemical Society

-------
   (4) The rates of metal exchange between the gill surfaces
 and test waters are fast when compared to the time re-
 quired for a bioassay test
   (5) The gill surfaces have a finite interaction capacity
 per unit weight.
   (6) Competitive inhibition exists between the hardness
 metals and the toxicants, which include the trace metals
 and hydrogen ion..
   A variety of experimental observations, both chemical
 and biological, will be utilized to substantiate the model

 Model Development
   Excessive mucous secretions, are often observed when
 fish are stressed by elevated concentrations of trace metals
 (5-7) and hydrogen ion (8).  In addition trace metals may
 be concentrated in the gill tissues (9,10) with the mech-
 anism of toxicity apparently being related to salt and water
 balances within the gill tissues (11,12). The physiology
 associated with the toxic action of trace metals is extremely
 complicated, and it is not the intent of this paper to discuss
 or even speculate as to the mode of action. This model
 utilizes competitive equilibria to predict changes in the'
 chemical activity of metal  species associated with gill
 surfaces. This associated ion activity correspondingly in-
 fluences the physiological function of the gills.
  Gill membranes consisting of phospholipids could pro-
 vide a surface of a net negative charge and the sites nec-
 essary for die formation' of Lewis acid-base complexes with
 the metal ions and hydrogen ion. The interaction may be
 classified as surface complexation (adsorption) or ab-
 sorption as long as exchange is rapid and it is reversible.
A schematic representation is shown in eq 1 with copper
                           C.
             Cu2* + 35=8"- —
(1).
 as the Lewis acid.  The surface is represented by ^S"~,
 which designates a group of Lewis base sites that collec-
 tively is capable of forming surface complexes with the
 metal species.  This approach permits application of
 chemical principles that have been successfully utilized in
 the interpretation of more precisely defined and controlled
 chemical systems. The surface complex is designated by
 MJCu""*2, and the equilibrium constant for the interaction
 is KCU- Utilizing the condition that complexation reactions
 of this type are rapid, an equilibrium expression maybe
 written'

          , KC* = NSCu-^l/ICu'+lMS"-)        (2)

 where braces  and brackets designate concentration of
 moles per kilogram and moles per liter,  respectively.
 Rearrangement of eq 2 provides

            |=SCu-"«l • ^c«HSn-|tCu2+]      ..  (3)

 which identifies a linear relationship between the con-
 centration of the surface complex and the concentration
 of Cu2* in solution, provided the concentration of sS"~
 remains constant A relationship of this type is in agree-
 ment with experimental observation where an increase in
total toxicant concentrations decreases survival time (13).
Implicit is the fact that a small fraction of the complexa-
tion sites  is occupied by Cu2*, and  thus l^sS*-) remains
essentially constant within the experimental uncertainties
of the bioassay test.           .
  Trace-metal speciation in natural water systems is de-
pendent upon pH  and which  completing  Uganda are
present.   Most bioassay test waters contain minimal
amounts of organic material, and thus these complexes will
not be considered.  A majority of the complexes involve
hydroxide and inorganic  carbon. Each of these species
         constitutes some finite fraction of the total trace metal in
         solution.  For Cu2+ this fraction, designated by «&,", is
         defined as
                        [Cu2*J/fCuTJ
                                                         (4)
        where CuT is total copper in solution.  Procedures for the
        evaluation of a values have been presented elsewhere (14)
        Combination of eq 3 and 4 provides
                                                         (5)
        which establishes the concentration of the surface complex
        in terms of speciation.
          Bioassay tests indicate that other species may be toxic
        and need to be included (2). For copper, these include
        CuOH*, Cu(OH)s.aq,  and possibily CujOHj*. All are
        capable of forming surface complexes with the strength
        of the interaction being species dependent Generalization
        of eq 5 provides

                                                        (6)
    - - - ,  — ^— M -_!»*«*«**» «B«W W*ftV**ABW«kMWAJQ VI U1Q BtUJ,
 complex for the ith copper species. Similarly K^ and o^
 are the surface complexation equilibrium constants and
 the fractions of total copper present as the ith species. An
 equation similar to eq 6 is applicable for the other trace
 metals.
   There are reports (I, 3,4) indicating that an increase
 in water hardness increases fish resistance to trace metals.
 Test  water concentrations  for the common hardness
 metals, calcium and magnesium, range from approximately
 10 to 1000 mg/L as CaCO3.  The corresponding concen-
 trations would be ICTMCT2 M. For these metals to exhibit
 a pronounced protective effect, their concentration gen-
 erally has to be greater than 10~3 M. These metals form
 few complexes with the ligands found in most test waters,
 and therefore the equated metal concentration is essen-
 tially equal to the total metal concentration.
  The GSIM model interprets the protective action of the
 hardness cations as a competition between these metal ions
 and the toxic species, in essence competitive inhibition.
 The total number of surface interaction sites is given by
eq 7, where ssS"- are the free sites, s=S(M)-*+* are those

ST » ssS»- + sSfM)"*2 + ssSCH)-"*1 + ssS(TM)    (7)
                                                        occupied by the hardness cations, ssS(H)~"+l are those
                                                        occupied by hydrogen ion, and =sS(TM) are those occupied
                                                        by the trace-metal species.  Charges for *=S(TM) are
                                                        omitted, and since Ca2* and Mg2+ exhibit similar chem-
                                                        istries, they are not differentiated.
                                                          For test waters of pH 6 or greater there appears to be
                                                        little hydrogen ion dependence, and with concentrations
                                                        of hardness cations many orders of magnitude greater than
                                                        the concentrations of the toxic trace-metal species, it is
                                                        assumed that (=S"- + ssSjM)-"*2) » (sSfH)-"*1 + ^S-
                                                        (TM)). The equilibrium expression for the hardness metal
                                                        interaction is

                                                                  KM - |^S(M)-"+2}/(lMi+]|^S''1)        (8)

                                                        Substitution into the simplified form of eq 7 with rear-
                                                        rangement provides
                                            - CIF
                                                                                                         (9)
                                                        which is designated as the competitive interaction factor,
                                                        CIF. Substitution into eq 6 provides a relationship suitable
                                                        for the interpretation of the copper toxicity data:
       NSCu,| -
                                                                                       + KM(M2+])     (10)
                                                                     Environ. Sci. Technol., Vol. 17,' No. 6. 1983  343

-------
         Table I.  Species Concentrations, Hardness, and ETC
         Values for Rainbow Trout (/)
                                                        Table II.  Predicted ETC Values for Cutthroat Trout
i
hard-
ness,
test
no.
1
2
3
4
5
6
7
8
9
10
11

«Cu =
(1 + *M!
[CuOH*]

P
6
6
6
7
7
8
8
8
9
9
9


>H
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0

Cu,«
Mg/L
22.2
39.5
82.2
32.5
137
16.2
138.6
83.1
14.8
35.2
16.0

mg/L as
CaCO
32
101
371
101
298
31
371
360
30
98
364

[Cu»*] + [CuOH»] =
EM»»]),«M = 6X 10
+ (Cu(OHVaql).

, CIF"



ETC,* ng/L
0.44
0.17
0.08
0.
0.
0.
0.
0.
0.
0.
0.

17
06
44
05
05
44
17
06









av
9.77
6.72
6.58
5.50
8.22
7.13
6.93
4.16
6.51
5.98
0.8











6.21 1 2.29
(Cu(OHVaq). *CIF =
'. cETC=ClF([Cu»»] +
11
         Defining {^^/((Kc^JN&r)) as die effective toxicant
         concentration, ETC, eq 10 becomes
 ETC • [Curta*/(l
                                           CIF[toxic species]
                                                  ..   (U)
         Application of GSfM
           Copper. Chemical speciation has successfully accounted
         for variation in toxicity of some metals (2,4, 15-17). In
         these studies the total metal concentrations is often much
         greater than the concentration of the most toxic species.
         Equation 11 is capable of accounting for the speciation;
         however, a value for Ku has to be assigned before quan-
         titative interpretation of the hardness can be presented.
         An empirical estimate of 5 x 10s comes from observed
         copper toxicity. The concentration of the hardness metals
         has to be greater than 10~3 M before a pronounced effect
         is observed.
           The data presented by Miller and Mackay (3) provides
         a way .to estimate this constant. They determined copper
         toxicity as a  function of hardness at constant pH and
         complexation capacity. As a consequence CuT may be
         compared directly. From these comparisons KM is calcu-
         lated to be (7 ± 5) x 10*.
           Equation 11 indicates that ETC should have a constant
         value for a given test species and constant time of exposure.
         For copper
ETCCu =
                                lCuOH+] + [Cu(OH)s.aq])  (12)
         Data presented by Howarth and Sprague (1) include a
         change in species  distribution as well as hardness (see
         Table I). The mean ETC^, value for rainbow trout is 6.21
         Mg/L with a standard deviation of 2.29.
           An extensive study utilizing cuttthroat trout provides
         another system for the application of GSIM (4).  Three
         hardness and three alkalinity concentrations were em-
         ployed, resulting in nine combinations. The toxic species
         are considered to be Cuz+, CuOH+, and Cu(OH)2-aq, and
         the fraction of the interaction surface available is regulated
         by the hardness metal ion concentrations. The values of
         ETCcu are listed in Table U  For comparison, the average
         96-h ETCcu for rainbow trout is 6.21 Mg/L, whereas the
         comparable value for cutthroat is 2.72 Mg/L.
           Copper species distribution and ETC values have been
         calculated for the  Miller and Mackay data  (3), and the
hardness/
alkalinity,
mg/L as
CaCO,
205/178
205/77.9
160/26.0
70/174
70/70
74.3/22.7
18/183
18/78.3
26.4/20.1


PH
7.73
7.61
7.53
8.54
7.40
7.57
8.07
8.32
7.64


GIF
0.089
0.089
0.11
0.22
0:22
0.21
0.53
.0.53
0.43

Cu,*
Mg/L
18.5
27.4
27.4
, 9.2
25.2
14.2
\ 1.53
7.09
5.21


ETC, ng/L
1.65
2.44
3.04
2.02
5.54
2.98
0.81
3.75
2.24
                                                                                                    av 2.72 ± 1.33

                                                                   0 Data from ref 4.96-h LC60. * Cu = Cu** +• CuOH* +
                                                                 Cu(OH),-aq, species distribution calculated by using for
                                                                 CuOHMogK,:          "
                                                                 96-h LCSO.
                                                                      6.48 and tor Cu(OH), log ff, = 11.78 (18),
                                                                 Table III. Predicted ETC Values for Rainbow Trout and
                                                                 Fathead Minnows .
hardness/
alkalinity,
mg/L as
CaCO,
PH
CIF
Cu,
Mg/L
ETC, Mg/L
ref
rainbow trout
12/10°
99/10°
49/28°
98/28°
12/51°
97/51°
v
300/205*
'7.1
7.0
7.3
7.2
7.4
7.3

7.35
0.62
0.17,
0.36
0.17
0.62
0.17

0.0625
12.7
38.0
17.0
30.5
3.6
21.7
av
62.2
7.87
6.52
6.10
5.18
2.22
3.69
5.26 i 1.89
3.88
3
3
3
3
3
3

19
fathead minnow
198/161*
31.4/15*
360/150*
20/9*
7.9
7.2
8.2
7.5
0.092
0.39
0.053
0.50
45.9
39.7
70.5
13.2
4.22
15.5 -
3.7
6.6
20
21
22
22
                               av 7.53 i 5.48
 " 15-day LCSO values. » 96-h LCSO values. 	


results are summarized in Table ffl.  Values for the 15-day
test are somewhat less than those observed for the 96-h
test with rainblows, 5.28 vs. 6.21 Mg/L.  The order of re-
sistance of the test fish to copper appears to be fathead
minnows > rainbow trout > cutthroat trout.
  The data presented in Tables I-III include a variation
in pH from 6 to 9, an alkalinity variation from 10 to 205
mg/L as CaC03, a hardness variation from 12 to 371 mg/L
as CaC03, and a variation in total copper by more than
a factor of 100. Application of the GSIM to these data has
identified an effective toxicant concentration for each test
animal. The variability in the predicted ETC values is
equal to or less than the observed experimental variability.
  This model is based on the premise that trace-metal
speciea bound to the gill surfaces cause impairment of
physiological functions. The amount of trace metal bound
is regulated by the chemical composition of the test waters.
Specifically a competition exists between the hardness
metals and the toxic species for interaction sites.
,  Zinc. The coordination chemistry of zinc is similar to
copper in many respects; however, the thermodynamic
stability constants are generally not as large. Zinc spec-
iation for a number of bioassay studies has been completed,
and a correlation exists between the sum of the concen-
         344  Environ. Sei. Techno!., Vol. 17. No. 6. 1983

-------
  Table IV.  Zinc Toxicity to Brook Trout and
  Rainbow Trout (23)   '   .
fish
size
mean
wgt, g
3.0
3.0
3.9
3.9
19.0
19.0
3.9
3.9
4.9
4.9
28.4
28.4
°96-h
hardness/
alkalinity,
mg/L as ZnT «
CaCO, pH mg/L
brook trout
46.8/41.8 7.63 1.55
177.6/170.2 7.41 6.14
47.0/42.8 7.58 2.12
179.0/170.1 7.17 6.98
44.4/42.5 7.38 2.42
169.7/43.0 ,7.31 4.98,
rainbow trout
46.8/41.8 7.63 0.370
177.6/170.2 7.41 2.51
47.0/42.8 7.58 0.517
179.0/170.1 7.17 2.96
44.4/42.5 7.38 0.756
169.7/43.0 7.31 1.91
LC50 values. * Zn = Zn* +
Zn,*
mg/L GIF
1.42 0.30
4.83 0.10
1.94 0.30
5.91 0.10
2.27 0.31
.4.71 0.11
0.339 0.30
1.96 0.10
0.475 0.30
2.56 0.10
0.712 0.31
1.81. 0.11
ZnOH*.
ETC,
mg/L
0.43
0.48
0.58
0.59
0.70
0.44
0.10
0.20
0.14
0.26
0.22
0.19

Table V. Cadmium Toxicity to Rainbow Trout (25?
hard-
ness,
mg/L as CdT, Cd,*
CaCO, mg/L mg/L CIF ETC, mg/L
20 0.091 0.083 0.50 0.042
80 0.358 0.262 0.20 0.052
320 3.69 1.66 0.058 0.096
320 0.677 0.618 0.058 0.036
  « pH 7.2, 48-h LC50.
Cd(OH),-aq.
Cd =
                                  av  0.056 ± 0.027

                                 + CdOH* +
 tnttions of Zn*+ and ZnOH* and the observed ferocity (I ?)•
 Utilization of the procedures outlined for copper provide
           ETCa, = dFtfZn8*] + [ZnOH*])       (13)

 A recent report by Holcoinbe and Andrew (23) presents
 additional zinc toxicity data for brook trout and rainbow,
 trout. This study was designed to test the influence of
 hardness.  The conditions were well regulated, and the
 results are presented in Table IV.   '
   Lead. A chronic bioassay study by Davies et al (24) has
 established that the maximum acceptable toxicant con-
 centration (MATC) for rainbow trout in hard water should
 range from 18.2 to 21.7, Mg/L dissolved lead. In soft water
 the MATC ranges from 4.1 to 7.6 Mg/L. Species distri-
 butions for the test waters indicate that soluble lead was
 dominated by Pb2* and PbOH+. The hardness of the test
 waters was 353 and  28  mg/L as CaC03, respectively.
 Calculated CIF values are 0.054 and 0.42, which indicates
 that soluble lead should be 7.8 times more effective in soft
 water. The MATC ratio is 4.3 with means and ranges from
 2.4 to 7.7 by use of the extreme values. A majority of the
                                   lead studies are of static design and there is sizable dif-
                                   ference between total lead and soluble lead.  -
                                    Cadmium.  A study designed to test theJnfluence of
                                 •  hardness on cadmium toxicity to rainbow trout (25) has
                                   been reported.  There is major variation in  hardness,
                                   20-320 mg/L as CaCO3, and the alkalinity of the waters
                                   was not reported.  Test water conditions were reported
                                   earlier (19), and these values were utilized to calculate the
                                   speciation. The results are summarized  in Table V.
                                    Given the range of hardness 20-320 mg/L and the range
                                   of total cadmium 0.091-3.69 mg/L (a factor of 40), the
                                   agreement is considered to be very good.
                                    Combination of Metals. The interpretation of metal
                                   toxicity resulting from the mixture of metals is of sizable
                                  current interest and will probably become more critical as
                                  tame goes on. The terminology put forth by Sprague (26,
                                  27) a useful for discussion purposes. When  the toxicity
                                  of a mixture corresponds to the sum of the fractions of the
                                  single components, the effect is referred to as "additive".
                                  When the effect is greater than or less than,  a "more-
                                  than-additive"  or "less-than-additive" effect is assigned.
                                    The GSIM assumes an additive effect  This results from
                                  the linear or near linear relationship between concentration
                                  ftf th« filirfa^A anAA^A^ «•**! *U- w—_•	LJ»-	'-..:.  *•
          	' ~~I——-™' *-*«* w**« «*w**v0|/wuuiu£ vuuvcuwciVKJni
 of these species in solution. To apply the model, the toxic
 unit concept (27) needs to be utilized.  This may be ac-
 complished through ratios of the ETC values for the re-
 spective metals.
   A paper by Eaton (28) reports 72- and 96-h LC50 data
 for mixtures of Cu,Cd, and Zn. The data are summarized
 in/Table VL  In a parallel study (28) it was observed that
 5030 Mg/L total zinc was required for the 96-h LC50. The
 pH, alkalinity, and hardness are the same, and thus 0.41
 toxic units are assigned to zinc. This value is equal to the
 ratio of the zinc required in the mixed-metal study to that
 required when zinc was the only toxicant (2050/5030).
 Another study from the same laboratory (20) reported a
 96-h LC50 total copper  value of 430 Mg/L.  From these
 values 0.36 toxic units are assigned to copper (154/430).
 The contribution from cadmium is difficult to assign since
 the acute  study for cadmium and fathead minnows re-
 ported the presence of sizable amounts of insoluble cad-
 mium salts (29). Using the lowest total dissolved cadmium
 concentration, 1400 pg/L, a hardness of 201 mg/L as
 CaCO,, a  pH of 7.7, and an alkalinity of 161 mg/L as
 CaCO,, 118 Mg/L is predicted  for the  Cd ETC value.
 Coupled with the value in Table VI, 0.22 toxic units are
 assigned to cadmium (26.3/118). Summation of the re-
 spective contributions provides a value near unity and may
 be significant.  There is  sizable uncertainty in the cad-
 mium contribution, however.
  Other trace-metal mixture studies (30,31) indicate an
 additive response.  There are insufficient chemical data
 available to do chemical speciation, and thus ETC cannot
be calculated.  In the cases where pH, alkalinity, and
hardness remain fairly constant, total conceptions may
be utilized as a measure of fish response.
Table VL Toxicity of Cu, Cd, and Zn Mixture to Fathead

metal
Cu
Cd
Zn
metal coricn.
Mg/L (total)
154
,320
2050

TM,6, Mg/L
9^52
299
1517
Minnows" (28)

CIF
0.088
0.088
0.088


ETC, Mg/L
6.84
26.3
133


TU
0.36
*" 0.22 r
0:41
  0 96-h results, alkalinity = 154 mg/L as CaCO,, hardness = 207 mg/L as CaCO,. pH 7 7  * Cu
Cu(OH)iaq;Zn=Zn" + ZnOH';Cd = Cd'* + CdOH'.                     '
                                                                                        sum  -0.99
                                                                      Environ. Set. Techfiol.. Vol. 17, No. 6. 1983  345

-------
   A sequence of papers (32-34) has reported the toxkity
of mixtures to guppies. In these experiments copper and
nickel appear additive, whereas copper and zinc appear
more than additive.  For the zinc studies (32), a 96-h LC60
value of 6.76 mg/L is reported. This value exceeds the
calculated zinc solubility of 3,2 mg/L. A lower concen-
tration of zinc would correspondingly decrease the degree
of apparent enhanced toxirity due to the mixture of copper
and zinc.
   Hydrogen Ion. It is difficult to precisely define the
tolerance of fish toward elevated hydrogen ion concen-
trations since  it appears to be dependent  upon many
variables. One report indicates that pH values near 5 may
represent a lower tolerance limit, however (35).  In addi-
tion, alteration of gill surface tissues is observed at pH
values below 5.2 (56). The concentration ranges where the
respective metals exert an influence, [Ca]  at 10"8 M; [Cu]
« 10-«-l
     33, 2023-30.
 (13) EPA-440/9-76-023, U.S. Environmental Protection Agency,
     Washington. DC. 1976.
 (14) Pagenkopf,  G. K.  "Introduction to Natural Water
     Chemistry'; Marcel Dekker. New York, 1978.
 (16) Andrew, R. W.; Biesinger, K. £.; Glass, G. E. Water Re*.
     1977,11, 309-15.
 (16) Sunda, W. G.; Engel, D. W.; Thuotte, R. M. Environ. Sci.
     Technol. 1978,12, 409-413.
 (17) Pagenkopf, G. K.  In "Zinc in the Environment, Part 2:
     Health Effects"; Nriagu, J. O., Ed.; Wiley-Interacienee: New
     York, 1980; pp 353-361.
 (18) Sunda, W. G.; Hanson, P. J. In "Chemical Modeling in
     Aqueous Systems"; Jeanne, E A., Ed.; American Chemical
     Society: Washington, DC, ACS Symp. Ser. No. 93.
 (19) Calamari, D.; Marchetti, R. Water Res. 1973,7,1453-64.
 (20) Mount, D. I. Water Res. 1968,2, 215-223.
 (21) Mount, D. I.; Stephen, C. E. J. Fish Res. Board Can. 1969,
     26,2449-2457.
 (22) Pickering, Q. H.; Henderson, C. Air Water Pollut.Int. J.
     1966,/O, 453-463.
 (23) Holcombe, G. W.; Andrew, R. W. EPA-600/3-78-094, U.S.,
     Environmental Protection Agency,'Washington, DC, 1978.
 (24) Davies, P. H.; Goettl, J. P., Jr.; Sinley, J. R.; Smith, N. F.
     Water Res. ,1976,10,199-206.
 (25) Calamari, D.; Marchetti, R.; Vailatis, G. Water Res. 1980,
     14, 1421-1426.
 (26) Sprague, J. B. Water Res. 1969, 3, 793-821.
 (27) Sprague, J. B. Water Res. 1970,4, 3-32.
 (28) Eaton, J. G. Water Res. 1973, 7,1723-1736.
 (29) Pickering, Q. H.; Cast, M. H. J. Fish Res. Board Can. 1972,
     29, 1099-1106.
346  Environ. Scir Technot., Vol. 17, No. 6. 1983

-------
                                          Environ. Sci. Tectmol. 1683. 17. 347-352
(30) Brown, V. M. Water Res. 1968, 2, 723-733.
(31) Brown, V. M.; Dalton, R. A. J. fish Biol. 1970,2,211-216.
(32) Anderson, P. D.; Weber, L. J. Toxicol. Appl. Pharmacol.
     1975, 33, 471-483.
(33), Anderson, P. D.; Weber, L. J. Proc. Int. Con/. Heavy Met.
     1973. 2, 933-953.
(34) ^Muska, C. F.; Weber, L. J. EPA 600/3-77-085, U.S. Envi-
     ronmental Protection Agency, Washington, DC, 1977, p
     71-87.
(35) Jones, J. R. E. "Pish and River Pollution"; Butterworths:
     London, 1964; 107-116.
(36) Daye, P. G.; Garside, E. T. Can J. Zool 1976,54,2140-55.
 (37)  Zitko, V. Proceedings of Toricity to Biota of Metal Forms
      in Natural Water, International Joint Commission, 1976
      pp 9-32.
 (38)  Zitko, V.; Carson, W. V.; Carson, W. G, Bull. Enoiron.
      Contam. Toxicol. 1973,10, 265.
 (39)  Perrin, D. D.; Sayce, I.  G. Talanta 1967,14, 833-42
 (40)  Harriss, D. K.; Ingle, S. E.; Magnuson, V. R.; Taylor, D. K.
      REDEQL-UMD, Department of Chemistry,  University of-
      Minnesota—Duluth, 1982.

Received for review November 25, mi. Revised manuscript
received October 25,1982.  Accepted February 17, /8S3.

-------
      Copper and Cadmium  Binding to Fish  Gills:  Modification  by

               Dissolved Organic  Carbon and  Synthetic  Ligands


                                  Richard C. Playle1 and D. George Dixon

                          Department of Biology, University of Waterloo, Waterloo, ON N2L 3G1, Canada

                                              and KentBurnison

                      National Water Research Institute, Environment Canada, Burlington, ON L7K 4A6, Canada

          Playle, R.C., D.G. Dixon, and K. Bumison.  1993. Copper and cadmium binding to fish gijls: modification by
                 dissolved organic carbon and synthetic ligands. Can. J. Fish. Aquat Sci. 50:2667-2677.
          Adult fathead minnows (Pimephales promelas) were exposed to 17 jig Cu • L"1 or 6 jig Cd • L"1 for 2 to 3 h in
          synthetic softwater solutions at pH 6.2 containing either naturally-occurring, freeze-drled dissolved organic carbon—f
          (DOC) or synthetic ligands such as EDTA. After exposures, gills were assayed for bound Cu or Cd.yAsa first I
          approximation,
          concentration c
         "relatively low'
          suggesting different gill binding sitesJCadmium accumulation  on gills was~more sensitive to increased'
         'concentrations of Ca and H+ than was Cu. Surprisingly, Cd bound to gills to the same or greater extent than
          did Cu: for synthetic ligands, Cd binds less well than Cu. This result corroborates previously published
          observations that Cd, unlike Cu,' is taken up at gills through high affinity Ca channels. Accumulation of Cd
          on fish gills was never associated with 14C-labeTled EDTA or 14C-citrate, indicating that free metal interacts
          with the gill while metal-ligand complexes usually do-not.

          Des adultes du T&e-de-boule (Pimephalespromelas) ont £t£ exposes a Cu en concentration de 17 tig • L~' ou du
          Cd en concentration de 6 tig • l~1 pendant 2 ou 3 h dans des solutions d'eau douce synth^tiquel pH 6,2 qiii
          contenaient du  carbone organique dissous  (COD) lypphilise' et qu'on trouve naturellement,  ou  des ligands
          synth&iques comme I'EDTA. La pe*riode d'exposition terminee, le Cu ou le Cd fixe" sur les branchies, eteit litre". En
          premiere approximation, le lac d'origine ou la taille moleculaire du COD n'agissait pas sur la fixation du Cu aux
          branchies, alors que la concentration  du COD avait un effet A une concentration as 4,8 mg  • L~1, le Cu ne
          s'accumulait pas sur les branchies des TStes-de-boules. Aux assez faibles concentrations testees, ni le Cu nt le Cd
          interfeVait avec la fixation des autres m&aux sur les branchies; cela donne a'Denser qu'il existe different sites de
          fixation sur les branchies. Plus que le Cu, ('accumulation de cadmium sur les branchies £tait davantage sensible
          aux hausses de concentration du Ca et du H+. Fait surprenant; le Cd se fixait aux branchies autant ou rrieme plus
          que le Cu : avec les ligands synth&iques, le Cd se fixe moins bien que le Cu. Ce resultat vient confirmer des
          observations dont il a deja 6t£ fait mention, a I'effet que le Cd, contrairement au Cu, est absorb^ au niveau des
          branchies par des voies a tres forte affinite1 pour le Ca. (.'accumulation de Cd sur les branchies des poissons n'a
          jamais &6 associee a I'EDTA marque* au' 4C ou au citrate marque1 au14C; cela indique que les atomes m&alliques
          libres passent en interaction avec les tissus des branchies alors que les complexes de m&aux-ligands ne le font
          habituellement pas.
          Received September 18, 1992
          Accepted May 18,1993
          (JB634)
                      Refu le Wseptembre 1992
                         Accept^ lei 8 mai 1993
          Waterbome  metals  generally show  their greatest
          toxicity to aquatic organisms in soft water of low
          alkalinity, low pH, and low dissolved organic carbon
(Spraguc 1987; Hemming and Trevors 1989; Spry and Wiener
1991).  This  relationship, can  be explained  by the usual
conceptual model of metal toxicity, which considers the free ion
to be the toxic form of the metal (Morel 1983; Pagenkopf 1983).
In essence, waterborae free metal ions must adsorb to gills before
they can either exert their toxic effect at the gilt surface directly,
or pass through the gills on their way to internal sites of toxic
   'Author to whom correspondence should be addressed. Current
address: Department of Biology, Wilfrid Laurier University,
75 University Ave. West, ON  N2L  3C5, Canada.        ,

Can. J. Fish. Aquat. Sci., \bl. 50.1993&.
action (Sirnkiss and Taylor 1989). Any process that prevents
initial adsorption on the gill surface by reducing either the
ambient free metal ion concentration, or the number of surface
binding  sites on the gill, will reduce toxicity of waterbome
metals.   -
  A number of conditions can affect the degree to which metals
bind to gill surfaces. For instance, hardness cations such as Ca2+
and Mg2+ compete for  metal  ion  binding sites,  and  metal
complexing agents such as dissolved organic carbon (DOC) may
reduce the free metal available to interact with an organism
(Morel 1983; Pagenkopf 1983). Low water pH may also affect
metal binding through H+ competition for binding sites, and by
affecting metal speciation (Campbell and Stokes  1985). In

                                                    2667

-------
 addition, H+ can displace metals from DOC, increasing available
 free metal ions (Cabaniss.and Shuman 1988a). A number of
 recent studies  using fish have shown  that  tbxicity  or
 accumulation of both Cu and Cd varies according to competition
 andcomplexation(forCu:Cusimanoetal. 1986; Hutchinson and
 Sprague 1986, 1987;  Playle et al. 1992; for Cd:  PSrt and
 Wikmark 1984;Cusimanoetal. 1986; Wicklund and Runn 1988; ,
 Pratapetal. 1989; Bentley 1991).
   In fish,  waterbome  Cu and Cd affect ion uptake through
 interactions at the  gill. Copper (12-50 jig • L'1) affects ion
 balance by reducing Na* and Cl~ influx at gills (Laurgn and
 McDonald 1985), mainly through effects on Na+-K+-ATPase
 (Lauren and McDonald  1987).  High  Cu  concentrations
 (200 jig • L"1) increase gill permeability, as shown by increased
 ion efflux, probably because of displacement of Ca from inter-
 cellular tight junctions (Lauren and McDonald 1985). Cadmium
 also affects ionoregulation, but through interference with Ca
 uptake at the gill,  largely through effects  on a high affinity
 Ca^-ATPase  (Verbost et al.  1989).  At Cd  concentrations
 
(=0.27 uM). Backgrounded was<0.1(33) fig • L-';withaddi;d
Cd (as CdCl2 • 2|HA Baker Chemical Co., Phillipsburg; NJ) die
concentration was 5.5 ± 0.1 (63) fig ' L-|(=0.05|jiM).AHhoujih
these Cu and Cd concentrations are ~10X higher than those
usually found in soft acidified surface waters (Hutchinson and
Sprague 1986; Johnson 1991; Spry and Wiener 1991), they gaj/e
measurable metal accumulation on gills within 2-3 h of exposure
(see Results). For one metal mixture experiment, zinc was add0.2 am. The DOC fractions were then freeze dried and storei^P
at room temperature in  polyethylene containers. The DOC'
fractions were converted to the soluble sodium form in Mill! ('[
(Millipore) water in the presence of AG50W-X8 cation exchange
resin (Na form, BioRad, Richmond, CA). After 20 h the solution!;
were filtered through 0.2 jim polycarbonate filters (Nuclepon!
Corp., Pleasanton, CA). The concentrated filtrate was added
directly to the aerated synthetic soft water used in the metal
exposure experiments.   Water DOC  concentrations  wen;
measured, after acidification and sparging to remove inorganic
carbon, using a Beckman model 915B total carbon analyzer. W«:
used DOC concentrations of <8 mg • L~' (= <8 mg C • L~'), thai;
are representative of soft lake waters (Spry and Wiener 1991). ||
  In experiments to determine metal-gill binding constants, we
used the  ligands  ethylenediaminetetraacetic  acid  (EDTA;
disodium salt; Baker analyzed reagent), nitrilotriacetic acid
(NTA; disodium salt, Sigma Chem. Co., St. Louis, MO), ethyl
enediamine (EN; Sigma), citric acid (AnalaR, BDH), glutamic
acid (Baker), oxalic acid (AnalaR, BDH), and salicylic acid
(Fisher). These ligands were chosen to ensure a wide  range oi
metal complexation strengths. Ligands were prepared as 1 nuVj
stock solutions in deionized water. Calculations of free and
ligand-bound metal concentrations were made using MINEQL
(version 2:1; Schecner 1991) on an IBM-compatible computer,
using measured pH (fixed) and water chemical concentrations.
Equilibrium stability constants at zero ionic strength (Km given
as K in this paper) were converted to values appropriate to thejg^
ionic strength of our water (/ ~ 1 X 10~4)  by the computelpr
program. Equilibrium constants were .from the MINEQL*
program, and are summarized in Table 1.
  At the end of each experiment the fish were killed with a blow
to. the head, both sets of gills (= gill baskets) were excised, anc
the gills were rinsed for 10 s in 100 mL synthetic soft water. Tlw
gills were then weighed (mean wet weight=0.07 g) and digestec

                            Can. J. Fish. Aqaat. Sci.. Vbi. SO. 199.

-------
                 TABLE 1. Log of equilibrium constants (1:1 binding to ligands) from the MINEQL+ computer program.
                 EN = ethylenediamine.  	•	'	''	    . .
                              EDTA
NTA
EN
Citrate     Oxalate   Glutamate    Salicylale
Cu-Kgand
Cd-ligand
Ca-ligand
H-ligand
18.8
16.3
12.5
10.0
13.1
9.4
8.2
10.3
10.5
5.6
0.1
10.0
7.3
5.3
4.7
6.3
5.1
3.4
3.6
4.2
8.3
4.8-
2.5
10.0
10.6
6.7
— -
13.4
for 8 h at 80°C in SX their weight of 1 N H2S04. Gill digests
were vortexed, centrifuged, and die supematent diluted 20X
with  deionized  water.  For  the  NTA-Cd and citrate-Cd
experiments, run 6 mo after all other experiments, 1 N HjSO4
added to gill samples was 10X the gill weight, with a later 10X
dilution with deionized water. For all experiments, gill and water
Cu and Cd concentrations were measured by graphite furnace
atomic absorption spectroscopy (AAS; Varian AA-1275 with
GTA-95 atomizer) using 10-uJL injection volumes, N2 gas, and
standard operating conditions documented by Varian. Water Na
and Ca concentrations were measured using a Perkin Elmer 4000
AAS, Zn concentrations using graphite furnace AAS, and pH by
a  Radiometer PHM82 meter  with  GK2401C combination
electrode/
  We compared  Cu and  Cd concentrations on gills of  fish
exposed to modified soft water with those of all fish exposed to
metals in standard soft water conditions (grey horizontal bands
in Figures) and  with those of all unexposed fish (dashed
horizontal band in Figures). Comparisons of gill metal concen-
trations  were made using ANOVA followed by the Tukey-
Kramer procedure. Data are plotted in the Figures  as means
±95%CI.

14C-Ligand Experiments
  Experiments were run to determine the contribution of free
versus ligand-bound metal to die metal load offish gills. If only
free metal binds to gills, then the metal alone would be detected
in gill samples. If metal complexed by a radiolabelled ligand
binds with gills, then both die metal and radiolabel would be
detected. We used  "C-EDTA  (tetrapotassium salt)  and
l,5-l4C-citrate for these experiments (Sigma; specific activity
11.8 and 60 mCi • mmoJ-', respectively). Cadmium was chosen
as die  test metal because less is  needed for detectable
accumulation in fathead gills, compared to Cu, using graphite
furnace AAS.                •
  The I4C experiments were run in a similar manner to diose
outlined above. Two or 3 minnows were placed in 1 L of
synthetic unaerated soft water for 2-3 h. The exposure water
contained 0 or 0.05 ujnol Cd (6 ng-L"1), 0 or  0.1 junol
I4C-EDTA, or 0 to 250 junol mixtures of "C-citrate and "cold"
citrate. We aimed for ~107 disintegrations per minute (DPM) of
I4C • L~', to yield reasonable counts in die gill samples. Gills
were excised as before, weighed, and digested for 8 h at 80°C in
10X their weight of 1 N H2SO4. Digests  were vortexed and
100 |iL was  diluted 10X with deionized water before Cd
analysis  by  graphite furnace  AAS. Remaining  solutions
(0.4-1 mL) were added to 16 mL of Cytoscint ES fluor (1CN
Biomedicals Inc., Irvine, CA) in 20-mL plastic scintillation vials
(Fisher). The vials were shaken, and tiien counted in a Beckman
LS 1701 liquid  scintillation counter  with automatic quench
correction. The I4C label in water was determined by counting a
1 -mL water sample in 16 mL of fluor.

Can. J. Fish. Aquat. Sci.. VW. 5ft 1993
                    To test detection of 14C in die gill samples,Nfive gill samples
                  giving background counts were  spiked with  10-200 uL of
                  -2 X 10s DPM "C-citrate • L~' solution. In addition, "C-labelled
                  anduacene (Sigma), a lipophilic polyaromatic hydrocarbon, was
                  used to measure the uptake of a l4C-compound known to cross
                  biological. membranes.   Ten   tnicrolitres   of  0.1   u,Ci
                  "C-andiracene • nL~' (=0.07 |unol)  in  dimethyl sulphoxide
                  carrier (DMSO; Omnisolv, BDH) was added to 1L of synthetic
                  soft water, to yield ~2 X 106 DPM • L~'. Three minnows were
                  held in diis water for 2 to 3 h. Ten microlitres of DMSO in 1 L
                  was tested widi '^C-citrate (-4 X  10* DPM • L-1) to determine
                  if DMSO contributes to citrate entry into gills. Gill and water
                  samples  were counted as before.
                    Because citric acid is a substrate in die tricarboxylic acid cycle,
                  it was possible  dial "C-citrate bound by fish gills could be
                  metabolized to I4CO, and expired. This possibility was tested
                  using  250-mL  solutions  in  sealed  jars  with  and without
                  "C-citrate, Cd, and fish. Water in die jars was acidified after 2 h
                 .to pH 3.7-4.0 with 1 N H2SO4 to liberate CO2 from die test
                  waters. Containers, with pleated pieces of filter paper soaked in
                  1 mL of .10% KOH, were suspended inside die jars to trap
                  evolved  COj. Filter papers plus die KOH were added to 16 mL
                  of fluor, and 200 p.L of glacial acetic acid was added to each
                  sample to reduce chemolumescence. Samples were counted after
                  storage in die dark overnight

                  Results

                  Time Course and Effects of H+ and Ca on die Accumulation
                      of Cd on  Minnow Gills
                    Cadmium accumulation on fathead minnow gills was linear to
                  about 2 h (r- 0.92, P < 0.001), after which accumulation
                  levelled  off (Fig.  1). Two- to 3-h exposures were therefore
                  appropriate for assessing initial accumulation of Cd on minnow
                  gills. The Cd accumulation pattern was die same whether or no!
                  0.1 jiM  EDTA  was present at twice me Cd concentration,
                  however no Cd accumulated on gills in 2 h if 0.25 jiM EDTA (at
                  5X the Cd concentration) was present (Fig. 1).
                    Increasing water Ca concentration from 35 to 95 \iM did not
                  significantly reduce Cd accumulation, but Ca concentrations of
                  1055 and 2000 u,M fully prevented Cd accumulation on gills
                  (Fig. 2). In Cd exposures run at  pH 4.83 ± 0.08 (7) with no
                  additional Ca, no Cd accumulated on gills (Fig- 2).

                  Effects of Natural Dissolved Organic Carbon on Metal
                      Binding to Gills
                    To determine if eitiier die source of dissolved organic carbon
                  (DOC),  or die size fraction of DOC, affected Cu accumulation
                  on gills, we tested diree size fractions of DOC from five lakes
                  and  Luther  Marsh.  As preliminary data using Luther
                  Marsh DOC  suggested dial 5 mg DOC-L'1  prevented  Cu

                                                                      2669

-------
                                                                                              3.6
                                                             4.0
                                                                               (0.05iiM)i
   . (A) did not affect Cd accumulation, whereas 0.25 »tM EDTAdid (•). O = fish noTexposed to Cd
    The grey horizontal band represents the 95% confidence interval of gills of all fish exposed to 6 u* Cd I • l«
                BackgreundandOUx^^

    0.8
                                                                                        DOC
                       pH6.3
pH4.8
3.3. Gill Cu concentrations of fathead minnows exposed for 2-3 h in
ft water (pH 6.2) to 17 fig Cu -L~' and various concentrations <|f
Fia. 2. Effects of Ca and H+ on Cd accumulation on gills of minnows
exposed to 6 ngCd •L~lfor2to3hinsoftwater(pH6.2).&I.055jiM
Ca prevented Cd accumulation on gills. No Cd accumulated on gills of
fish held at pH 4.8 (no added Ca). Vertical lines represent the 95%
confidence interval (from left to right, n = 5,6,6,5). For all figures *
**. ***=significantly below metal-exposed gills (grey horizontal band)
and +, ++,  +++  =  significantly  above  background  gill  metal
concentrations (dashed horizontal lines) for P < 0.05, < 0.01, < 0.001,
respectively.


accumulation on minnow gills (data not shown), we tried to test
that amount of each DOC fraction, although not all size fractions
for all lakes were examined.
  Accumulation of Cu on gills of fish exposed to 17 ngCu- IT' in
the presence of DOC indicated that DOC concentration was the
main factor in decreasing Cu binding to gills (Fig. 3). All gill Cu
concentrations  were   reduced   to background   at  DOC
£4.8 mg- L-', regardless of DOC size  source or fraction, and
were significantly reduced relative to gills from fish exposed to
Cu in the absence of DOC (F < 0.001). Below 4.8 mg- L~" little
or no protective effect of DOC was observed, except for two
values  at 3.7 mg-L-'. These  two results were  from  the

2670
no.
soft water (pH 6.2) to 17 fig (
dissolved organic carbon (DOC). Three size fractions of fteeze-dried
DOC from  various lakes were used: • = 1000-30000 Da, •!=
>30 000 Da but <0.2 jim, and A * >0.2 pin. Gill Cu was reduced
significantly at all DOC concentrations a4.8 mg - L~'. Hie data fit|a
straight.line: gill Cu-=  -0.26 • [DOC] + 2.52 (r = -0.77. P < 0.001if.
Vertical lines represent the 95% confidence interval (»=5 to 6 fish gills i.
Dashed horizontal lines represent the 95% O of background gill Cu
(n = 67). The grey horizontal band represents the 95% O of all
Cu-exposed fish gills (n = 74) in the absence of added DOC (DOC
<2mg • L~!). Background and Cu-exposed gill Cu concentrations were
significantly different (P < 0.001).
                 1000-30 000 Da DOC fraction from Salmon L. and William's
                 Bay; neither was significantly different from background gill Cui
                 and both were significantly lower than Cu-exposed gills in
                 absence of added DOC (P < 0.001). These two values sugg
                 that lower molecular weight DOC fractions complex Cu to i
                 greater  degree  than do  higher molecular  weight fractions'.
                 Otherwise it appears that, as a  first  approximation,  DOC
                 concentration, not source or size fraction, is the dominant factor
                 in determining metal accumulation on fish gills.             ||
                   Next, we looked, at the effects of DOC  on  Cu and C<1 *"
                                                                         I
                                              Can. J. Fish. Aguat. Sci.. Vol. 50,1993'

-------
                                                                        Smg-L-'DOC
                          0  '	^	'   *" 0
                                           Cu, Cd.Zn           ,         Cd     Cu.Cd,Zn
                       FIG. 4. (A) Gill Cu concentrations of fathead minnows held for 2-3 h in soft water (pH 6.2)
                       with either 17 |ig Cu -L~' or 17,6, and 19 jtg Cu, Cd, and Zn • L~',  in the presence
                       (stripes) or absence of DOC. Five milligrams DOC per litre reduced gill Cu to background;
                       (B) gill Cd concentrations of fathead minnows held for 2-3 h in soft water (pH 6.2) with
                       either 6 u,g Cd • L"1 or 17,6, and 19 u.gCu,Cd, andZn • L"1 ,in the presence or absence
                       of DOC. Five milligrams DOC per litre (stripes) did not significantly reduce Cd
                       accumulation on the gills.
accumulation on minnow gills, either present as individual
metals or as a mixture of Cu, Cd, and Zn. Once again,
5 mg DOC- L"1  (1000-30 000 Da Luther Marsh) prevented
Cu accumulation on fathead gills (Fig. 4A). The simultaneous
presence of 6 jtg Cd- L*1 and 19 jig Zn- L~'  did not affect Cu
deposition, in either the absence or presence of DOC. This
result indicates either strong binding of  Cu  to the gills
(relative to Cd and Zn), excess metal binding sites on the gills,
or different binding sites for each metal. Background gill Zn
concentrations were so high (20-40 jig Zn- g wet tissue"1) that
the 19 jig Zn- L"1 exposures did not significantly alter gill Zn
concentrations (data not presented).
  Cadmium  accumulation was  unaffected  by simultaneous
exposure to Cu and Zn (Fig. 4B), again indicating either strong
Cd binding, or excess or different metal binding sites  on gills.
With 5 mg DOC- L~( there was no significant reduction in the
amount of Cd on the gills (Fig. 4B), in marked contrast to the
results for Cu (Fig. 4A). Later experiments to determine the
concentration of DOC needed to prevent Cd accumulation on
gills, yielded 0.29 ± 0.15 (6) and 0.30 ± 0.08 (6) |ig Cd • g wet
dssue" (  ±95%  CI) for 5.1 and 7.7 mg DOC -L'1 of the
1000-30 000 Da Luther Marsh fraction, respectively. Both these
concentrations  were significantly  below values  from Cd-
exposed  fish in the  absence of  DOC  (P< 0.001, <0.01,
respectively), and were not significantly above background.

Complexation of Cu and Cd Using Synthetic Ligands
  We used synthetic Hgands to estimate conditional metal-gill
equilibrium constants (AT) for Cu and Cd. The approach is based
on equilibrium  constants for metal-ligand complexes in soft
water, such as those published in Morel (1983) and used in
computer  programs like  MINEQL* (Schecher 1991).  For
example, the Cu-EDTA complex (log K = 18.8) forms about
300X more strongly than does Cd-EDTA (log K - 16.3), and
about 0.5 X 10*X more strongly than does the Cu-NTA  complex
(log  K =  13.1; log K values from MINEQL*. summarized in

Can. J. Fish. Aquat. Set., W. SO, 1993
 Table 1). Copper binds to NTA about 5000X more strongly than
 does Cd (iog XCII-NTA = 9.4). Note that in all cases Cu binds to
 these Hgands better than does Cd. By exposing fish to metals and
 Hgands  at  known  concentrations   in  water  of  defined
 composition, and measuring metal accumulation on gills, metal-
 gill stability constants can be estimated.

 Copper                .          .
   To estimate Cu-gill stability constants, we exposed fathead
 minnows to 17 jig Cu- L~' (0.27 |iM) in  soft water in the
 presence of various ligands: 0.25 |iM EDTA prevented Cu from
 accumulating on gills (Fig. 5), indicating that enough Cu was .
 bound in the Cu-EDTA complex to render Cu unavailable to bind
 to the gills. MINEQL* calculations yielded 0.019 jtM free Cu,
 with the remaining Cu in the Cu-EDTAform. Likewise, 0.25 u,M
 NTA prevented Cu accumulation (Fig. 5). Only about 0.030 |*,M
. Cu was calculated to be in the free form in the presence  of
 0.25 \iM NTA(MINEQL+ calculation). 2.5 uMNTAalso reduced
 gill Cu to background (0.61 ± 0.16 (4)); only about 0.0002 (iM
 Cu was calculated to be free at this NTA concentration.
   Interpretation of results using  ligands that bind  Cu more
 weakly than EDTA and NTA is complicated by the  fact that Ca
 and H* compete with Cu for these ligands (e.g. log K values are
 similar,  Table 1). Thus, not all the added ligand is  available to
 bind Cu. For example, only 0.04 |iM of the 0.25 jiM  ethyl-
 enediamine was in the CuEN form (-16%), the rest of the ethyl-
 enediamine being in the H~ or rF~ form (MINEQL* calculation).
 In contrast, >96% of the 0.25 \iM EDTA "or NTA was available
 to bind  Cu (above). Ethylenediamine, citric, and  oxalic acid
 (25 p.M) were needed to  prevent Cu accumulation on gills
 (Fig. 5), for which 0.006,0.007, and 0.044 \tM, respectively, of
 the 0.27 jtM Cu in solution, was calculated to be free. In addition
 to these data, we had conducted a series of experiments with
 glutamic and salicylic acids, in which 25 \M glutamate reduced
 gill Cu accumulation slightly (but not significantly; to  1.84 ± 0.54
 (6)), and 250 \iM salicylic acid was needed to reduce gill Cu to
 1.31 ± 0.28 (5), not significantly above background.

                                                     2671

-------
                      Fio.  5.  Amount of  Cu  on the gills of  fathead minnows  exposed  to  Cu
                      (0.27 jiM = 17 fig • L~l) and ligands for 2-3 h in soft water at pH 6.2.0.25 >M EDTA
                      and NTA prevented Cu accumulation on the gills, whereas higher concentrations of the
                      other ligands were needed to do so. From left to right, n = 10,4,9,9,9,6,5,5,6,5,5.
                      The horizontal bands designated by dashed lines or grey shading represent, respectively, the
                      95% CI for background gill Cu, and Cu-exposed gills in the absence of ligands.'
Cadmium                          .
  The results for Cd binding to gills in the presence of synthetic
ligands were surprising. Although Cd binds less well to synthetic'
ligands than does Cu (Table I), Cd bound to minnow gills as well
or better than did Cu. Between 2 and 5 X more EDTA than Cd
was needed to prevent Cd accumulation on gills (Fig. 6; see also
the Cd time course experiment, Fig. 1). This much EDTA would
leave between 0.001 and 0.005 uM of the 0.05 p-M total Cd in
solution as free metal (MINEQL+ calculation). Extremely low
concentrations   of  free  Cd  resulted in  significant   Cd
accumulation on gills, in contrast to the  0.006 to 0.04 ftM free
Cu that was not enough free  Cu to result in increased gill Cu
concentrations (above).
  Both Ca and H* complex with NTA and citric acid, reducing
the amount of ligand available  to bind  Cd. NTA (5 ftM) and
2500 \iM citric acid significantly reduced Cd accumulation on
minnow gills (Fig. 7). Free  Cd under these conditions  was
calculated to be 0.03 and 0.0002 |iM, respectively. The citric
acid results agree with the Cd-EDTA results: virtually all Cd in
solution must be complexed, otherwise measurable Cd accu-
mulation occurred. The Cd-NTAresults agree with the Cu-oxalic
acid results, in that 0.03-0.04 u.M of free metal was not enough
to result in metal accumulation on the gills. Considering all the
Cd and Cu data on a molar basis, there needs to be the same or
less free Cd than Cu if no metal is to accumulate on the gills.
That is, the affinity of Cd for the gills is the same or greater than
that of Cu.

14C-Ligand Experiments
  To this point the experimental design and our interpretation of
the results were based on the assumption that free metals interact
with fish gills, while metal-ligand complexes do not. Copper-
EDTA and Cu-NTA complexes  apparently do not bind to fish
gills (Fig. 5), but  Cu accumulation in the presence of low
concentrations of ethylenediamine, citric acid,  or oxalic  acid
could be interpreted as Cu-Hgand accumulation on or in gills.
Similarly, elevated Cd accumulation on gills for the lower ligand
concentrations (Fig. 6, 7) could have been due to Cd-ligand
accumulation on or in gills. To ensure  the correctness of our
interpretations, .we wanted to verify that the free metal, not the
metal-ligand complex, was responsible for elevated gill melal
concentrations in our experiments. Numbers offish used in there
experiments were small, due to the expense of the radiolabellod
ligands.                                              II
  Three fathead minnows exposed to "C-EDTA (0.1 fim'sl
EDTA = 2.4 X10« DPM) in 1  L of soft water (no  Cd) hjjp^
background gill Cd values of 0.11-0.13 u,g Cd • g~',  arlclH
36-48 DPM  per gill sample  (Fig. 8),  the  same  level of
radioactivity  as fish  exposed to  Cd  without radiolabl;!
(29-43 DPM per gill sample).  Thus, "C-EDTA  did  nut
accumulate on minnow gills. In Cd plus "C-EDTA exposure!';,
gills accumulated Cd but  there was no increase in gill DPIJ!
(Fig. 8). MINEQL+-calculated Cd-EDTA was 90% of total CJL
Assuming 1:1 .Cd:EDTA binding, the expected counts if 90% of
the Cd accumulation was due  to Cd-EDTA would be SOOJj-
1 800 DPM (gill samples = 0.07 g wet weight each; allowing a
generous (50%) margin for counting inefficiency and loss of
sample). To ensure that these seemingly large counts would be
detected, we added  10-200  ILL of 0.1  ftM  "C-citrate
(= 2.45 X 10* DPM • L-') to gill samples from control fish. Gill
DPM  versus radiolabel  added was a  linear  relationshijp
(r a 0.999, n  = 5). A  10-jtL spike (1 X 10'6 ujnol citrate)
yielded 22 DPM above background, and a 20-p.L spike yieldeli
44 DPM above background. An additional 20-40 DPM (1.5-2X
background counts) would therefore easily be detected in our
system.                     , •                        |
   "C-labelled -  .anthracene,   a   lipophilic   polyaromatio
hydrocarbon,  was also used to test  detection of 14C. About
2 X 106 DPM of 14C-anthracene (0.07 junol) was added to 1 it
of soft water, and three fish were exposed to this solution far
2-3 h. These fish had gill counts of 370-550 DPM (Fig. 9)J
whereas two control fish had gill counts of 39 and 46 DPM. Tw
fish exposed to the dimethyl sulphoxide (DMSO) carrier
0. 1 nmol MC-citrate (= 4.2 X 106 DPM • L~') had gill counts
49 and 58 DPM (Fig. 9); anthracene's entry was therefore due t
-------
I
       0.8
       0.6
       0.4
       0.2
                                                                 0.6 r
  *
                    nui*'iM
                                        Cd  = 0.05

FlG. 6. Amount of Cd on gills of fathead minnows exposed to Cd
(6 M-g • L~' = 0.05 nM) and EDTA for 2-3 h in soft water at pH 6.2. It
took between 2 and 5X more EDTA than Cd to prevent Cd accumulation
on the gills. From left to right, n - 13,4,5,4. The horizontal bands
designated by dashed lines or grey shading represent, respectively, the
95% CI for background gill Cd, and Cd-exposed gills in the absence of
ligands.
"C-citrate had background gill Cd concentrations and a mean
gill radioactivity value of 58 DPM. Nine fish exposed to Cd alone
had gill Cd concentrations of 0.31 u,g Cd • g~' (wet tissue) and
mean gill DPM values of .44.  Three fish  held in  O.I  t*M
I4C-citrate plus 0.05 (iM Cd showed Cd accumulation on gills,
but no 14C-citrate  entry.  However,  MINEQL* calculations
showed 98% free Cd under these conditions, so binding of free
Cd, as opposed to Cd-citrate, would be expected. Similarly, there
was no evidence of l4C-citrate on gills offish held in 5 jiM citrate
with 0.1  jiM >4C-citrate: again, plenty of free Cd would be
available to bind  at the gills  (80% of total Cd) under these
conditions.
  With Cd plus 25 pM cold and labelled citrate (16.35 X 10*
DPM in  1 L) no counts due to  I4C were detected in the gill
samples (Fig. 10). Here, 57% of the Cd would be bound by citrate
(MINEQL* calculations). Under these conditions, again with a
generous provision for counting inefficiency, an additional 18,
21, and 31 DPM would have been expected in the samples if 57%
of the Cd accumulation was  due to Cd-citrate entry. These
increases would have been detected in our system, but the gill
samples were within the background range. Finally,  250  jiM
citrate (19.44 X 10* DPM in 1 L) did not prevent Cd  accumu-
lation on the gills, in accord with results presented in Fig. 7. Here,
95% of the Cd was calculated to be in the Cd-citrate form, and
9,6, and  6 DPM above background would have been expected
if 95% of the Cd accumulation  on the gills had been due to
Cd^citrate. These values are at the detection limit of thus method,
but the gill DPM values were within background. In total, there
was no evidence that '4C-citrate accumulates on gills.
  There was, however, the possibility that 14C-citrate was taken
up at the gills and metabolized to 14CO2 in the tricarboxylic acid
cycle. This scenario was tested by trapping UCO2 in 10% KOH

Can. J. Fish. Aquat. Sci, Vol. 50.1993
 FIG. 7. Amount of Cd on gills of fathead minnows exposed to Cd
 (6 |tg • L'1 = 0.05 nM) and NTA or citric acid (2-3 h in soft water,
 pH 6.2). NTA (5 |iM) and 2500 pJvt citric acid were necessary to prevent
 Cd accumulation on gills. These experiments were done -6 mo after
 those presented in previous Figures,  which shifted Cd values:for
 background (dotted horizontal lines, n = 12) and Cd-exposed gills in the
 absence of added ligands (grey horizontal band, n =  12). From left to
 right, n = 5,5,5,9,5,5,   •
 in sealed jars containing 250 mL of a 0.1 u,M "C-citrate solution
 (6.7 X 106 DPM • L-'). The KOH sample from three jars con-
 taining one fish each yielded 17 270 ± 1 140 DPM (mean
 ± 1 SEM). With 6 ng Cd  • L~« in solution, KOH from jars con-
 taining fish yielded 13310 ± 1180 DPM. KOH from two sealed
jars (no fish) with magnetic stir bars set to mimic agitation of
 water due to minnows swimming gave 18 210 and 23 820 DPM.
 KOH  from a single, unstirred sealed jar (no fish),  yielded
 3 240 DPM, and KOH alone gave 80 DPM. "C caught in the
 KOH traps was unrelated to the* presence  of fish, and appeared
 to be  due to breakdown or volatilization of the  14C-citrate,
 encouraged by agitation of the water either by fish or by stir bars;
 It is therefore unlikely that metabolic breakdown of MC-citrate
 bound to gills was responsible for low counts of gills from the
 HC-citrate plus Cd experiments.


 Discussion               -
  The most surprising result of our work was the strong binding
of Cd to fathead minnow gills relative to that of Cu. Synthetic
ligand experiments (Fig. 5,6,7) indicated that the concentration
of free Cd required for significant accumulation of Cd on gills—
using graphite .furnace technology — was the same or less than
the concentration of free Cu required for significant Cu binding.
'If adsorption on the gill surface was the sole factor determining
binding, and was similar to binding to other ligands, then Cd
would be expected to bind about 10-100X less well to gills than
would Cu (Table 1).  The probable reason for the difference is
uptake of Cd through high affinity Ca channels, in addition to
surface binding, whereas Cu uptake is likely through general
surface binding plus possible uptake through low affinity ion
channels. Both Cd and Ca cations have an ionic radius of -1.2 A,
and Verbost et al.  (1989) showed  that  any treatment that
decreased gill  Ca accumulation had a similar effect on Cd
'accumulation, strongly suggesting that both cations enter gills
by the same pathway.
  Reid and McDonald (1991) also found that Cd bound to fish
gills better than did Cu. Radiotracers bound to excised rainbow
. trout gills in the order LaJ+> Ca2* = Cd2* > Cu2+. Lanthanum was
expected to bind best, due primarily to its triple charge; they
explained the weaker than expected binding of Cu to gills
through the coordination chemistry of Cu2* (Reid and McDonald

                                                     2673

-------
                                                                                          300
                                           -                .        -.|ig-
                                         A plus Cd. No additional counts due to 14C-EDTA were
or u.i jun "U-BUJ A piusca. NO additional counts due to "C-EDTA were detected. Three
fish per treatment, with tKe order of gill Cd and DPM preserved left to rightin each treatment
   S
        600
       400
       200
                control    14C-citrate   14C-anthracene

                            + DMSO  •    -+  DMSO
 FIG. 9. Disintegrations per minute (DPM) of fathead minnow gills from
 two fish exposed to soft water only (control), two fish exposed to
 0.1 fiM l4C-citrate plus dimethyl sulphoxide (DMSO), and three fish
 exposed to 0.07 |iM 14C-andiracene plus DMSO. Accumulation of
 14C-labelIed anthracene in or on gills was easily'measured by our
 technique.                              •               '
  1991). From our viewpoint, the explanation needed is why Ca
  and Cd bound better to the gills than did Cu. As presented above,
  active uptake of Ca and Cd at high affinity sites can explain their
  results as well as ours. Radiotracers such as lo9Cd can be used at
  very low concentrations of total metal (e.g., 0.003 u,M Cd, about
  a 10X lower concentration than we used), but at such low Cd
  concentrations exposure times necessary for adequate gill counts
  need to be about one day (Wicklund Glynn et al. 1992). There-
  fore, graphite furnace AAS is a good method to measure metal
  uptake at gills, because it allows both low metal exposure
  concentrations and short exposure duration.
   Significant morphological changes in fish  gills occur at
  >10u,gCd  -L-',  even in synthetic soft  water  (Karlsson-
  Norrgren et al. 1985)! Low concentrations of Cd do not lead to
  general increases in gill permeability to ions (increased leakiness
  due to displacement of Ca from cell membranes; Verbost et al.
"1989). Therefore, the ameliorative effects of Ca against  low
  concentrations  of Cd are  probably  due to specific Ca-Cd
  antagonism, not to a general  reduction  in  gill permeability

  2674                                           '
                                      (Pratap et al.  1989; Verbost et al.  1989; Fig. 2). Reduced Oi
                                      influx (ReidandMcDonald 1988) can be explained by inhibition
                                      of basolateral  CaJ+-ATPase. This inhibition results in increase)!
                                      intracellular Ca, with a resultant reduction in Ca entry at cell
                                      surfaces(Verbostetal.l989),andrnayeventuallyleadtocelldeatli!
                                        Our dissolved organic carbon (DOC) results indicate, as a first
                                      approximation, that source and size fraction of DOC do not
                                      determine the  protective effect of DOC against Cu binding t<>
                                      gills. If DOC concentration  was  £4.8  mg-L"1, Cu waij
                                      prevented from binding to minnow gills (Rg. 3). This result
                                      agrees with those of Cabaniss and Shuman (I988b), who founc'i
                                      that source may be less important in explaining variations in Cu
                                    .  binding by dissolved organic matter than are chemical factor!
                                      such as pH, alkalinity, and ionic strength. Cadmium binds tc
                                      DOC about 10X less well than does Cu (Morel 1983; Albert;
                                      and Giesy 1983; Fig.  4). Estimates of Cu-DOC binding  varj
                                      considerably, with log K values ranging from 6 to 11 for watei
                                      with pH > 6 (Morel 1983). Our own estimates of Cu and Cc
                                      binding constants to DOC are made in our companion papa
                                      (Playleetal.1993).
                                       Our experiments were run for 2-3 h, long enough for Cd anc
                                      Cu accumulation on gills to stabilize (Fig. 1; Playle et al. 1992)
                                     Measured metal accumulation is probably a  combination ol
                                      surface-adsorbed metal, and subsequent metal transfer across
                                     gill surfaces. This process of adsorption followed  by entry is
                                     assumed to occur for metal uptake by algae (Morel et al. 1991]
                                     and is likely the process at fish gills. Verbost et al. (1989) usec
                                     an EDTA rinse to remove surface-bound I09Cd and 43Ca, anc
                                     found that both tracers were associated with the inside of the gil
                                     epithelium. No  differentiation between  surface-bound  anc
                                     interior metal was made in our study. Our premise was that anj
                                     process  affecting initial  adsorption at the gills  —  mainly
                                     competition or metal complexation—would reduce initial meta
                                     adsorption and therefore accumulation and toxicity. Results ol
                                     SkowroAski et al. (1992) indicated that similar competition anc
                                     complexation effects (by K* and Cl~, respectively) are seen id
                                     either 5- or 90-min Cd exposures, even though approximately!
                                     20% more Cd accumulated in their cyanobacterium during the
                                     longer exposure.                                         I
                                       Metal binding proteins  such as metallothionein may afford
                                     some  protection against  Cu over time,  but production o
                                     metallothionein was likely not a factor in our short experiments

                                                                  Can. J. Fish. Aquat. Sci, Vol. 50.I993t

-------
             _   0.6
             •a

             |   0.4
                 0.2
             o
             1
                              gill Cd
     - gill DPM
                                                                                              300



                                                                                              200

                                                                                              gill 0PM

                                                                                              100
                       0.1 iM  0.05 (Ml
                     "C-dtrate    Cd
                            260
                                                     5             25
                                             ftM cold plus 1*C-cltrate; O.OS §M Cd
Flo. 10. Gill Cd (grey bars) and disintegrations per minute (DPM, clear bars) of gill samples from fish exposed for 2-3 h in soft water
to combinations of f4C-citrate, "cold" citrate, and Cd (6 u,g • L~' = 0.05 jtM). Wider bars at left represent mean values ±95% CI
(R = 9); otherwise each bar represents one gill sample, with order preserved left to right in each treatment There were no additional
counts due to 14C-citrate.                                                          •
 In any case, the gill is the first organ affected by waterbome Cu;
 induction of  metallothionein-Iike proteins  will  only have
 protective effects once gills have regained normal regulatory
 function (Laure'n and McDonald 1985). Longer-term studies
 would indicate if acclimation to Cu and Cd occurs through
 changes in metal-gill binding characteristics (Reid et al. 1991),
 changes that could lead to reduced metal concentrations on the
 gills (McDonald et al. 1991).
   Interpretations of our results are based on the premise that
 some  ligands can outcompete  gills for metals, whereas other
 ligands that form weak metal complexes will not. Implicit in our
 analyses and calculations is die assumption that metal-ligand
 complexes do not bind to  gills. Otherwise, measured gill metal
 concentrations may actually represent meial-Hgand  accumu-
 lation on gills. There is reasonable agreement that EDTA and
 NTA do not pass through cell membranes easily (Jackson and
 Morgan 1978; Harrison etal. 1984; Patt and Wikmark 1984; Nor
 and Cheng 1986; Daly et al: 1990; Block and PSit 1992), due
. both to their negative charges and relatively large size. Our
 present data also gave no indication of "C-EDTA entry into fish
 gills (Fig.  8), although entry of 14C-anthracene  was easily
 measured (Fig. 9).
   The argument becomes  less clear when ligands such as citric
 acid are used. Citric acid is a small molecule and is a substrate
 in the tricarboxylic  acid cycle. Studies with a  number of
 organisms have shown that Cu and Cd accumulation or toxicity
 is not reduced as much as expected when small, weak ligands
 such as citric acid or glycine are used to complex the metals.
 Competition between ligand and organism for the metal ion, as
 weU as entry (and toxicity) of the small metal-ligand complexes
 have been discussed in most of these studies, with conclusions
 favouring  ligand  competition  (Zitko  and  Carson  1976;
 Knezovichetal. 1981; Harrison etal. 1984), metal-ligand entry
 (Guy and Ross Kean 1980; Borgmann and Ralph 1983; Part and
 Wikmark 1984), or a combination of both  (Giesy et  al. 1977;
 Poldoski 1979; Laegreid et al. 1983; Nor and Cheng 1986; Daly
 et al. 1990; Femandez-Pinas et al. 1991). Block and PM (1986)
 and Block et al. (1991) even showed that l09Cd transfer through
 perfused gills was enhanced when cpmplexed with xanthate, due
| to the lipid-soluble  nature  of the complex formed. Some
 confusion has undoubtedly occurred because ion  selective
 electrodes, used frequently to measure free metal ions, do not
 take into account metal that is only weakly bound to a particular

 Can. J. Fish. Aquat. Set. Vol. 50,1993
ligand and would therefore act as a supply of free metal to
another ligand (e.g., gill)  with  a higher binding constant
(Florence 1977; Giesy et al. 1977).
  In the absence of any knowledge of the relative strengths of
metal-gill and metal-ligand complexes, competition between
the gill  and ligands  for  free metal,  or uptake of  small
metal-ligand complexes, are both valid arguments for metal
accumulation or  toxicity in the presence of weak ligands.
However, once it is determined how little free metal is needed
for significant metal accumulation on gills (this study), and the
strengths of Cd-gill and Cu-gill binding are calculated (Playle et
al. 1993), the simpler interpretation of metal accumulation in the
presence of weak ligands becomes competition between gills
and ligands for free metal ions. For example, the Cd and citrate
work of PM and Wikmark (1984) suggested that the Cd-citrate
complex entered  perfused  rainbow trout  gills, but by their
calculations 0.01-0.04 pM  Cd (of a total concentration of
9.5 u,M) was still free in solution—likely enough free metal to
explain their results  through entry of free Cd alone. In our
experiments, just 0.0001-0.03 uM free Cd was needed to result
in significant accumulation of Cd on gills. As another example,
this time for freshwater shrimp, the reported toxicity of  the
Cu-glycine complex and lack of toxicity of the Cu-(glycine)2
complex  (Daly et al. 1990) can be explained by the stability
constants of the two complexes. When the concentration of
glycine is high enough, the Cu-(glycine)j complex (log K= 16.0;
MINEQL+) probably out-competes the freshwater shrimp for Cu,
whereas at lower concentrations the weaker Cu-glycine complex
may not Qog JTs 8.6).
  Our present work gave no indication of l4C-citrate entry along
with Cd (Fig. 10), suggesting that only free metal interacts at the
gills. 14C-citrate itself did not enter the gills, even in the presence
of DMSO (Fig. 9),  a strong  surfactant,  and there was  no
indication that l4C-citrate was assimilated and then lost as I4CO2.
Although metal-ligand entry could be occurring at a slow rate
by diffusion, this route of entry does not need to be invoked to
explain metal accumulation on gills. Depending on the relative
strengths of metal-ligand and metal-gill complexes, and ligand
interactions with cations such as H* and Ca2+, different concen-
trations of ligand will be needed to prevent metals from binding
to fish gills (Fig. 5,6,7). Metal-gill and metal-ligand stability
constants, along with ligand concentrations, should therefore be
useful for predicting metal accumulation on gills, and hence

                                                    2675

-------
  bioavailability and toxicity.    .
    In summary, measurement of actual metal accumulation on
  minnow gills has allowed us to determine that Cd binds to gills
  as well or better than does Cu, probably because of Cd uptake in
  high affinity Ca  channels. This result emphasizes that physio-
  logical processes at the.gill must be considered when predicting
  metal  accumulation on fish gills. Dissolved organic carbon
  protects against Cu binding to gills better than it does against Cd
  binding, which is related to generally greater affinity of Cu to
 .synthetic ligands, and the strong binding of Cd to fish gills. Hus^
  information, and  similar information for other metals, should be
  useful in predictive models for metal accumulation on gills, and
 .therefore should be useful to predict metal toxicity to fishJDucu^
  lation of metal-gill stability constants from our synthetic ligand
  results, and their use in a model to calculate metal binding to gills
  offish exposed to Cu and Cd in natural lake water, is the subject
  of our companion paper (Playleetal. 1993).


  Acknowledgements

    We thank Drs. Chris Wood and Gord McDonald, McMaster Uni-
  versity, Hamilton, Ontario, for the use of their graphite furnace for metal
 analyses, and Dr. Bruce Greenberg, University of Waterloo, for supply-
 ing die I4C-anthracene. This research was funded by an Operating Grant
 (No.  8155} from  the Natural Sciences and  Engineering Research
 Council of Canada to D.G. Dixon, by a Research Grant front the Dorset
 Research Centre, Ontario Ministry of the Environment, to D.G. Dixon,
 and by an KB. Eastburn Postdoctoral Fellowship to R. Playle.


 References

 ALBERTS, JJ., AND J.P. QESY. 1983. Conditional stability constants of trace
     metals and naturally occurring humic materials: Application in equilibrium
     models and verification with field data, p. 333-348. In R.F. Christman and
     ET. Gjessing (ed.) Aquatic and terrestrial humic materials. Ann Arbor
     Science. Ann Arbor. MI.            ••         •
 BENTLEY, PJ. 1991. Accumulation of cadmium by channel catfish (Ictalunu
     punctatus): influx from environmental solutions. Comp. Biochern. Physiol.
     99C: 527-529.
 BLOCK, M.,  AND P. PART. 1986. Increased availability of cadmium to perfused
     rainbow trout (Salmo gairdneri. Rich.) gills in the presence of the
     complexing agents diethyl ditniocarbamate, ethyl xanthate and isopropyl
     xanthale. Aquat  Toxicol. 8:295-302.
 BLOCK. M, AND P. PART. 1992. Uptake of lwCd by cultured gill epithelial cells
     from rainbow trout (Oncorhynckus myUa). Aquat Toxicol. 23:137-151.
 BLOCK, M., A.  WICKLUND GLYNN, AND P. PART. 1991. Xanthate effects on
     cadmium  uptake and  intracellular distribution in  rainbow trout
     (Oncorhynchta mytiss) gills. AquaL Toxicol 20:267-283.
 BOROMANN, U., AND KM. RALPH. 1983. Complexation and toxicity of copper
'     and the free metal bioassay technique. Water Res. 17:1697-1703.
 CABANISS, S.E, AND M.S.SHUMAN. 1988a. Copper binding by dissolved organic
     matter: L Suwannee River fulvic acid equilibria. Geochim. Cosmochim.
     Acta 52:185-193.                 ,
 CABANISS. S.E.. AND M.S. S HUMAN. 1988b. Copper binding by dissolved organic
     matter: II. Variation in type and source of organic matter.  Geochim.
     Cosmochim. Acta 52:195-200..
 CAMPBELL, P.G.C., AND P.M. STOKES. 1985. Acidification and toxicity of metals
     to aquatic biota. Can. J. Fish, Aquat Sci. 42:2034-2049.
 CUSMANO, R.F., D.F. BRAKKE, AND G.A. CHAPMAN. 1986. Effects of pH on the
     toxicities of cadmium, copper, and zinc to steelhead trout (Salmo gairdneri).
     Can. J. Fish. Aquat. Sci. 43:1497-1503.
', DALY, H.R., LC. CAMPBELL, AND B.T. HART. 1990. Copper toxicity to Paratya
     aiutraliensis: I.  Influence of nitrilotriacetic acid and glycine. Environ.
     Toxicol. Chem. 9:997-1006.
 FERNADEZ-PINAS, F., P. MATEO, AND I. BONILLA. 1991. Binding of cadmium by
     cyanobacterial growth media: free ion concentration as a toxicity index to
     the cyanobacterium Noaoc UAM 208. Arch. Environ. Contam. Toxicol.
     21:425-431.     ,                 .                    ,
  FUEMMING, C.A., AND J.T. TREVORS. 1989. Copper toxicity and chemistry in the
     environment: a review. Water Air Soil Pollut. 44: 143-158.

  2676                      -                           '
 FLORENCE, ^M. 1977. Trace metal species in fresh waters. Water Res
     681-687.
 GIESY, J.P. JR., G J. LEVERSEE, AND D.R. WILLIAMS. 1977. Effects of natuniuV
     occurring aquatic organic fractions on cadmium toxicity to Simocephaliu
     terrulatiu (Daphnidae) and Gambusia afftnis (Pccciliidae). Water Res. Ill:
     1013-1020.
 Guv, R.D.. AND A. Ross KEAN. 1980. Algae as a chemical speciation monitot'-I.
   .  A comparison of algal growth and computer calculated speciation. Water
     Res. 14:891-899.                                           I
 HARRISON, S J., N.W. LEPP, AND DA Pram. 1984. Uptake of copper by excised
     roots IV. Copper uptake from complexed sources. Z. Pflanzenphysiol. 1 ii3:
     445-450.                         '         •     •
 HUTCHINSON, NJ., AND J£. SPRAOUE. 1986. Toxicity of trace metal mixtures to
     American flagfish (Jordanella floridae) in soft, acidic water and impli-
     cations for cultural acidification. Can. J. Fish. Aquat. Sci. 43:647-655.' |
 HUTCHINSON, NJ., AND J.B. SPRAGUE. 1987. Reduced lethality of Al, Zn and (pu
  y mixtures to American flagfish by complexation with humic substancesjin
     acidified soft waters. Environ. Toxicoi. Chem. 6:755-765.           I
 JACKSON. G.A., AND JJ. MORGAN. 1978. Trace metal-chelator interactions aitd
   •'nhytoplankton growth in seawater media: theoretical analysis and c«n-
   "parison with reported observations. LimnoL Oceanogr. 23:268-282.   ||
 JOHNSON, M.G. 1991. Temporal trends in metal concentrations in rivers, coasua-
     zone waters, and sediments of northern and eastern Georgian Bay. J, Great
     Lakes Res. 17:241-254.       :                               II
 KARLSSON-NORRGREN, L., P. RUNN, C. HAUX, AND L. FOUJN. 1985. Gdmhnji-
     induced changes in gill morphology of zebrafish, Brachydanio rerip
     (Hamilton-Buchanan), and rainbow trout, Salmo gairdneri Richardson, j.
     Fish Biol.  27:81-95.                                       •  I
 KNEZOVJCH, If.. F.L. HARRISON, AND J.S. TUCKER. 1981. The influence if
     organic dictators  on the toxicity of copper to embryos of the Pacific oyster,
     Craaostrea gigas. Arch. Environ. Contam. Toxicol. 10:241-249.     ||
 LABORED. M.. J. ALSTAD, D. KLAVENESS, AND H.M. SEIP. 1983. Seasonal vari-
     ation of cadmium toxicity toward the alga Salenastrum capricomutwn
     Printz in two lakes of different Humus content Environ. Sci, Technol. llj:J
     357-361.             '    -                                 -   |
 LAUREN. D J., AND D.G. MCDONALD. 1985. Effects of copper on branchial ioncj-
     regulation in the rainbow trout, Salmo gairdneri Richardson. J. Com]).
     Physiol. B 155:635-644.                                      ||
 LAUREN, DJ.. AND D.G. MCDONALD. 1987. Acclimation to copper by rainbow
     trout Salmo gairdneri: biochemistry. Can. J. Fish. Aquat Sci. 44:105-11 Jj.
 MCDONALD, D.G.. CM. WOOD. R.G. RHEM, M.E. MUELLER, D.R. MOUNT, AND
     H.L. BERGMAN. 1991. Nature and time course of acclimation to aluminum
     in juvenile brook trout (Salvelinusfontinalis). I. Physiology. Can. J. Fislil
     Aquat Sci. 48:2006-2015.                                     II
 MOREL, F.M.M. 1983. Principles of aquatic chemistry. John Wiley and Sons'
     Toronto. ON. 446 p.                               'II
 MOREL, F.M.M., RJM. HUDSON. AND N.M. PRICE. 1991. Limitation of produc
     tivity by trace metals in the sea. Limnol. Oceanogr. 36:1742-1755.     |
 NOR, Y.M., AND H.H. CHENG. 1986. Chemical speciation and bioavailability of
     copper uptake and accumulation by Eichornia. Environ. Toxicol. Chem
   '5:941-947.            -                                     ]
TAQENKOPF, O.K. 1983. Gill surface interaction model for trace-metal toxicitji
'    to •fishes: role of complexation, pH, and  water hardness. Environ. Sci!
     Technol. 17:342-347.                                         ||
 PART, P., AND G. WIKMARK. 1984. The influence of some complexing agentis
     (EOTA and citrate) on the uptake of cadmium' in perfused rainbow troui!
     gills. Aquat Toxicol. 5:277-289.
 PLAYLE, R.C., R.W. GENSEMER, AND D.G. DDCON. 1992. Copper accumulatioi
     on gills of fathead minnows: influence of water hardness, complexation am
,    pH of the gill micro-environment. Environ. Toxicol. Chem. 11:381-391.
 PLAYLE, R.C., D.G. DDCON, AND' K. BURNISON. 1993. Copper and cadmiun
     binding to fish gills: estimates of metal-gill  stability constants  am
     modelling of metal accumulation. Can. J. Fish. Aquat Sci. 50:2678-2687
 POLDOSKI, J.E.  1979. Cadmium bioaccumulation assays. Their relationship u
     various ionic equilibria in Lake Superior water. Environ. Sci. Technol. 13
     701-706.
 PRATAP, H.B., H. Fu, R.A.C. LOCK. AND S.E. WENDELAAR BONOA. 1989.1
     of waterborne and dietary cadmium on plasma ions of the i           •—•*•
     chromis mossambiciis in relation to water  calcium levels. Arch. Environ j3S£*
     Contam. Toxicol. 18:568-575.
 RETD, S.D., AND D.G. MCDONALD. 1988. Effects of cadmium, copper, and low
     pH on ion fluxes in the rainbow trout, Sabno gairdneri. Can. J. Fish. Aquat
     Sci.  45:244-253.
 REID, S.D.,  AND D.G.  MCDONALD.  1991. Metal binding activity of the gills o
     rainbow trout (Oncorhynchus mykiss). Can. J. Fish. Aquat. Sci. 48
     1061-1068.

                                  Can. J. Fish. Aquat. Sci. Vol. 50.199.

-------
SHD, S.D., D.G. MCDONALD, AND RJt RHEM. 1991. Acclimation to subtotal
   ' aluminum: modifications of metal-gill surface interactions of juvenile
    rainbow trout (Oncorhynchus myttii).  Can. J. Fish. Aquat Sci. 48:
    1996-2005.
SCHECHER, W. 1991. MINEQL* version Zl. Environmental Research Software.
Sonoss, K-, AND M.G. TAYLOR.. 1989. Metal fluxes across the membranes of
    aquatic organisms. Rev. Aquat. Sci. 1:173-188.
SKOWRONSKL T.. S. SZUBINSKA, M. JAKUBOWSXL AND B. PAWUK. 1992. Cad-
    mium availability to the cyanobacterinm Synechocyais aquatiiis in
    solutions containing chloride. Environ. PoUut 76:163-167.
SPRAOUE, JJJ.1987, Effects of cadmium on freshwater fish, p. 136-169. In J.O. *
    Nriagu and J£. Sprague fed.) Cadmium in the environment John Wiley
    and Sons, Toronto, ON.
SPRY, D J., AND J.G. WIENER. 1991. Mettl bioavailability and toxiciry to fish in.
    low-alkalinity lakes: A critical review. Environ. Pollut. 71:243-304.
VEBBOST, PM.. J. VAN Rooti, G. FUK, RA.C. LOCK, AND S£. WENDELAAR
    BONGA. 1989. The movement of cadmium through freshwater trout
    branchial epithelium and its interference with calcium transport J. Exp
    Biol. 145:185-197.                            .      «~    *~«"
WIOOUND, A., ANDP.RUNN. 1988. Calciumeffectsoncacunimn uptake, redistri-
    bution,  and elimination in minnows. Phoarau phoximu.  ««-limnM to
    different calcium concentrations. J^quat-Toxicol. 13:109-122.
WKKUWD GLYNN, A, C HAUX, AND C HOOSTRAND. 1992. Chronic toxiciry and
    metabolism of Cd and ZntajuvemUemiinKwsCP/wcrinwptoJinuj) exposed
    to a Cd and Zn mixture. Can. J. Fish. Aquat ScL 49:2070-2079.
Znxo, V., AND W.G. CARSON. 1976. A mechanism of the effects of water hard-
    ness on the lethality of heavy metals to fish. Chemosphere 5:299-303.
Can. J. Fish. Aquat. ScL. Hjt 5ft 1993
                                                                                                                                  2677

-------

-------
         Copper  and Cadmium Binding to  Fish Gills: Estimates  of
           Metal-Gill Stability Constants and Modelling of Metal
                                            Accumulation
                                  Richard C. Playle1 and D. George Dixon

                          Department of Biology, University of Waterloo, WateHoo, ON N2L3C1, Canada

                                            > and Kent Burnison
                                %                              ~                          .
                      National Water Research Institute, Environment Canada. Burlington, ON L7R 4A6, Canada

          Playle, R.C., D.C. Dixon, and K. Bumison. 1993. Copper and cadmium binding tofish gills: estimates of metal-gill
                 stability constants and modelling of metal accumulation. Can. j. Fish. Aquat Sci. 50:2678-2687.
          Fathead minnows (Pimephales promelas) were exposed to 17 jig Cu • L"1 or 6 jig Cd • L~' in synthetic soft water
          in the presence of competing ligands. Measured gill  metal concentrations correlated with  free metal ion
          concentrations, not with total metal. Langmuir isotherms were used to calculate conditional metal-gill equilibrium
          constants and the number of binding sites for each metal. Log Kc»4i\ was estimated to be 7.4 and the number of
         Cu binding sites on a set of gilIs (70 mg, wet weight) was -2 x 10~* mol (-30 nmol • g wet weight'1). Log Ku-*\\
         was -8.6, and the number of Cd binding sites on minnow gills was -2 X 10~10 mol (-2 nmol • g wet weight"').
         Stability constants for H* and Ca interactions at metal-gill binding sites and for metal interactions with dissolved
         organic carbon (DOC) were estimated using these metal-gill constants. All stability constants were entered into
         the M1NEQL* aquatic chemistry program, to predict metal accumulation on fish gills using metal, DOC, and Ca
         concentrations, and water pH. Calculated metal accumulation on gills correlated well with measured gill metal
         concentrations and with LCso values. Our approach of inserting biological data into an aquatic chemistry program
         is useful for modelling and predicting metal accumulation on gills and therefore toxicity to fish.

         Des Tites-de-boule (Pimephalespromelas) ont && exposes a du Cu ou du Cd en concentration de 17 pg • L"1 ou
         6 fig' L~', respectivement, dans une eau douce synth&ique,  en presence de ligands concurrents. Les
         concentrations mesurees de mettl sur ies branchiesitaient en correlation avec celles des ions me'talliques fibres,
         mais pas avec celles du metal total. Des isothermes de Langmuir ont servi a calculer lesconstantes conditionnetfes
         d'equilibre metol-branchies ainsi que le nombre de sites de fixation dans le cas de chaque metal. Le log KMmncMe
         a ete evalue* a 7,4 et le nombre de sites de fixation du Cu sur une paire de branchies (70 mg poids frais) &ait
         d'environ 2 x 10~>mol (-30 nmol • g poids frais'1}. Le log Khunanchfea &6 evalue" a environ 8,6, et le nombre de
         sites de fixation de Cd sur Ies branchies £tait d'environ 2 X 10~10 mol (-2 nmol • g poids frais-'). Les constantes
         de stability des interactions H* et Ca sur Ies sites de fixation des mdaux avec le carbone organique dissous (COD),
         ont ete evaluees a partir de ces constantes metoux-branchies, Toutes Ies constantes de stabilitl ont ft£ versees
         dans le programme de chimie de I'eau MINEOL* dont on s'est servi pour prevoir {'accumulation des metaux sur
         les  branchies a partir des concentrations des mftaux, du COD et du Ca ainsi qu'a partir du pH de I'eau.
         L'accumulation catculte de me'taux sur les branchies Itait en bonne correlation avec la concentration de me'taux
         mesuree sur les branchies ainsi qu'avec  les valeurs de CL». Notre methode d'insertion de donnees biologiques
         dans un programe de chimie de I'eau cpnstitue une facon utile de modeliser et de prevoir ('accumulation des
         metaux sur les branchies, done de la toxicit£ de ces me'taux pour les poissons.
          Rece/ved September 18,1992
          Accepted May 18, 1993
          (JB635)
                      Refit le ISseptembre 1992
                         Accept le 18 mat 1993
   It has long  been realized .that metal-organism stability
   constants would be useful in interpreting and predicting the
   toxicity  of  waterbome  metals  to aquatic organisms
(Biesingcr and Christensen 1972; Zitko and Carson 1976). To
date, conditional stability constants for Cu have been determined
for algae (Xue  and Sigg 1990) and for completing  agents
released by algae (Van Den Berg et al. 1979; Xue and Sigg 1990)
and zooplankton (Fish and Morel 1983). These Cu stability
   'Author to whom correspondence should be addressed. Current
address: Department of Biology,  Wilfrid Laurier University,
75 University Ave. West, Waterloo, ON N2L 3C5, Canada.

2678                  .         '
constants fall in the range 107-10U, indicating that 50% of the
Cu binding sites would be occupied at aqueous Cu concen-
trations of 10-7-10-» M.
  For fish, Reid arid McDonald (1991) determined metal-gill
stability constants for La, Cu, Cd, and Ca, using excised gills of
rainbow trout Their values, lO^-lO", were determined using
radiolabelled metals in mM metal solutions. These values are
low compared to those  for algae, most likely because of the
relatively high metal concentrations used. The conditional nature
of stability constants makes them dependent on the metal con-
centrations used  during their  determination:  high  metal

                            Can. J. Fish. Aquot Sci., Val. 50,1993 A

-------
   concentrations yield low stability constants because even low
   affinity sites are able  to bind excess metals. At low metal
   concentrations, only high  affinity sites are able  to bind the
   limited amount of metal, yielding higher stability constants.
    Strength of metal binding is not the sole determinant of metal
   interactions with aquatic organisms. Dissolved organic carbon
   (DOC), Ca concentration, and pH are important modifiers of
   metal bioaccumulation  and toxicity in freshwater organisms
   (Yan et al. 1990; Meador 1991; Playle et al. 1992,1993). Dis-
   solved organic carbon  can complex metals,  rendering them
  unavailable to interact with an organism, and calcium competes
  with metals for binding sites, as does H*. In addition, water pH
  determines the proportion of total, unbound metal that is in the
  cationic form, generally thought to be the toxic species of metals
  (Morel 1983; Pagenkopf 1983).
    The objective of the present study was to determine metal-gill
  stability constants for low concentrations of Cu and Cd, and to
  use  these constants, plus  metal-DOC, Ca-gill, and  H-gill
  constants, in a computer model to predict metal accumulation on
  fish gills. This approach has been used for algae (Xue and Sigg
  1990) but has not yet been applied to fish. Metal accumulation
  on minnow gills in the presence of complexing ligands such as
  EDTA were determined  in our preceding paper (Playle et al.
  1993). We used the lowest concentrations of Cu and Cd (17 and
  6 jig • L~', respectively) that were feasible to yield measurable '
  metal deposition on fish gills, using graphite furnace atomic
 absorption spectroscopy. The gill accumulation data plus cal-
 culated free-metal concentrations in the water were used to
; determine the metal-gill stability constants. Free metal con-
 centrations were calculated using MINEQL* (Schecher 1991),
 an aquatic chemistry computer program based on equilibrium
 constants. Finally, our estimates of the metal-gill stability con-
 stants were assessed by comparing calculated metal accumu-
 lation on gills with measured gill metal concentrations of
 minnows exposed to Cu and Cd in water from a series of Ontario
 lakes.                   •.

 Materials and Methods

 Ligand  Experiments

   Detailed descriptions of the ligand experiments are given in
the preceding  paper (Playle et al. 1993) and the results are
summarized in Tables 1 and 2 of the present paper. In brief,
acclimated adult fathead minnows (Pimephales promelas) were
exposedto 17j»,gCu -L'1 (0.27(jAf)or6u,gCd- L-' (0.05jiM)
in synthetic soft water (Na, Ca -50 uA!) at pH 6.2,19°C. Fish
were  killed at the end of the 2-3-h exposures, their whole gill-
baskets were removed, weighed, then digested in 1 N H2SO4 for
8 h at 80°C. The supernatant of the gill digests was diluted with
deionized water, then analyzed for Cu and  Cd using graphite
furnace  atomic absorption spectroscopy. Experiments were run
in the presence or absence of freeze-dried dissolved organic
carbon (DOC, 0-8 mg-L"1)  or the synthetic ligands  EDTA,
nitrilotriacetic acid (NTA), ethylenediamine (EN), and citric,
glutamic, oxalic, and salicylic acids (0-2500 u,M).

Calculation of Gill Stability Constants
  We used Langmuir isotherms to calculate Cu- and Cd-gill
stability constants (KC^M, Kci^. In order to make these calcu-
lations, estimates of free Cuz* and Cd1* concentrations'in the
synthetic soft water, with and without added ligands, were made

Can. J. Fish. Aquat. Sci., VbL 50,1993
   using the MINEQL* program. Input values were 0.27 jiM Cut
   0.05 pM Cd, 70 jiM Na, 35 M.M Ca, fixed pH 6.2,19°C, anil
   concentrations of the added ligands (0-2500 jiM).  Solution
   ionic strength was calculated by MINEQL* (I~ 1 X 10'4), an
-------
                  TABLE 1. Gill Cu concentrations (above background) of fathead minnows held for 2-3 h in soft water
                  containing 0.27 fiM total Cu, plus ligands. Summarized from Playle et al. (1992,  1993). EN =
                  ethylenediatnine.
Ugand FreeCu2* GillCu logtfCu-ligand
((iM) ((iM;MINEQL+ calculation) (junol-g'1 wet tissue) (from MINEQL*)
EDTA


NTA
1
EN


Citrate


Oxalate


Glutamate


Salicylate




No ligands
No ligands.
0.15 '
0.25
0.30
0.25
2.5
0.25
2.5
25
0.25
2.5
25 ,
0.25
2.5
25
0.25
25
25
0.25 -
2.5
25
100
250
•

no added Cu
0.112
0.019
0.0002
0.030
0.0002
0.214
0.080
0.006
0.230
0.122
0.007
0.246
0.199
0.044
0.247
0.210
0.101
0.252
0.249
0.225
0.176
0.127
0.253

0.028
0.019 18.8
0.000
0.000
0.000 13.1
0.000
0.036 10.5
0.017
0.013 '
0.027 7.3
0.020
0.008
0.018 5.1
0.024
0.008
0.038 8.3
0.034
0.020
0.021 • 10.6
0.024
0.027
0.020
0.012
0.026 —
•
0.000 —
                  TABLE 2. Gill Cd concentrations (above background) of fathead minnows held for 2-3 h in soft water
**!- '«-** ,, . J . \ * 4
Ugand FrceCd2* -
(|iM) (nM;MINEQL* calculation)
EDTA 0.05
0.10
0.25
0.5
5.0
NTA 0.1
0.5
14.4
4.83
1.41
0.64
0.05
48.7
46.4
,, GillCd logKCd-ligand
(nraol • g~ ' wet tissue) (from MINEQL*)
2.0
2.4
0.4
0.0
0.2
2.3
1.6
16.3
. •


'
9.4

                  EN          0.05
                  Citrate        0.1
                             250
                            2500
                  Oxalate       0.05
                  Glutamate     0.05
                  No ligands
                  No ligands,
                    no added Cd
49.3
49.1
 2.60
 0.24
493
49.3
49.3
 0.99
.2.7
 1.2
 0.4
 2.1
 2.2
 2.0

 0.0
5.6
5.3

3.4
4.8
can be substituted for Cd. Implicit in these calculations is the
assumption that there is enough metal and time for the gill sites
to become saturated. In practice, the metal concentrations and
exposure  times used in  our experiments adequately fulfilled
these assumptions (Playle et al. 1992,1993).
               Lakewater Exposures and Toxicity Tests
                 All stability constants were entered into the MINEQL* aquatic
               chemistry program, to calculate metal accumulation on fish gills.
               To test the model, we wanted to measure metal accumulation on
2680
                                             Can. J. Fish. Aquat. Sci., Vol. SO, 1993

-------
 s
 s
 I
                                                  0.06
                                                  0.04
                                                  0.02
                                                       3
                                                       i
                    0.1
                                  0.2
                                                0.3
FIG. 1. Plot of measured gill Cu against free Cu2* concentrations. Gill
Cu concentrations were from fathead minnows exposed 2-3 h in
synthetic soft water with 17 u,g Cu • L"1 (0.27 jiM) and 0-250 u.M of
synthetic ligands (data from Playle et al. 1992,1993). Free Cu2* was
calculated using MINEQL* and known concentrations of ligands and
ions (Table  1). Gill Cu varied directly with free Cu2* (r=0,858,
/><0.001).The equation of the line was y=6.23*+0.86 (y in jigCu • g
tissue'1) ory=0.09&C+0.014 (y in u,mol • g tissue'1).
gills when minnows~were exposed  to  17 fig Cu • L'1 and
6 |tg Cd • L"1 in natural soft water, as opposed to synthetic soft
water. In addition, we wanted to correlate metal accumulation
on gills with metal toxicity.
  Water was collected from five lakes ME of Peterborough,
Ontario, in June, 1991. The lakes were: Salmon L., Brooks Bay
and Sharps Bay of Jack L., Anstruther L., Loucks L., and Wolf L.
(all approximately N44°42', W78*13'). Twenty-L stainless steel
containers (Spartenburg Steel Products, Spartenburg, SC) were
used to collect the surface water.  These  containers were then
pressurized with N2 gas  and the water  was filtered through
1.0 jim glass fibre filters (A/E, 142 mm, Gelman Sciences Inc.,
Ann Arbor, MI) to remove algae. The filtered water was stored
in 40-L stainless steel containers  (Simgo Inc., Toronto, ON),
brought back to Waterloo, transferred to 4-L polyethylene con-
tainers, and stored at 4°C. These water samples were used for
combined Cu and Cd metal exposures to adult fathead minnows,
conducted in the same manner as described for the synthetic
softwater experiments. A synthetic soft  water plus a 10-uM
(total)  mixture of glutamic, citric, and oxalic acids was also
used (3.3 |iM of each), to mimic  a 5 mg DOC • L'1 solution
(Campbell and Stokes 1985).
  Toxicity tests using a geometric series of concentrations of the
Cu and Cd mixture were run in the lake waters, using juvenile
fathead minnows. These fish were <1 mo old, and had been held
in soft water for 1 wk before the experiment. During acclimation
they were fed live brine shrimp and  fine powdered fish food
twice daily. Ninety-six h LCSOs were determined using duplicate
exposures of five fish per 100-mL solution in un-aerated poly-
ethylene urine cups. Eighteen juveniles (blotted dry) weighed.
individually yielded an average weight of 1.6 ± 0.5 mg. LC50
values were calculated using a trimmed Spearman-Karber analysis.


Results

Calculation of Metal-Gill Stability Constants
Copper                                  .
  Previously, we measured Cu accumulation on gills of fathead
minnows exposed to 17 ftg Cu • L"1 (0.27 u,M) in synthetic soft
water,  in the presence or absence of synthetic ligands (Playle
et al. 1992,1993). Because we knew the water chemistry and the

Can. J. Fish. Aquat. Sri.. VW. 50,1993
                                                                 0.4
  3
  I
  S
  i
     0.2
                                                                                                                      I
                                                                                  20
                                                                                                40
                                                                                                              80
                                                                                     freeCd"(nM)
FIG. 2. Plot of measured gill Cd against free Cd2* concentrations. Gill
Cd concentrations were from fathead minnows exposed 2-3 h in
synthetic soft water with 6 jig Cd • L'1 (0.05 |*M) and 0-2500 |iM c'f
ligands (data from Playle et al. 1993). Free Cd2* was calculated usinj
MINEQL* and known concentrations of ligands and ions (Table 2). Gill
Cd varied directly with free Cd2* (r=0.732, P < 0.01). The equation of
the line was y ~ 0.004* + 0.163 (y in ng Cd • g tissue'1) ory = 0.03ijr
+ 1.451 (y in nmol • g tissue'').


ligand concentrations, we were able to calculate free Cu concen
trations using the MINEQL* program. Gill Cu concentration!;
varied directly with free Cu2*, the major component of free Cu"
gill Cu versus free Cu2* gave a straight line (Fig. 1; r=0.8581
P  < 0.001). Gill Cu concentrations did not correlate with total
Cu, which was constant at 0.27 uM                        11
  To calculate log £0,1111 and the number of Cu binding sites on
the gills, a Langmuir plot of Cu adsorption was constructed. Free
Cu2* (micromolar) divided by gill Cu (micromoles Cu per grant
wet tissue, minus  background gill Cu, Table 1) was plotted!
against free Cu2* (micromolar). The equation for the line wal
y a 35.38*  + 1.28.  The inverse of the slope of the line is the!
number of gill binding sites, = 0.03 |imol  • g wet tissue'1. The
inverse of the intercept = X> (binding site number), therefon
K  = 26 L • (tmoi'1 = 26 X 10* L •  mol'1, and log K^n = 7.4
Average total gill weight for the minnows was 0.07 g (wet tissue
Playle et al. 1993), therefore total binding sites for Cu on the gill
of our fathead minnows was about 2X10~9 mol.

Cadmium
  Accumulation of Cd on gills of fathead minnows exposed t<
6 (Jig Cd • L'1 (0.05 \iM) in the presence or absence of synthetu
ligands (from Playle et al. 1993) was plotted against free Cd2
concentrations calculated using MINEQL* (Fig. 2). Althougl
there were not as many data points as there were for Cu, gill G
still varied directly with free Cd2* (r * 0.732, P < 0.01). It wa
difficult to obtain intermediate values for the curve, because st
little Cd was used in the experiments (e".g., 0-50 nM is jus
one-fifth the scale for Cu in Fig. 1).
  The log Kca-pii and the number of binding sites for Cd were
calculated using a Langmuir plot. Gill Cd concentrations above
background (Table 2) were used for these calculations.
Langmuir  isotherm,  calculated in nmoles, yielded the H  ^
y = 0.44*+1.06. The number of gill binding sites was the inverse^
of the slope, = 2.27 nmol • g wet tissue'1. From the inverse o]j
the intercept, K = 0.42 L • nmol'1 = 0.42  X  10' L • mol-1, and
         , = 8.6. Average gill weight was 0.07 g wet tissue; thui
                                                             the number of gill binding sites for Cd was about 2 X 10 I0 mo
                                                             per minnow.
                                                                                                                 268

-------
                  TABLE 3. Input data for the metal-gill interaction model. Initial log K values were as calculated in the text.
                  Final log AT values were determined using MINEQL4 to best fit, within the constraints of the initial log K
                  values, calculated gill metal with measured gill metal concentrations for the synthetic soft water system.
                   Binding sites
Complex
Initial log*
Final tog K

cu:giii
2X KT'riiol per fish

Cd-gill
2X10 10mol per fish

DOC ,
Img-L «5X10~8mol


Cu-gillcu
Hf-gillcu
Ca-gillcu
Cd-gillcd
tf-gillcd
Ca-gillcd
Cu-DOC
cd-Doc
H+-DOC

7.4
<5.6
<3.5
8.6
>6.l
5.0
8.4-9.4 "
6.4-8.4
4.0

7.4
5.4
3.4
. 8.6
6.7
5.0
9.1
7.4
4.0
Calculation of Ca, H*, and DOC Stability Constants
  Calcium and H* can compete with Cu and Cd for binding sites
on gills. By using the log Aincoi-^ii values calculated above, and
data presented in Playleetal. (1992,1993), the stability constants
for Ca and H* were estimated.
  Copper accumulation on minnow gills was not affected at
pH 4.8 (15.8  |iM H*) compared to pH 6.3 (0.5 nM H*; Playle
et al.  1992). Inserting K^^a = 10M, free Cu = 2.53 X 10~7 M
(MINEQL*   calculation),   and   H*= 15.8X10-'  M  into
Equation 5 yields log KH^n < 5.6. Copper accumulation was not
affected by 2 000 u,M Ca (Playle et al. 1992): using Equation 5,
Cu = 2.53 X 10-7M,Ca=2 X 10-JM,and£o*ai= 107-4, yields
  Cadmium accumulation on the gills was reduced at pH 4.8
(Playle et al.  1993).  Inserting K^n  =  10", free Cd =
4.93 X10-*  M   (MINEQL*  calculation),   and  H*  =
15.8 X lO'6 M into Equation 6 yields log KH^a > 6.1. ForCa,
95 \tM did  not reduce Cd accumulation on  gills, whereas
1050 nM Ca did (Playle et al. 1993). Using Equations 5 and 6,
respectively, log K&^u < 5.3, but >4.3, the mean of which is
l°g KCHHI ~ 5.0.
  Five to 6 mg DOC • L~' prevented Cu accumulation on gills,
by complexing all the 0.27 pM Cu in solution (Playle et al.
1993). Thus, there were at least 0.05 u.mol  binding sites per
mg DOC • L"'. These metal binding sites must have had a higher
affinity for Cu than did the Cu binding sites on the gill, or the
DOC would not have been able to prevent Cu accumulation on
the gills. The log KCU-DOC value for the 0.05 u,mol of high affinity
binding sites per mg • L'1 DOC must therefore be about 1-2 log
units higher than the log KC»^\ = 7.4 value. Our initial estimate
of the stability constant was log K^DOC - 8.4-9.4, which is in
the mid-range of published values (Van Den Berg and Kramer
1979;Morel 1983). Note that direct attack of the Cu-gill complex
by DOC is unlikely (adjunctive pathway of ligand exchange; see
Hering and Morel  1990). More likely is the disjunctive pathway,
that first involves  dissociation of Cu from the Cu-gill complex,
the free Cu then being able to react with DOC in solution.
  Cadmium has consistently been reported to bind <10X as well
to DOC  than does Cu (Alberts and Giesy 1983; Tuschall and
Brezonik 1983). Thus, an initial estimate of the log ATca DOC value
of 10-100X less than for Cu would be 6.4-8.4. The log AT for
humic substances  is 3 to 4 (DeWit et al. 1991), therefore we set
log £H-DQC = 4. We had no information on the interactions of Ca
with DOC.
2682
          Model Development
            Our purpose in determining metal-gill and other stability
          constants was to use these values in a computer model to predict
          metal accumulation  on fish gills, and thereby predict metal
          toxicity. In essence, bringing biology into an aquatic chemistry
          program. The model we chose to use was MINEQL* (Schecher
          1991), an aquatic chemistry program based on log K stability
          constants.
            Our initial estimates of log K stability constants and the
          number of metal binding sites are summarized in Table 3. To
          input these data into the MINEQL* program, three null com-
          ponents were defined (gillcu, gillcd, and DOC). We simulated
          two different sites for Cu  and  Cd binding on the gills (gillcu,
          gillcd) on the basis of our previous results (Playle et al. 1993)
          that indicated, for the low metal  concentrations we used, no
          competition between Cu and Cd for gill binding sites. Metal, Ca,
          and H* interactions with gillcu, gillcd, and DOC were defined in
          the program as 1:1 reactions (Table 3).
            We used Na = 70 fiM, Ca = 35 jiM, Cd = 0.05 uM, Cu =
          0.27 M-M, fixed pH 6.2, and 19°C for water chemistry parameters
          of the synthetic soft water..For log K values for which we just
         . had upper or lower values, or a range, we chose final log K values
          after optimizing to best fit the observed data in Playle et al. (1992,
          1993). In essence, the model was able to mimic the greater effect
          of H* and Ca on Cd binding to gills, and the greater effect of
          DOC on Cu binding to gills. Of note in Table 3 are die final
          log ^mem-roc values, that we optimized to log £0*00: = 9.1 (in
          the model, Cu binds to DOC 50 X better than it binds to the gill),
          and log KCVDOC - 7.4 (in the model, Cd binds to DOC 50X less
          well than does Cu, and Cd binds to DOC 16X less well than Cd
          binds to gills). We chose 1 mg • IT1 DOC = 0.05 jxmole binding
          sites,  to allow enough binding sites for all the Cu and Cd in
          solution at 6 mg DOC • L~l (e.g., no competition by the metals
          for the sites). In general, as long as a ligand is in excess relative
          to metals in solution, metals do not compete with each other
          (Morel 1983).

          Metal Accumulation and Toxicity to Fish Exposed to Cu and
              Cd in Soft Lake Water
            Of the natural lake waters that we used in the Cu  and Cd
          exposures, Salmon.L., and Brooks and Sharps bays contained
          more  Ca, more DOC, and had higher pH than did Anstruther,
          Loucks, and Wolf lakes (Table 4). The synthetic soft water plus
          a 10 u,M (total) addition of glutamic, citric, and oxalic acids had
          water chemistry similar to that of synthetic soft water without

                                       Can. J. Fish. Aqual. Sci, Vol. 50,1993

-------
TABLE 4. Lake water and synthetic soft water characteristics, and juvenile fathead minnow 96 h LCSOs. Water characteristics were measured
^-/Aq JfpNrtA^S- 96hLC50 I
Water
Salmon Lake

Brooks Bay

Sharps Bay

Anstruther Lake

LoucksLake

Wolf Lake

SSW+ ligands

SSW

pH
8.05+0.01
(12)
7.80+0.01
(12)
7.90+0.02
(12)
7.10+0.03
(12)
7.04+0.01
(12)
6.91+0.03
(12)
6.60+0.03
(8)
6.69+0.01
(20)
DOC
6.4±0.1
(3)
9.8
(2)
7.9
(2)
7.2+0.6
(3).
6.6
(2)
6.0+0.2
(3)
1.8
(2)
25
(2)
Ca
695+20
(4)
445±10
(4)
590+20
,(4)
128+2
(4)
92+2
(4)
94±1
(5)
36+1
(3)
37±1
(H)
' j
Na
34+6
(4)
76±9
(4)
90±32
(4)
39±5
(4)
37±4
(4)
48±12
(4)
87+23
(3)
91 + 11
(ID
Cond.
85

—

78

24

22

19

—

_

Cu
2.3

3.1

6.6

4.1

1.4

0.8

—

2.1 ±0.3
(32)
Cd
0.1

0.1

<0.1

, 0.1

6.3

0.3

—

<0.1
(33)
Al
10.7

12.8

6.5

12.8

13.3

9.6

—

4.4

Fe
<50

<50

<50

<50
>
<50

<50

_'

<50

Zn
12.0

4.0

2.5

15

4.2

1.6

—

-2.8±OJ
(10)
Cu
>126

102
(78^133)
115
(85-156)
50.0
(31.9-54.4)
27.8
(18.4-42.0)
31.2
(21.9-44.6)
10.3
(6.8-15.6)
9.4
(7.6-11.5)
Cd
>32

25.3
(19.1-334)
27.8 ||
(19.8-39)0)
10.4 || ,
(7.8-14JO)
6^ 1
(4.2-1 IJI)
6.9 ||
(4.0-11.'))
1-1
(0.4-3.01
1.0
(0.6-1.7]
added ligands. Minnow exposures in both these synthetic soft
waters were at pH -6.7, to match the pH of the softer of the
natural waters.
  Copper- and Cd 96-h  LC50 values  for juvenile fathead
minnows exposed in the various waters are given in Table 4. In
general, LCSOs for the Cu and Cd mixtures were lowest (metals
were roost toxic) in softer water. Metals were most toxic in the
synthetic soft water, and  were only slightly less toxic in the
presence of the 10 pJVf ligand mixture. Metals were least toxic
in water from Salmon L., with too few mortalities to accurately
determine the LC50 (Table 4).
  Cadmium accumulation on gills of adult minnows exposed to
6 M.g Cd - L"1 and 17 (ig Cu  • L"1 was highest in Wolf L. water
(significantly above background, and not significantly different'
from metal-exposed fish in synthetic soft water alone; Fig. 3),
and lowest in Salmon L. water (not significantly above back-
ground). Fish exposed to the lake waters without added metals
did not have gill Cd concentrations significantly above back-
ground (0.18 ± 0.02 u,g Cd - g wet tissue'1; ±95% CI, n = 17).
Fish exposed to Cd and Cu in synthetic soft water plus 10 u-M
(total) glutamic, citric, and oxalic acids had 0.64 ± 0.23 (6) u-g
Cd • g wet tissue'1, and fish exposed to the metals in synthetic
soft water without added ligands had gill Cd concentrations of
0.54 ± 0.05 (9) jig Cd • g. These exposures were run at pH 6.6
and  6.7, respectively; their gill Cd concentrations were not
significantly different from those offish exposed in synthetic
soft water at pH 6.2.                                 ''
  Copper accumulation on  gills of adult minnows was never
above background in fish held in lake waters supplemented with
Cu and Cd (Fig. 4), probably because DOC exceeded 5 mg • L"'
in all lake waters (see Playle et al.  1993). In lake water without
added metals, gill Cu was 0.80 ± 0.08 (17), also not significantly
above background. Minnows exposed to Cu and Cd in synthetic
soft water plus the 10 jiM ligand mixture (pH 6.6) had gill Cu
concentrations of 2.58 ± 0.18 (6) (ig Cd: g wet tissue'1, while
fish exposed to the metals in synthetic soft water without added
ligands (pH 6.7) had gill Cu concentrations of2.16 ± 0.30(9) u,g
Cu • g"1. Neither concentration was significantly different from

Can. J. Fish, Aqaat. Set., Vol. SO. 1993
             S
                          Lake
FIG. 3. Gill Cd concentrations of fathead minnows exposed 2-3 h in Cd-
and Cu-supplemented lake .water. Cadmium deposition on gills was
highest in the softest water. S=Salmon L., B=Brooks Bay, SH=Sharps
Bay, A=Anstruther L., L=Loucks L., and We Wolf L. Vertical lilies on
bars=95% CI; n=6 for each bar. Dashed horizontal lines represent the
95% CI of gilt Cd for minnows held hi synthetic soft water in the absence
of added metals (n = 38). The grey horizontal band represents the 95%
CI of gill Cd for fish  held 2-3  h  in  synthetic soft water with
6 u.g Cd • L"1 (» = 45). *, **, *** = significantly below metal-expose*
gills (grey band), and +, ++, +++ = significantly above background gill
metal concentration  (dashed lines)  for P< 0.05, <0.01, <0.001,
respectively^
gills of minnows exposed to Cu in synthetic soft water at pH 6.2.

Model Testing

  Final values for the log K stability constants (Table 3) were
used to assess the usefulness of the model in predicting gill metal
concentrations and metal toxicity.  In the model we  used the

                                                       2683

-------
                                              CafeiM)
                                                    (mg-L")
SH
                                         W
FIG. 4. Gil! Cu concentrations of fathead minnows exposed 2-3 h in Cd-'
and Cu-supplemented lake water. There was no significant accumu-
lation of Cu on die gills, likely because DOC was a 6 mg • L~' in each
lake water. Dashed horizontal lines represent the 95% CI of gill Cu for
minnows held in synthetic soft water alone (n=38). The grey horizontal,
band represents the 95% CI of gill Cu for fish held 2-3 h in synthetic
soft water with 17 ng Cu • L"1 (n = 45). Lake names and other details
are given in caption of Fig. 3.
   140
FIG. 5. Measured and modelled gill Cd, for fathead minnows exposed
2-3 h to 6 |ig Cd • L"' and 17 u.g Cu • L~' in natural soft waters and
synthetic soft water (SSW) with or without 10 ji.M added ligands (at'
pH 6.7). Lakes are defined in the caption of Fig. 3. Gill Cd is expressed
as a percentage of the mean gill Cd concentration (100%) for fish held
in synthetic soft water plus Cd  at pH  6.2 (grey bar, ±95% CI).
Background gill Cd = 0% (+95% CI, dashed line). Calculated gill Cd
correlated well with measured gill Cd (see text for details). Measured
gill Cd values are given with then-95% CI.
DOC, Ca, and Na values (molar), and pH (fixed) for the eight
waters listed in Table 4. The model mimicked a system open to
atmospheric CO2. Although we did not measure CO32~ in the lake
waters, for the model we assumed its concentration was equal to
that of Ca (Table 4), which yields approximately the same values
as if the more rigorous Henderson-Hasselbalch equation were
used to calculate HCO3~. For the comparison of measured and
calculated gill Cd and Cu, measured values from Fig. 3 and 4
were converted to percentages, with background gill metal=0%,
and gill metal concentrations of fish exposed to metals in
synthetic soft water (at pH 6.2) = 100%. The calculated gill metal
concentration in synthetic soft water (pH 6.2, assuming back-
ground DOC = 1 mg • L~') was used as the model 100% value.
                                                              •5
                                                              £
                                                              3
                                                                                      Lake
                                      FIG. 6. Measured and modelled gill Cu for fathead minnows exposed
                                      2-3 h to 17 u.g Cu • L'1 and 6 u,g Cd  • L'1 in natural soft waters and
                                      in synthetic soft water (SSW) with or without added ligands. Lakes are
                                      defined in the caption of Fig. 3. Gill Cu is expressed as a percentage of
                                      the mean gill Cu concentration (100%) for fish held in synthetic soft
                                      water plus Cu (grey bars). Background gill Cu = 0% (dashed line).
                                      Calculated and measured gill Cu correlated well (see text for details).

                                        Modelled gill Cd agreed well  with  our measured values
                                      (Fig.  S), with the largest discrepancy for Sharps Bay. The cor-
                                      relation coefficient between measured and modelled gill Cd was
                                      0.9008 (P < 0.01). Measured gill Cd versus the juvenile fathead
                                      minnow LCSOs for Cd (Table 4) yielded a correlation coefficient
                                      of -0.897 (P < 0.01). Modelled gill Cd versus the LC50 values
                                      yielded a better correlation (r = -0.960; P < 0.001). These data
                                      indicate that model-calculated gill Cd concentrations could rea-
                                      sonably be used to predict acute Cd toxicity to fish.
                                        Measured and modelled gill Cu also  agreed well (Fig. 6;
                                      r - 0.926, P < 0.001), in spite of the fact that none of the gill Cu
                                      concentrations from the natural lakewaters  was above back-
                                      ground (Fig. 4). Measured gill Cu versus the log of the juvenile
                                      fathead minnow LC50s for Cu yielded r= -0.830 (P< 0.05; gill
                                      Cu vs. linear LCSOs was not significant, r = -0.662, P > 0.05).
                                      Modelled gill Cu had a slightly better correlation with the log
                                      LCSOs (r = -0.895, P < 0.01; the correlation with linear LCSOs
                                      was r = -0.727, P < 0.05).  As was the case with Cd, modelled
                                      gill Cu concentration was a  good indicator of acute Cu toxicity.

                                      Discussion

                                        In the present study, we  have better defined interactions of
                                      Cu and Cd with fish gills, as influenced by dissolved organic
                                      carbon (DOC), pH, and Ca. We used Langmuir isotherms to
                                      estimate conditional metal-gill stability constants. These were
                                      log Kca^a - 7-4 and log KCA^= 8.6. There were about 10X more
                                      of the weaker Cu sites on the gills (-2 X 10~9 mol per fish) than
                                      Cd sites (~2 X lO"10 mol per fish).
                                        Interactions of Cu and  Cd with the gill and  with DOC, and
                                      competing reactions with Ca and H*, are illustrated conceptually
                                      in Fig. 7, along with their log K values. Cadmium binds to the
                                      gills better than does Cu, and is more affected by competition
                                      from  Ca and H4. These interactions are likely a consequence of
                                      active Cd uptake through high affinity Ca channels in addition
                                      to general surface binding. Cadmium reduces Ca uptake, mainly
                                      through effects on basolateral Ca2*-ATPase activity (Verbost
                                      et al.  1989). The data of Reid and McDonald (1988) support the
                                      existence of higher affinity binding sites for Cd compared to Cu.
                                      Calcium competition reduces Cd uptake at gills, and therefore
2684
                                                                                           Can. J. Fish. Aqaal. Sci.. W. 50,1993

-------
  the subsequent distribution to the rest of the fish body (Wicklund
  and Runn 1988). Transfer  of metals through the basolateral
  membrane is probably by diffusion (Verbost et al. 1989).
    Copper, on the other hand, appears to bind to weaker sites on
  the gills, and possibly also enters through ion channels (Fig. 7),
  but binds more strongly to DOC than does Cd (log XCU-DOC=9.1;
  log Kctwc = 7-4)- The binding order for DOC reflects relatively
  strong binding of Cu to ligands. The effect of low concentrations
  of Cu at die gills is an inhibition of Na* influx, mainly through
  effects on Na+-K*-ATPase (Lauren and  McDonald  1987).
  Relatively more inhibition of this effect was seen by carbonate
  complexation than by Ca (Lauren and McDonald 1986), which
  agrees with our results of less protective effects of Ca on Cu
  accumulation, compared to effects on gill Cd (Playle et al. 1992,
  1993). Note that in the conceptual model illustrated in Fig. 7 the
  gill mucus layer is  ignored.  Gill mucus probably binds metals
  and slows metal access to the gill (Part and Lock  1983).
  However, once a metal exceeds the completing and sloughing
  ability of mucus, the metal will become available to react at the
  gill epihelium; this is the situation illustrated in Fig. 7.
    There are few log KB*^,, values with which to compare our
  values. Reid and McDonald (1991) found log AT values of about
  3.5,3.0,3.0, and 2.4 for La, Cd. Ca, and Cu, respectively. These
  values are low, probably because they used mM concentrations
  of metals  in their experiments  instead of environmental,
  sub-micromolar concentrations. Log XT values are sensitive to the
  metal concentration at which the experiments are run, and it is
  difficult to extrapolate from a stability constant obtained at high
  metal:ligand ratio to a stability constant for a low metahligand
  ratio (Perdue and Lytle 1983). However,  Reid and McDonald
  (1991) found that Cd bound to gills better than did Cu, which'
 agrees with our results. Xue and Sigg (1990)  determined
  log tfciMipe - 8-8 (at pH 6.0), which is -10X higher than ours for
  fish at about the same pH. These workers were successful in model-
  ling Cu binding to algae using a computer program based on
 log K values, their input log  KQ,^^ values, and the number of
 binding sites (2 funol •  g dry  algae'1) they determined for algae.
   We inserted our metal-gill, H*-gill, Ca-gill, and metal-DOC
 stability constants (Fig. 7; Table 3) into MINEQL+ (Schecher
  1991). The model adequately reflected measured metal accumu-
 lation on the gills (Fig. 5,6).  In addition to Cu and Cd bound to
 gills, the MINEQL* output included Ca and H* bound to gill sites
 (an indication of competition for the sites),  unoccupied gill sites,
 free metal in solution (an indication of the pool of metal available
 to bind at the sites),  and metal bound to DOC (representing the
 extent of metal complexation). This simple model considered
 only 1:1 metal-ligand interactions, and ignored Ca interactions
 at metal binding sites on DOC.  At the low  metal concentrations
 we used, 1:1 binding would be expected: Cabaniss and Shuman
 (I988a) showed that  Cu-DOC interactions were >90% 1:1
 binding.  Not incorporating Ca-DOC  interactions probably
 would have had little effect on the model, because both Cu and
 Cd would bind much more  strongly to DOC than would Ca
 (Kemdorff and Schnitzer 1980; Cabaniss and Shuman 1988a;
 Daly et al. 1990). It should be noted that our stability constants
 represent average conditional stability constants, akin to con-
v tinuous multiligand binding models (e.g., Perdue and Lytle 1983;
 Grimm etal. 1991) or "quasiparticle" models (Sposito 1981).
    Stability constants change  with solution chemistry (pH, ionic
 strength) as well as with metal and ligand concentration (Perdue
 and Lytle 1983; Cabaniss and  Shuman 1988a). Van Den Berg
  and Kramer (1979)  and Morel  (1983) determined that there is

  Can. J. Fish. Aquat. Sri.. Vol. 50.1993
 Cd'
                 Cu
        DOC
             9.1
                           Cu—JA
     4.0
 Cd
   \
     7.4
       X
        DOC
          V
             4.0
              \
    Ca*
 FIG. 7. Conceptual illustration of Cu, Cd, Ca, and H* interactions at fish
 gills and with dissolved organic carbon (DOC). The numbers are the log
 K conditional stability constants for the interactions. In water (left), Cu
 and Cd may become complexed by DOC, and also face competition
 from Ca and H+ for negative binding sites on gill surfaces. As well as
 binding to the general gill surface, Cu may enter the gill through low
 affinity channels. Copper disrupts Na*uptake at the basolateral Na
 pump. The stability constant for Cd is higher than for Cu: if surface
 binding was solely responsible for metal reactions at the gills, Cd wouM
 be expected to bind less strongly than Cd. Uptake of Cd through hig'L
 affinity Ca channels  is likely the .explanation  of the higher-thaii'-l
 expected Xbi-tiii. Also illustrated is the inhibition of the basolateral Ca
 pump by Cd.         .     '  , •
• about a one log unit decrease in stability constants as pH
 decreases by one unit However, we used the approach of Grimm
 et al. (1991) and Cabaniss and Shuman (1988a) where changes
 in stability constants are considered to be due solely to H|
 competition. That is, the use of one metal-gill stability constant!
 with another H* stability constant for competition, as opposed to
 different metal-gill constants  for  each pH. In spite of the
 qualifications and simplifications used in the model, relative
 amounts of Cu and Cd deposition on the gills were reasonable
 well predicted (Fig. 5,6). Copper was complexed better by DOC
 than was Cd, and Cd binding to the gills was affected more by
 Ca than was Cu.                                          I
   We calculated approximately 0.05 u,mol metal binding site!;
 per mg DOC. Reported number of Cu binding sites for DOC van/
 from about 10 iimol per mg DOC (Cabaniss and Shuman 1988a'>
 to 0.2 (imol per mg DOC (Hering and Morel 1990).  Our low
 number of binding sites is indicative of the low metal concen-
 trations we used in our experiments. The sites were relatively
 strong metal binding sites (log /^-DOC=9.1, log Acd-bbc=7-4)1
 which may represent metals binding with carboxyl groups (Tagil
 et al. 1991). Sunda and Hanson (1979) determined log ACU DOC f
 9.0 for river water at pH 5.95, very similar to our Cu result!
 Hering and Morel (1990) used 0.05' u,M Cu concentrations i
 their experiments, and found slightly stronger Cu binding t
 humic acid (log  ^cu-ten* «* = 10.1). Holm and Curtiss (1990)
 used 0.05,0.2, and 10 jiM Cu in their experiments, and found
 log K= 10.1,8.5, and 5.5, respectively, fornatural organic matter
 in ground water, demonstrating the dependence  of stability
 constants on the metal concentration used in their determination!
 Grimm et al. (1991) determined (mean) log KCM-DOC F 4.2-4.9, !i

                                           .'          2685

-------
  much lower stability constant than those listed above, but they
  used experimental Cu concentrations s5 (iM. Holm and Curtiss
  (1990) determined that complexes with natural organic matter
  should dominate Cu speciation at the low Cu concentrations they
  used. In agreement, our measured and modelled gill Cu values
  (Fig. 6) were mainly influenced by DOC concentration.
    Cadmium has been shown to bind less well than Cu to DOC
  and fulvic and humic acids (e.g., Florence 1977; Kemdorff and
  Schnitzer 1980; Alberts and Giesy 1983; Lund et al. 1990; Sahu
  and Banerjee 1990). This binding order is related to general
  metal-ligand binding strength. If Cu binds better to DOC and
  humic acids, then Cu toxicity would be expected to decrease in
  response to organic material more than does Cd toxicity (e.g..
  Winner 1985; 1986 for Cu, Cd toxicity to Daphniapulex). In our
  work,  freeze-dried DOC protected against Cu deposition on
  minnow gills at concentrations &4.8 mg-L"1 (Playle et id. •
  1993), and insignificant Cu accumulation was  observed  in
  natural lake waters containing £6 mg DOC -L'1  (Fig. 6),
  whereas Cd still bound  to fish gills in the presence of & 6 mg
  DOC • L~' (Fig. 5). Because of these binding results, we suggest
  that toxicity of the metal mixtures in soft waters (Table 4) was
  due mainly to Cd.
    In the  model we ignored differences  in DOC source or com-
  position, because concentration of DOC adequately determined
  gill Cu concentrations (Playle et al. 1993). Although DOC from
  different sources can bind or detoxify metals to varying degrees
  (e.g.. Nor and Cheng 1986; Sahu and Banerjee 1990), DOC from
  similar sources (Lei. same size of watershed and vegetation types)
  generally binds metals in a similar manner (Oliver et  al. 1983;
  Cabaniss and Shuman 1988b). Source  of DOC may not be as
  important as pH, alkalinity, or  ionic strength in  determining
  metal binding to DOC (Cabaniss and Shuman 1988b), and these
  workers suggested that experimental effort should be expended
 in defining these effects, not on documenting variations in DOC
 binding properties with season or location.
   Synthetic  soft water plus 10 jiM (total) citric, glutamic, and
 oxalic acids did not reduce Cu or Cd accumulation on fathead
 minnow  gills, nor did it alter metal toxicity  (Table  4). This
 mixture was meant to be a synthetic DOC analogue, similar to
 but simplified from those  used by Sposito  (1981) and by
 Campbell and Stokes (1985). Campbell and Stokes (1985) found
 that  a  22 (iM DOC analogue  complexed -70%  of a 6 p-g
 Cu • L~' solution (pH 6). With  3X that amount of Cu in our
 experiments, and half the iigand concentration, only about 12%
 of the Cu would be expected to be complexed, not enough to
 reduce Cu accumulation  on  gills. However, the  model did
 calculate some reduction  in free Cu (to about 60%  of total;
 Fig. 6). It is not surprising that the 10 jiM (total) Iigand solution
 did not reduce Cu or Cd bound to gills or reduce metal toxicity,
 and it does not adequately reflect 5 mg DOC • L"1 in our system.
   The model should be expanded to include other metals. It may
 be possible  to calculate stability constants from previously
 published Iigand work, but some metal-gill stability constants
 have already been published. For example, Wilkinson et al.
 (1990) used salmonid data to estimate log KM^u - 6.5. Con-
 ditional stability constants  for  organic material  and Al are
I available: log KAI-DOC = 6.2 (Urban et al. 1990), log J^AHM*.* =
 7.8 (Shuman 1992). Constants are also available for other metals
 such as  Mn (log  AWooc  =  3.8; Urban et al.  1990),  Fe
 (log Kft.Doc = 9.4; Urban et al. 1990), and Co (log KC**,.* =
 5-6; Van  Loon et al. 1992). Weaker complexing metals such as
 Ni and Zn would likely bind to DOC about 10X less well than -

 2686
  does Cu (Morel 1983). Lin and Benjamin (1992) indicated some
  of the confounding effects of Iigand addition to a multiple-metal
  system. A Iigand may complex a metal that is.otherwise bound
  to a (gill) binding site, which would free that site (e.g., increase
  free gill sites) to which another metal may bind. Models such as
  ours will be useful to predict complex interactions of this type.
   In an expanded model, competition for gill binding sites by Ca
  and H+need to be considered for each metal using appropriate
  log K values. Metals will be less toxic in hard water, and very
  acidic conditions will keep metals off gills. Of course, H+ itself
  can be disruptive to gills. Simulations using our model indicated
  that H* would protect against metal binding  at gills before it
  would displace metal from DOC. By the time enough metal was
  displaced from DOC for the metal to interact at the gills (pH < 4),
  H* itself would likely become toxic. This point illustrates the
  usefulness of the model in integrating multiple effects such as
  speciation, competition, and complexation, a situation  where
  models relying on correlation (Yan et al. 1990; Meador 1991)
  are often inappropriate.
   Output from the present model is toxicant bound to the gills,
  the target organ, which then must be translated into toxicity to
  the fish. This last step requires the most study. Together, the
 model output would give an indication of all types of competition
 for gill binding sites, of which Ca interactions can be considered
 beneficial, and most others as potentially toxic. In our lake water
 experiments, measured and calculated,gill'Cu and Cd concen-
 trations did correlate with toxicity, but it is not known how much
 metal accumulation is necessary to affect fish survival on a
 long-term basis. Presumably, a constant accumulation (gill
 "dose") will be toxic, such as was found  for whole-body Cd
 accumulation in/fya/«//aazreco(Borgmannetal.  1991).
   In summary, measurement of metal deposition on minnow
 gills has allowed us to estimate conditional stability constants of
 Cu and Cd binding to gills, the number of metal binding sites on
 the  gills,  plus interactions  .with Ca,  H*. and DOC. This
 information, and similar information for other metals, will be •
, useful in predictive models for metal accumulation on gills, and
 will therefore be useful to predict metal toxicity to fish.

 Acknowledgements

   We thank Drs. Chris Wood and Gord McDonald, McMaster Uni-
 versity, Hamilton, Ontario, for the use of their graphite furnace for metal
 analyses. We thank Dr. David Lean for help  collecting lake water
 samples. This research was funded by an Operating Grant (No. 8155)
 from the Natural Sciences and Engineering Research Council of Canada
 to D.G. Dixon, by a Research Grant from the Dorset Research Centre,
 Ontario Ministry of the Environment, to D.G. Dixon, and by an E.B.
 Eastbum Postdoctoral Fellowship to R. Playle.

 References

 ALBERTS, JJ., AND I.P. GIESY. 1983. Conditional stability constants of trace
    metals and naturally occurring humic materials: Application in equilibrium
    models and verification with field data, p. 333-348. In R.F. Christman and
    E.T. Gjessing fed.] Aquatic and terrestrial humic materials. Ann Arbor
    Science, Ann Arbor, MI.
 BIESINGER, K.E.. AND G.M. CHRlSTENSEN. 1972. Effects of various metals on
    survival, growth, reproduction, and metabolism ofDaphnia magna.J. fish.
    Res. Board Can. 29:1691-1700.
' BORCMANN.U., W.P.NORWOOD, AND I.M. BABIRAD. 1991. Relationship between
    chronic toxicity and bioaccumulation of cadmium in HyaMla azleca. Can.
    J. Fish. Aquai. Sci. 48:1055-1060.
 CABANISS, S.E., AND M.S. SHUMAN. I988a. Copper binding by dissolved organic
    matter: 1. Suwannee River fulvic acid equilibria. Geochim. Cosmochim.
    Acto 52: 185-193.

                              Can. J. Fish. Aquat. Sci.. Vol. 50.1993

-------
 CABANES. S.E.. AND M.S. SKUMAN. 1988b. Copper binding by dissolved organic
     matter Q. Variation in type  and source of organic  matter. Geochim.
     Cosmcchiro. Acta 52:195-200.
 CAMPBELL. P.G.C., AND P.M. STOKES. 1985. Acidification and toxicity of metals
     to aquatic biota. Can. J. Fish. Aquat Sci. 42:2034-2049.
 DALY, HJU'MJ. JONES, B.T. HART, AND I.C. CAMPBELL. 1990. Copper toxichy
     loPanaya australieruir. HL Influence of dissolved organic matter. En won.
     Toxicol. Chem. 9:1013-1018.
 DEWIT, J.CM., MM. NEDERLOF, WJi VAN REMSDUK, ANDL.K. KOOPAL. 1991.
     Determination of H* and metal ion affinity  distributions for humic
     substances. Water Aii Soil PolluL 57-58:339-349.
 FISH, W., AND F.M.M. MOREL.  1983. Characterization of organic copper-
     complexing agenis released by Daphnia magna. Can. J. Fish. Aquat Sci.
     40:1270-1277.                       .
 FLORENCE, T.M. 1977. Trace metal species in fresh waters. Water Research 11:
     681-687.
 GRIMM, D.M., L.V. AZARRAGA, I_A. CARRHRA, AND W. SUSETYO. 1991. Con-
     tinuous multUigand distribution model used to predict the stability constant
     of Cu(Q) metal  complexatioh with humic material from fluorescence
     quenching data. Environ. Sci. Technol. 25:1427-1431.
 HERING. J.G.. AND FMM. MOREL. 1990. Kinetics of trace metal complexation:
     ligaad-exchange reactions. Environ. Sci. Technol. 24:242-252.
 HOLM. TJt, AND CD. CURTISS m. 1990. Copper complexation by natural
     organic matter in ground water, p. 508-518. In D.C. Melchior and &L.
     Bassen led.] Chemical modeling of aqueous systems 1L American Chemical
     Society, Washington, DC
 KERNDORFF, H.. AND M. SCHNTTZER. 1980. Sorption of metals on humic acid.
     Geochim. Cosmochim. Ada 44:1701-1708.
 LAUREN.DJ, AND D.G. MCDONALD. 1986. Influence of water hardness, pH, and
     alkalinity on die mechanisms of copper toxicity in juvenile rainbow trout,
     SaSino gairdneri. Can. I. Fish. Aquat Sci. 43:1488-1496.
 LAUREN, D J., AND D.G. MCDONALD. 1987. Acclimation to copper by rainbow
     tro«,Saimo^<2irtin*ri:biocbernistry.Can. J.Fish.AquaLSci.44:105-111.
 LIN, C.-F...AND MM. BENJAMIN. 1992. The effects of strongly complexing
     ligands on the adsorptive partitioning of metal ions. Water Res. 26:
     397-407.
 LIN. H.. AND DJ. RANDALL. 1990. The effect of varying water pH on the
     acidification of expired water in rainbow trout J. Exp. Biol. 149:149-160.
 LUND. W., LA. HELBAK, AND ttM. SHIP. 1990. Studies of the complexation
     properties of aquatic mtmic material by. differential pulse polarography. Sci.
     Total Environ. 92:269-281.
 MEADOR, LP. 1991. The interaction of pH, dissolved organic carbon, and total
     copper in the  determination of ionic copper and toxicity. Aquat Toxicol.
     19:13-32.
 MOREL, F.MJM- 1983. Principles of aquatic chemistry. John Wiley and Sons,
     Toronto, ON. 446 p.
NOR, Y.M., AND H.H. DENG. 1986. Chemical speciation and broavailability of
     copper uptake and accumulation by Eichornia. Environ. ToxieoL Chem.
     5:941-947.                        ''         <         •
 OLIVER, E.G., EM. THURMAN, AND RX. MALCOLM. 19831 The contribution of
     lumic substances to the acidity  of colored natural waters. Geochim.
     Cosmochim. Acta 47:2031-2035.
PAGENKDPF, O.K.  1983. Gill surface interaction model for trace-metal toxicity
     to fishes: role of complexation, pH, and water hardness. Environ. Sci.
     Technol. 17:342-347.
 PART. P.. AND RA.C. LOCK. 1983. Diffusion of calcium, cadmium and mercury
     in a mucous solution from rainbow trout Comp. Biochem. Physiol. 76C:
     259-263.  -
PERDUE, EM., AND C.R. LYTLE. 1983. A critical examination of metal-ligand
     complexation models:  application  to defined multiligand mixtures,
     p. 295-313. In R.F, Christman and Egil T. Gjessing [ed.] Aquatic and
     terrestrial humic materials. Ann Arbor Science, Ann Arbor. ML
 PLAVLE.R.C.. ANDC.M. WOOD. 1989. Waterchemistry changes in the gill micro-
     environment of rainbow trout:  experimental observations and theory. J.
     Comp. Physiol. B 159:527-537.
 PLAYLE, R.C., R.W. GENSEMER, AND D.G. DDCON. 1992. Copper accumulation
     on gills of fathead minnows: influence of water hardness, complexation add
     pH of the gill micro-environment Environ. Toxicol. Chem. 11:381-39ll!
 PLAYLE, R.C, D.G. DKON, AND K. BURNISON. 1993. Copper and cadmium
     bindiDgtofishgilkimodificationbydissolvedorganiccarbonandsynthetic
     ligands. Can. J. Fish. Aquat Sci. 50:2667-2677.                   II
 REID. S J).. AND D.G. MCDONALD. 19.88. Effects of cadmium, copper, and lo'v
     pH on ion fluxes in the rainbow trout. Salmo gairdneri. Can. J. Fish. Aqua'L
     ScL 45:244-253.                             .               II
 REID, SJ>., AND D.G. MCDONALD. 1991. Metal binding activity of the gills <«f
     rainbow trout  (Oncorhyndua mytiu). Can. J. Fish. Aquat  Sci. 4{!:
     1061-1068.                                      •            II
 SAKU. S, AND D.K. BANERJEE. 1990. Complexation properties of typical soil and
     peat humic acids with copperdl) and cadraium(II). Int J. Environ. Anal
     Chem. 42:35-44:               •         .                     -|
 SCHECHER.W. 1991.MIr^L+versic«2.1.EnvmjiimentalReseafchSoftwan
 ———-—n "• •••*•••'*••••***(*,  v«.tMwif A,i,bu*nuiuu«iuBiivc5caicnoonwBrc
 SRUMAN. M.S. 1992. Dissociation pathways and species distribution of.alumi-
     num b«ind loan aquadcM vie acid. Environ. Sci. Technol. 26:593-598!
 Sposrro, G. 1981. Trace metals hi contaminated waters. Environ. ScLTecbnoJl
     15:396-403.               '       ,  ;                          |
 SUNDA. W.G.. AND PJ. HANSON. 1979. Chemical speciation of copper in rive
     water effect of total copper. pH. carbonate, and dissol ved organic maud
     •. |J|*T f OA  f_ T? * *^	». » * MI    >  *    i *< •  '.         .     'It
     p. 147-180. In EA. Jenne [ed] Chemical modeling in aqueous systemii
     ACS Symposium Series 93. American Chemical Society, Washington, DC!
TAGA, M.. S. TANAKA, AND M. FUKUSHMA. 1991. Evaluation of copner(LTJ|
     binding ability of humic acids in peat using sulphopropyl-sephadex C-2ii
     catiop exchanger. Anal. Chim. Acta 244:281-287.                   II
TUSCHALL.JJL JR..ANDPJ. BREZDNK. 198i Complexation of heavy metals b|
     aquatic humus: a comparative study of five analytical methods, p. 275-294!
     /n RJ.  Christman and E.T. Gjessing [ed.] Aquatic and terrestrial hnmii:
     materials. Ann Arbor Science, Ann Arbor, ML                       jl
URBAN, N JL, E. GORHAM, IX. UNDERWOOD. f£. MARTIN. AND J.G. OGDEN m!
     1990. Gepchemical processes controlling concentrations of Al Fe, and Mil
     in Nova Scotia lakes. LimnoL Oceanogr. 35:1516-1534.              II
VAN DEN BERG, CM.G..ANDJJL KRAMER. 1979. Conditional stability constant!
     for copper ions with ligands in natural waters, p. 115-132, to EA. Jenws1
     [ed] Chemical modeling in aqueous systems. American Chemical Society!
    'Washington, DC.                                               j
VAN DEN BERG. GM.G., P.T.S. WONG. AND Y.K. CHAU. 1979. Measurement ol;
     complexing materials excreted from algae and their ability to ameliorate'
     copper toxicity. J. Fish. Res. Board Can. 36:901-905.                I
VAN LOON. LJt. S. GRANAGHER, AND H. HARDUF. 1992. Equilibrium dialysis]
     ligand exchange: a novel  method for determining conditional stability
     constants of radionuclide-humic acid complexes. Anal. Chun. Acta 268:
  •  -235-246.   •-.                                    -             |
VERBOST. PAt. J.  VAN ROOD. G. FLK. R-A.C. LOCK. AND S.E. WENDELAAR
     BONGA. 1989. The movement of cadmium through freshwater troui
     branchial epithelium and its interference with calcium transport. J. Exp!
     Biol. 145:185-197.                         ...
WKXLUND, A.,ANDP.RUNN. 1988. Calcium effects on cadmium uptake, redistri-
     bution, and elimination  in minnows, Phoxinus phoiinus, acclimated to
     different calcium concentrations. Aquat Toxicol. 13:109-122.         ]
WILKINSON, KJ., P.G.C. CAMPBELL, AND P. COUTURE. 1990. Effect of fluoride
     complexation on aluminum  toxicity towards juvenile Atlantic salmod
     (Salmo solar)- Can. J. Fish. Aquat Sci. 47:1446-1452.               I
WINNER, R.W. 1985. Bioaccumulation and toxicity of copperasaffectedby inter]
     actions between humic acid and water hardness. Water Res. 19:449-455)
WINNER, R.W. 1986. Interactive effects of water hardness and humic acid on the
     chronic toxicity of cadmium toDapkniapulex. Aqnat Toxicol. 8:281-293!
XUE, R-B^ AND L. SKW. 1990. Binding of Cu(II) to algae in a metal buffer!
     Water Res. 24:1129-1136.                                      |
VAN, ND., GI_ MACKE, AND PJ. DILLON. 1990. Cadmium concentrations o]
     crustaceanzooplanlctonofacidifiedandnonacidifiedCanadiaashieldlakes!
     Environ. Sci. Technol. 24; 1367-1372.                             j
Zmco. V.. AND W.G. CARSON. 1976. A mechanism of the effects of wate;
     hardness on the lethality of heavy metals to fish. Chemosphere 5:299-303
 Can. J. Fish. Aquat. Sci., *W. 50.1993
                                                                                                                                      268

-------
                                      f
        Pergamon
                Computers A Ctaselencei Vol. 20, No. 6. pp. 973-1021.1994
                             Copyright O 1994 Ebevier Science Ltd
0098-30M{94)EOOI6-M        Primed io Grot Briuin. AH righuiwmd
                                      0091-3004/94 $6700 +0.00
           WHAM—A CHEMICAL  EQUILIBRIUM  MODEL AND
              COMPUTER CODE FOR WATERS,  SEDIMENTS,
                 AND SOILS INCORPORATING A DISCRETE
             SITE/ELECTROSTATIC MODEL OF ION-BINDING
                             BY HUMIC SUBSTANCES
                 /
                                         E. TIPPING
           Institute of Freshwater Ecology, The Ferry House, Ambleside. Cumbria LA22 OLP, UJL

                              (Received and aeeefud I December 1993)

       Abstract—WHAM (Windermere Humic Aqueous Model) b designed to calculate equilibrium chemical
       speciation in lurface and ground waters, sediment*, and sous. The model is suitable especially for problems
       where the chemical specUtion is dominated by organic nutter (humic substances). WHAM combines
       Humic Ion-Binding Model V with a simpk inorganic specUtion code for aqueous solutions. Precipitation
       of aluminum and iron oxides, cation-exchange on in idealized day mineral, and adsorption-desorpu'oa
       reactions of fulvic acid also are taken into account. The importance of ion accumulation in the diffuse
       layers surrounding the mimic molecules is emphasized. Model calculations are performed with a BASIC
       computer code running on a Personal Computer.

       Key Wards'. Chemical equilibrium. Chemical speeiation, Humic substances. Ion-binding, Sediments, Soils,
       Waters.                .
               INTRODUCTION

Humic substances are recognized to play an import*
ant part in the chemical speciation of waters, sedi-
ments, and soils, but describing their interactions
with protons, metal  ions, and other species  under
environmental conditions has proved difficult  in the
absence  of an appropriate speciation code.  Until
recently, the main attempt to do so has been with the
program GEOCHEM (Mattigod and Sposito,  1979),
reprcsentine^fssolved humic matter bv * «di«aif"i *f
lovLjnotecular  weight  organic compounds with
known Vjndmg propeoies^Although useful as a guide
to die interactions,  this approach  is not entirely
satisfactory, because it ignores the macroionic nature
of fulvic and .humic  acids. More recently, a "com-
petitive Gaussian" model for fulvic acid has been
introduced into -the MINTEQ code (Allison and
Perdue. 1994), .but the available  database for metal
interactions is'limited (Susctyo and others, 1991). No
codes seem to be available to deal comprehensively
with solid-phase organic matter.
  The  Windermere  Humic   Aqueous   Model
(WHAM), presented here, can be applied to equi-
librium speciation  problems involving waters, sedi-
ments, and soils, with humic substances present in
dissolved or paniculate .form. It follows from the
development and partial validation-of Humic Ion-
Binding Model V (Tipping and Hurley, 1992; Tipping
I993a, I993b, 1993c). Although  Model V could be
          . combined with existing inorganic speciation codes in
           order to solve chemical equilibrium problems, a new
           organic-inorganic  combination that  takes  into
           account the solving demands of Model V in the
           overall scheme was developed, father than attempting
           to incorporate Model V into an existing code. This is
           important especially for concentrated systems (sedi-
           ments and soils) where the accumulation of ions in
           (he diffuse layer surrounding the humic molecules
           takes on major importance. Moreover, a purpose-
           built model should be better-suited for incorporation
           into dynamic models of natural aquatic, soil, and
           catchment systems.

                       MODEL DESCRIPTION

             WHAM is a combination of several submodels.
           These are Humic Ion-Binding Model V (Tipping and
           Hurley, 1992; Tipping, I993a, 1993b), and models of
           inorganic solution chemistry, precipitation of alumi-
           num and  iron oxyhydroxides, cation-exchange on a
           representative day, and the adsorption-desorption
           reactions  of fulvic acids (Tipping and Woof, 1990,
           1991). The interrelationships among these submodels
           are depicted in Figure 1.

           Model V: specific binding by humic substances
          .  Humic compounds are represented by hypothetical
           size-homogeneous, rigid, molecules, which carry pro-
           ton-dissociating groups  that can bind metal ions
                                             973

-------
974
                                           E. TIMING
either singly or as bidentate pairs. The interactions
are described in terms of intrinsic equilibrium con-
stants and electrostatic terms, The former refer to the
(usually hypothetical) situation where the humic sub-
stances have zero electrical charge. The latter lake
into account the influence on binding of the variable
humic charge; binding strength is enhanced when the
metal species and the humic molecule carry opposite
charges, and diminished when the charges have the
same sign. The electrostatic terms take the general
form e'2**2 where w is the electrostatic interaction
                                                 factor (dependent upon ionic strength), z is the charge
                                                 on the combining ion in question, and Z is the humic
                                                 charge. For example, the dissociation of a proton by
                                                 the following reaction

                                                              RAH *>RA -' + H*         t  (I)

                                                 is characterized by the variable equilibrium quotient,
                                                                    RAH'
                                                                                   t*Z
                                                                                            (2)
                           DISCRETE SITES
                           binding ond
                           dissociolion
                                                     DIFFUSE UYER
                                                          botonclng
                                                          counlerions
                                     -2wzZ
                                                              Donnan
AI(PH)5


Fe(OH)3
                   K
                     so
                                     BULK  SOLUTION
                                 A  +  B
                                                                     Donnan
  CLAY
counterions
balancing

 ZCUY
                                  \
                                  L
                                                              Donnan
                           binding and
                           dissociation
                                                          balancing
                                                          counlerions
                           DISCRETE SITES  FA  DIFFUSE UYER
                                                                                             * v
                                                                                                 \i
                                       SOIL/SEDIMENT
                                            SOLIDS
       Figure 1. Functional relationships in WHAM. FA»fulvic acid, HA ^ humic acid. Species in bulk
       solution ait ia equilibrium with futvic and humic discrete sites, and also with counterion specks in diffuse
       layers. In version of model for waters (WHAM-W), aluminum and iron (oxy)hydroxides and day cation
       exchanger are omitted. In version for soils and sediments (VYHAM-S), some of fulvic acid can be present
       in both solid and dissolved forms. Humic and fulvic acids, and clay cation-exchanger, together with their
                         diffuse layers, have net zero charge, as does bulk solution.

-------
                                Equilibrium chemical spcciaiion by WHAM
                                             915
                     Table I. Default parameter values for fulvic acid (FA) and humic acid
                     (HA) used in  WHAM-W Version  1.0 and WHAM-S  Version 1.0.
                                  DL P diffuse layer. See text for details.

"A

p|CA
mV
"P*^A
ApK
•ApK
f
/„


pKn

j
y
f
K»
'
mol type A groups per g
mol type B groups per g
central pK for type A groups
:. central pK for type groups
range of type A pKV5-' A 7
,. range of type BpK's
'' defines w by w «• flogw/
proximity factor
molecular weight
molecular radius (nm)
iiii in terms of P^MHA
charge factor for DL volume
factor for max DL volume
distribution parameter
adsorption parameter '
adsorption parameter
FA
4.73 x IO->
2J7xlO->
3.26
9.64
3.34
5.52
-103
0.4
1500
0.8
3.96pKtlM
10*
0.25
I
10*
1
HA
3.29 x 10-»
1.65 x 10-'
4.02
8J5
1.78
3.43
-374
OJ
15.000
1.7Z
3pKiniA -~ 3
10*
0^5
—
— ,
—
 where Kta, is the intrinsic constant: Because w is
 positive and Z nearly always negative, Equation (2)
 shows that the greater (more negative) is the humic
 charge, the less does reaction (I) proceed, because of
 the tendency of the humic charge to hinder protein
 dissociation.
   The  proton-binding  groups of the humic sub-
 stances are heterogeneous, having a range of intrinsic •
 pK  values.. Two  types  of.  acid  group  are
 distinguished, denoted by A and B. Within each type
 there are  four different groups, present in equal
 amounts, the pK values of which are described in
 terms of a median value, pKA or pK,, and a factor,
 ApKA or ApKj, that defines the range of the values.
 For example, the four type A groups have pK values
-given by  •
                                              0)
 where I = 1,2,3,  or 4. The humic content of type
 B sites (nB) is fixed at one-half the content of the
 type A (nA) sites. This rather strict arrangement of
 proton-binding groups is  selected with  a view to
 simplifying the description of the binding of other
 ions.    '           •
   Competitive metal binding takes  place at single
 proton-binding sites (monodcntate), and at bidentale
 sites formed by pain of proton-dissociating sites. The
 extent to which, .bidentate binding can lake place is
 constrained by  the proximity factor (/^), which
 defines,  on the  basis of molecular geometry, the
 likelihood of pairs. of proton-binding. groups being
 close enough to form bidentatc sites: Thus a proxim-
 ity factor of 0.4 means that 40%  of the proton-
 binding  groups  are sufficiently  close to pair up.
 .Because there are' 8  proton-dissociating sites,  36
 different bidentate sites can .be formed theoretically,
 but in order to speed computations, only 12 represen-
 tative pairings— present in equal abundances — are
 allowed in the model. Monodentate metal binding is
described with intrinsic equilibrium constants for the
metal-proton exchange reaction:
        /MHz+Ml*»^Mtz**-|»+H*.     (4)
The negative logarithms of these intrinsic constants
are denoted by pKMHA and pK-tutu tot type A and
type B sites respectively. The smaller is the value of
pKMH> the stronger is the binding of the metal species
in question. The appropriate pK values are added
together to obtain the value for bidentate binding.
The model permits the binding of the first hydrolysis
product (e.g. CuOH4 in the situation of Cu1*) as well
as the parent species, the pKtm values for the first
hydrolysis product being assumed equal to those of
the parent This assumption has not been tested fully,
but has been shown to account adequately for copper
binding results (Tipping and Hurley, 1992; Tipping,
1993a).  It has been  discovered that plC^a  »
sufficiently well-correlated with pKiou that only one
of them needs to be specified to  describe the binding
of a given species. The selected parameter is pKUHA,
because-this has been determined to exert the most
influence when fitting published data. sets. Values of
PKMHA have  been derived from literature data for
aluminum, alkaline earths, heavy metals, europium,
and actinidcs  (lipping and Hurley. 1992; Tipping
1993* 1993b, 1993c), and also have been estimated
from   linear   free  energy  relationships  (Tipping,
1993c). In all, parameter values for 36 parent species
are currently available.
  To implement Model V, it is necessary to know the
following parameters, which characterize the proton-
binding  characteristics of the humic substance in
question; itA, n,, pKA, pKB, ApKA, ApKB, P (empiri-
cal parameter defining w, by w = P tog,,,/, where / is
ionic strength), and/p,. Sets of parameter values have
been published for different humic materials (Tipping
and Hurley, 1992; Tipping, 1993a, 1993b) and sets of
"best-average" default values have been arrived at for
fulvic and humic acids (Table 1). Generally, different
 CAGEOM/6—F

-------
976
                                            E. TIPPING
    samples of the same type of humic material (i-C- fulvic
    or humic  acid) give similar sets of these Model V
    parameters (Tipping and Hurley, 1992; Tipping,
    1993b). In addition, a pKMtu value is required for
    each species that can  bind at the discrete  sites, as
    explained in the previous  paragraph; default pKHHA
    values  are included  in the  SSED Database (see
    Appendix  I).

    Model  V: nonspecific binding by counterion accumu-
    lation
      The modification  of binding strength  at  specific
    sites by electrostatic effects (discussed  previously) is
    the result of the accumulation of an excess of counter*
    ions in a  diffuse layer adjacent to the molecular
    surface, and  this nonspecific  process can  be con-
    sidered to contribute to the total binding. For waters,
 .   such accumulation is important in using Model V to
    account for observed  competition effects involving
    species, such as alkaline earths,  that show weak
    specific binding (Tipping,  I993a). For  more concen-
    trated systems such as sediments  and soils, where
    humic concentrations can  be high (tens of grams per
    liter), counterion accumulation thus may account for
    the greatest part of some species. Conventional elec-
    trostatic theory  (Gouy-Chapman, Debye-Huckel)
    for charged interfaces in aqueous  solutions  regards
    the diffuse layer as being, in principle,  infinite in
    extent,  and ionic distributions are described using
 .   Boltzmann statistics (Tanford, 1961; Hiemenz, 1977).
    This description is the basis for the modifying terms
    mentioned in the previous section, of the form e ~*az.
    To compute  counterion accumulation (and co-ion
    deficit) with this theory, it is necessary to integrate for
    the entire volume of  the diffuse layer, or, more
    practically, for some truncated version thereof (Bolt,
    1982). By this approach, the modification of specific
    binding and counterion accumulation  would  be
    described with the same model. However, the inte-
    grations required are demanding computationally,
    and the theory does not take readily into account
    interference among the diffuse layers of neighboring
    panicles, which must be significant in concentrated
    systems such as sediments and soils. In the situation
    of humic substances, heterogeneity among molecules
    adds an extra problem in applying the fundamental
    theory. For these reasons, it is considered justified to
    adopt a much simplified description  of counterion
    accumulation in Model V and WHAM:
   •'   In Model V, the diffuse layer is regarded as a zone
    of defined thickness around the humic molecules and
r*  average concentrations of counterfoils within that
    zone are  considered.  Co-ions are excluded  com-
    pletely. There is no mathematical connection with the
    term  modifying interaction at discrete sites. An
    approximate thickness of the diffuse layer is taken to
    be the ionic-strcnglh-dependent Debye-Huckel par-
    ameter, K (Hiemenz, 1977).  With  the thickness
    defined, the volumes of the diffuse layers are calcu-
    lated from geometrical formulae, assuming the humic.
                                                  molecules to be spheres. It is necessary to assume
                                                  reasonable values for the humic radius and molecular
                                                  weight. The volume is given by
                                                  where ATA, is Avogadro's number, M is humic mol-
                                                  ecular weight, and r is the radius  of the  humic
                                                  molecule. With the diffuse layer volume so-defined,
                                                  ionic distributions are calculated easily, using Don-
                                                  nan expressions [see Eqs. (8) and (9)] and requiring
                                                  that the total counterion charge within the  diffuse
                                                  layer exactly balances the charge resulting from the
                                                  humic ionizable groups, modified by specific binding
                                                  of protons and metal ions. Also, the bulk solution is
                                                  charge-balanced perfectly (Fig. 1).
                                                    For dilute systems,  especially  at  higher ionic
                                                  strengths,  the diffuse layer volumes  are small  in
                                                  comparison with the total. However, this is not the
                                                  situation  for more concentrated systems at lower
                                                  ionic strengths. For example, the diffuse layer volume
                                                  of a solution of Smg/1 fulvic acid  with an ionic
                                                  strength of 0.1 M is calculated to ». • 4 x 10~s liters
                                                  per liter of total water. In contrast, iu an organic soil
                                                  with a water content of 75%, the humic acid concen-
                                                  tration might be 50 g per liter of total water, and die
                                                  ionic strength would be tow, say 0.001 M. This would
                                                  correspond to a diffuse layer,volume of ca 10 liters
                                                  per liter, an obvious impossibility. Model V attempts'.
                                                  to handle this problem with an empirical factor,^, f (
                                                  that restricts the diffuse layer volume. Diffuse layer )
                                                  volumes calculated with Equation (5) are designated
                                                  maximum volumes, Pbnurand Kn~.H for fulvic and
                                                  humic acid respectively, and the actual volumes Vw
                                                  and ym are calculated  from
                                                                      /DI/PI
                                                                            »-*
                                                             011  /Dc-f-
                                                  lne value of/PL is set to a number between 0 and 1.
                                                  that is the asymptotic maximum total diffuse layer
                                                  volume. On the basis of preliminary results with acid
                                                  soils, a default value of 0.25 is used for/w., but this
                                                  is not an optimized value. To suppress restrictions on
                                                  diffuse byer volume,/n. can be set to infinity.
                                                    A further difficulty arising with the Model V diffuse
                                                  layer is the small requirement for. counterfoils at low
                                                  net humic charge* which can lead to the paradoxial
                                                  situation in which counterion concentrations are cal-
                                                  culated to be lower than bulk solution values. This
                                                  has been noted for aluminum-rich soils, where the
                                                  specific binding of Alu can make thejiumic charge
                                                  go to near zero. This seems to follow from a violation
                                                  of the implicit assumption that the humic charge is
                                                  spread evenly over  the molecular surface. To take
                                                  into account  such "discretization" of humic charge
                                                  an adjusting  factor, Kz, is introduced  to force the
                                                  diffuse layer volume to diminish as charge decreases.

-------
                                Equilibrium chemical speciation by WHAM
                                            977
 Thus  ihe  maximum  diffuse  layer  volumes  are
 modified as follows;
                      'Dm..
                                             (7)
 where ZM,, is the modulus  of Z. Experience in
' modeling acid soils has suggested a value of 10' for
 Kz, but as in  the  situation  of/K., this is  not
 considered an optimized value.
   Although overlapping of diffuse layers is con-
 sidered in WHAM, this is restricted to the diffuse
 layers of like  molecules, that is fulvic acid diffuse
 layers overlap only with each other, and those of
 bumic acid behave likewise. For concentrated  sys-
 tems, clearly this  is artificial, but in  view of the
 simplifications  and  approximations  involved  in
 attempting to describe the diffuse layers, it does not
 seem justified to elaborate the model further. On a
 more  practical  level,  keeping  the diffuse  layers
 separate simplifies the  description  of fulvic acid
 adsorption-desorption reactions.
   In published work to date on Model V, and its
 predecessor Model IV, counterion accumulation has
 been, assumed  to depend only upon counterion
 charge. However, work in progress with soils suggests
 that in order to explain major cation  distributions
 (Na*. Mg'+,  Ca*+,  etc.) it may be necessary to
 introduce some selectivity. Although this has yet to
 be  established firmly, the possibility is allowed in
 attempting to present as comprehensive as possible a
 picture of the WHAM models. In  the nonselective
 Donnaii model,  diffuse  layer concentrations  of
 counterion  species  i  are  related  to   solution
 concentrations as follows:
                                             (8)
 where z^l) is the modulus of the charge on species
 i. and R is the ratio required for the sum of the:
 counterion charges to balance the humic charge Z.
 Selectivity is introduced by writing;

               <*(')
 where K*,(l) is the selectivity coefficient. Only a few
 values have been assigned, again.from fitting- with
 organic acid soils, and the values selected are not
 different greatly .from unity.
    The description of the diffuse layer within  Model
 V is a crude attempt to obtain a practical model,
 and  consequently there are several  uncertainties
 associated with it, as follows:

    (a) The relative amounts of accumulation of differ*
       ent counterions depend upon the diffuse layer
       volume; the smaller the volume, the more are
       tons of higher charge favored. Therefore uncer-
       tainty in this volume translates to uncertainty
       in nonspecific binding.
  (b) Although reasonable results can be obtained
      with the model for suspensions of soil solids,
      experimentally it is difficult to make measure-
      ments at the  high  effective  concentrations
      encountered in real soils. Extrapolation to such
      conditions relies on  the maintenance of the
      assumed overlap of diffuse layers, as described
      by Equation (6).
  (c) The overlapping of diffuse layers in concen-
      trated systems must alter the potential field at
 t     the humic surface, and thereby change the ionic
      distribution that is assumed implicitly in apply-
      ing the modifying  term e ~J"Z to account for
      the electrostatic influence on binding at specific
      sites (see previous description). This is ignored
      in the model.

It is clear that each of these uncertainties gets more .
serious as concentration is increased, and predictions  **•
about ionic distributions in real soils and sediments
are only first approximations. Nonetheless, correct
trends are likely to be captured.
  Parameters required for diffuse layer calculations
are humic molecular weights and radii,/ott JSTZ, each
of which is approximated in the present'versions of
WHAM (Table 1). If selectivity among counterions is
considered, as in  the soil/sediment version of the
model then  values of K^(i) for each counterion
species are required also  (see Appendix 1).

Fixed charge cation exchanger (day)
  This is included to allow some account to be taken
of the presence of soil and sediment days. The same
approach as for the bumic diffuse layer is taken,
except that the day is assumed to have a flat surface
for the purposes of calculating diffuse layer volume,
and selectivity is  possible. The total diffuse layer
volume is constrained along with those of the FA and
HA, by extending Equation (6). Default values for
the cation exchange capacity and surface area of the
clay are lO^equivg'1 and lOOm'g*1 respectively,
these  being  representative  of  values  given  by
Talibudeen (1981). In the  present version of the
soil/sediment database (see  Appendix. I), all  the
selectivity coefficients for the clay are set to! (cations
and neutral species) or 0 (anions).

Precipitation ojaluminum andiron(III)oxyhydroxtdcs
  WHAM allows these  two solid phases to form.
Straightforward solubility products are used, and are
compared with the appropriate ton activity products,
to decide whether precipitates are present. The solid .
phases are not considered to have active surfaces, that
is they cannot bind  ions, nor is there 'any explicit
reaction with humic substances. In using the model,
 these points must be borne in mind—thus WHAM in'
 this version is best for situations in which the solid itr-
 phases can be considered to be dominated by organic
 matter. A solubility product and enthalpy are needed
 for each reaction. There are ranges of values possible

-------
978-
                                                 E. TIPPING
      for these, because different forms may be present,
      with  different solubilities. Values typical of amor-
      phous precipitates are included in the database (see
      Appendix 1), but these can be altered for  specific
      problems if necessary.
        The inclusion  of aluminum and  iron(III)  oxy-
      hydroxides in the model reflects the environmental
      systems, acid waters, and soils, for which the model
      first was designed. Future applications may  require
      account to be taken of the formation and dissolution
      of other inorganic phases such as carbonates and
      sulfides.  Descriptions of these reactions could be
      incorporated readily into the code.

   ,   Sorptlon offulete add by soil and sediment solids
        Tipping and Woof (1990, 1991) showed that net
      electrical charge is important in the interactions of
      fulvic acid with soil solids, and employed a modeling
      approach that included a simple description of the
    ,  hydrophobicity of the fulvic molecules. This allows
 I  '  the distribution of fulvic acid between the aqueous
7- • and solid phases  to be estimated. The fulvic  acid
    ' molecules are considered to consist of a series of ten
     subtractions, each having  the same  ion-binding
     properties, but differing in hydrophobicity. Thus,
     desorption depends on whether the net fulvic charge
     is  sufficient  to overcome the hydrophobicity.  The
     model has three parameters. The abundance distri-
    ' bution of the ten fulvic fractions is described with j,
     whereas the adsorption of each fraction is character-
f  feed by ft and A».  Default  parameter values are
     included  in the SSED database (see Appendix 1);
     these have been determined to be typical for organic
     matter release from acid organic soils, but their use
     for higher pH systems  or sediments has not been
     investigated. In running the model, it is possible to
     override  the adsorption-desorpu'on  model,  and
     simply specify the concentration of dissolved fulvic'
     add.

     Chemical speclalion In the solution phase
       This concerns the humic-free solution phase, in
     which only  inorganic  species,  and  low molecular
     weight organic: if required, are recognized. Complcx-
     ation reactions are formulated in terms of up to three
    ^ master species, as in the PHREEQE code developed
     by Parkhurst and others (1980). Activity coefficients
 ,   are calculated with the extended  Debyc-Huckel
     equation. Equilibrium constants and  enthalpies for
     the reactions considered are listed in the 'databases
     for WHAM (see  Appendix I). These  values were
     obtained primarily from the compilations of Baes and
    . Mesmer (197(6). Smith  and Martell (1977),  Nord-
     strom and others (1990), Read and others (1991) and
     Mattigod and Sposito (1979). Some values also were
     taken  from Buffle (1988). Maes, De Brabandere, and '
     Cremcrs  (1988),  Sunda 'and  Hanson  (1979)  and
     Tipping, Woof, and Hurley (1991). Users can assem:
      ble their own databases if required. When only a few
                                                 reactions are of interest, it is advantageous to use a
                                                 small database in order to reduce execution times.

                                                       SOLVING THE WHAM EQUATIONS

                                                  There are two versions of WHAM; WHAM-W for
                                                 waters, and WHAM-S for soils and sediments. The
                                                 solving algorithms differ slightly, mainly because of
                                                 the different convergence characteristics required for
                                                 dilute and concentrated systems. The essence of the
                                                 problem is to distribute known total concentrations
                                                 of master species among the different chemical forms
                                                 considered. The  master  species are mostly parent
                                                 cationic and anionic species (H*, Na*. AP*. d',
                                                 POJ~, etc.). The criteria for solution are  mass
                                                 balance, charge balance  (except for WHAM-W in
                                                situations where pH is fixed), and correct net charge
                                                 on fulvic and mimic acids. In outline, the procedure
                                                 is as follows:                      -

                                                   (1) Set initial trial values of tonic strength, master
                                                       species1 activities,  fulvic,  and humic net
                                                       charges.
                                                   (2) Calculate activities and concentrations of all
                                                       inorganic species.
                                                   (3) Calculate binding to fulvic and humic acid,
                                                       and clay cation-exchanger (WHAM-S only).
                                                       Calculate net fulvic and humic charge from
                                                       amounts specifically bound.
                                                   (4) Calculate new ionic strength.
                                                   (5) Perform mass  balances on  all components,
                                                       calculate the charge ratio in the bumic-free
                                                       solution phase (unless pH is fixed), and com-
                                                       pare trial and calculated values of fulvic and
                                                       humic charge. If all these meet the criteria for
                                                       convergence, and if ionic strength has reached
                                                       a constant value, proceed to step (9), (10), or
                                                       (11), depending on the status of the precipi-
                                                       tation reactions. If the criteria are not met,
                                                       proceed to step (6).
                                                   (6)  Improve activity  values by the continued-
                                                       fraction approximation, using ratios of calcu-
                                                       lated and known total concentrations. The
                                                       activity of H* is  improved on the basis of
                                                       charge imbalance.                  '
                                                   (7)  Improve ZFA and ZHA  by  interpolating
                                                       between the previous trials and the calculated
                                                       values from (3).
                                                   (8)  Return to step (2).
                                                   (9)  If  the  solubility  of  Fe(OH),  has   been
                                                       exceeded,  return'to step (2), and repeat the
                                                       procedure, except that the activity of Fe**
                                                       now is obtained from the activity of H* and
                                                       the solubility product. (WHAM-S only.)
                                                  (10) Repeat   (9)   for   A1(OH),  precipitation.
                                                       (WHAM-S only.)
                                               -t> (11) Calculate the distribution of FA between solid
                                                       and aqueous phases (WHAM-S only).

                                                Computer codes  in Turbo BASIC to perform the
                                                 described tasks are given in Appendices 2 and 3. Code

-------
                               Equilibrium chemical spccialion by WHAM
                                            979
 verification has been achieved by checking individual
 segments against hand calculations, by comparing
 WHAM  outputs  with  those  from  a  version, of
 PHREEQE (Parkburst, Thorstenson. and Plummer.
 1980) into which Model V  has been  incorporated
 (M. Crawford. British Geol. Survey, pen. comra.),
 and by comparing WHAM outputs for inorganic
 speciation problems with  outputs  from WATEQ2
 (Ball, Jenne, and Nordstrom. 1979). The robustness
 or the algorithms has been investigated by performing
 computations  on  60 problems representative of
 natural waters, sediments,  and  soils, and no signifi-
 cant difficulties have been encountered. The usual
 source of failure Jo execute (program crash, or no
 convergence) is an inappropriate selection of starting
 pH, and this is rectified easily by editing the input file.
         ILLUSTRATIVE CALCULATIONS

Soft freshwater containing trace metals (Appendix 4)
  This is a fixed pH calculation and so the final
charge balance is not unity. When the model is used
in this mode, care has to be taken to supply approxi-
mately correct values for concentrations of major
electrolyte ions, in order to ensure reasonable values
of ionic strength. Because the system is dilute with
respect to humic substances, the humic-free solution
phase accounts for most of the total volume (993%).
As is usually the situation, the humic molecules have
negative charges, and these are balanced in the diffuse
layers almost entirely by sodium, magnesium, and
calcium.  However, such accumulation accounts for
only a few percent of these major cations. The two
trace metals, Ni and Cu, show contrasting behaviors,
with less than 50% of the Ni being bound by humic
substances, but 99.8% of the copper.  Because these
metals have appreciably greater affinities for,the
humic substances than do Na,  Mg. or Ca, their
binding is mainly by specific coihpiexation; diffuse
layer accumulation is insignificant Nonetheless, the
tabulated values of v (moles of a species bound per
gram  of humic substances) show that, by virtue of
their much higher solution activities, the alkaline
earths occupy more  binding sites than the trace
metals.            .

Sediment containing trace metak (Appendix 5)
  The sediment is assumed  to  be 90% by weight
water, and to contain appreciable organic matter and
day. The volume of solution phase is significantly less
than the total volume, the combined diffuse layer
volumes accounting for  17% of the  total water
volume. Despite the greater mass of clay present, the
organic matter dominates the calculated chemical
speciation. Sodium is recognized mainly (63%) in the
solution  phase, but the alkaline earths are  bound
predominantly to the solids, because  of counterion
accumulation and binding at specific sites. The large
amount of Fe(HI)  in the system is calculated to be
 present mainly (73%) as an oxyhydroxide precipitate,
 with nearly all the rest specifically bound by humic
 acid. Nickel is mostly sediment-bound, or bound to
 dissolved fulvic acid, so that only 0.4% is present as
 free Ni1*. Copper has the same son of distribution,
 but only 2 x 10~*% is calculated to be present as
 the free divalent cation. The model output includes
 Kt values, expressed in terms  of the Model V
 "architecture", that is  the total amount of master
 species bound per gram of solids divided  by the
 total  concentration in the aqueous phase, which
 includes that bound to dissolved fulvic acid. The
 apparent K< values are what would be calculated by
 assuming  all the water in the system to be  bulk
 solution, that is they are values that might be
 calculated conventionally from experimental  results.

 Add soil containing radioelements (Appendix  6)
  The soil is assumed to be  75% by weight water.
 Because of the high concentrations of humic sub-
 stances and the low tonic strength, nearly 25% of the
 water is calculated to be present in the diffuse layers.
 In this situation, the concentration of dissolved fulvic
 acid has been calculated with the model, and deter-
 mined to  be 46mg/l,  a reasonable value  for an
 organic soil solution. Sodium  is predominantly in
 solution, but this is not true of the other monovalent
 cation,  Cs, because  of the  different selectivity
coefficients for counterion accumulation (see Appen-
dix 1). Aluminum, frequently the dominant inorganic
element in acid  soils, is bound specifically to the
 humic substances; only 0.02% is free in solution. The
 Kt values for the radioelements Co, Sr, and Cs are
 determined  mainly  by  nonspecific  counterion
accumulation; higher pH values would be needed for
 specific binding by humic substances to contribute
 significantly to  the  distributions of Co and Sr,
 whereas the model does not permit specific binding of
 Cs by humics. However, even in organic-rich soils
 there may be enough frayed-edge illite to bind a large
 proportion of trace amounts of Cs. WHAM might be
 able to handle such situations with a small amount of
 the day cation exchanger, having a high selectivity
 for Cs. The Kt value for amcricium reflects to a large
 extent the distribution of fulvic add between the solid
 and solution phase, because strong completing by
 both solid-phase and dissolved humics means that the
 concentration of free Am14 is only 0.01% of the total.

              EXECUTION TIMES

  A 386  machine operating  at .25MHz,  with a
 numerical coprocessor,  took 21, 264, and 80 sec
 respectively  to run the problems. Speeds can  be
 increased considerably if the database size is reduced,
 because this allows smaller arrays to be employed.
 However, to  achieve  maximum speeds,  such a
 reduction requires the renumbering of species, as it is
 the identifying number of the last species, not the
 number of species per se that determines array size.

-------
                                               E TIPPING
 As an example, when the sediment problem discussed
.previously was run with  the identifier for the last
 required species changed to 160, instead of the value
 of 400 used with the default database, the execution
 time was reduced by 70%.
   A planned application of WHAM is in dynamic
 ecosystem models, where it will be used in order to
 compute chemical specialion on, for example, daily
 timesteps. Under these circumstances, the model can
 be made  to run more quickly  by storing values of
 variables  from one ttmestep to use as starting trial
 values in the next, because step-to-step changes in the
 values usually will be small.              .


            AVAILABILITY OF WHAM

   Executable versions of the  programs,  guidance
 notes for users, and details of database  documen-
 tation are available. A charge will be made to cover
 materials, postage/and handling. The author should
 be contacted for further information.

 Acknowledgments—This  work was funded by the  UK
 Ministry of Agriculture, Fisheries, and Food. I am grateful
 lo Lorenzo Giusti  (University of Lancaster, U.K.)  for
 helpful comments during (he development of the programs.


                  REFERENCES

Allison.  1. A., and  Perdue.  E  Mn  1994,  Modeling
   metal-humic interactions with MINTEQA2: Proc. 6th
   Intern.  Meeting. Intern. Humic Substances Sec. (Ban,
   Italy), in press.
 Baes. C F, and Meaner, R. E, 1976, The hydrolysis of
   cations: John Wiley & Sons, New York. 489 p.
Ball, J. W., Jenne,  E. A., and Nordstrom. D. K., 1979.
   WATEQ2-* computerized chemical model for trace
   and major element speciation and mineral'equilibria of
   natural waters, 4i Jenne, E A.. edn Chemical modeling
 .  in aqueous systems: Am. Chem.Soc.Symp. Ser., v. 93.
   p. 8I5-435.  '    ,
 Bolt, G. H., 1982, The ion distribution in the diffuse double:
   layer, At Boll. O. H., ed.. Soil chemistry B. Physico-
   chemical models: Ehevier, Amsterdam, p. 1-26.
 Buffle. J., 1988, Complexation reactions in aquatic systems:
   an analytical'approach:  Eflis  Horwood,  Quchesler,
   692p.
 Hiemenz, P. C. 1977. Principles of colloid and surface
   chemistry; Dekker, New York, 516 p.
 Maes. A.,  De Brabandere, J., and Owners, A,  1988, A
   modified method for the measurement of the stability of
   europium humk acid complexes in alkaline conditions:
   Radiochim, Ada, v. 44/45, p. 51-57.
 MatUgod. S. V.. and Sposiio. C., 1979. Chemical modeling
   of trace metal equilibria in contaminated soil solutions
      using the computer program  GEOCHEM, in Jenne,
      E A., ed., Chemical modeling in aqueous systems: Am.
   ,   Chem. Soc. Symp., Washington. D. C, p. 83T-456.
   Nordstrom, D. K, Plummer, N. C, Langmuir, D., Busen-
      berg, E. May, H. M, Jones, B. F., and Parkhurst, D. I_
      1990, Revised chemical equilibrium data for major
      water-mineral interactions and  their limitations. In
      Melchior, D. C. and Bassett, R. L, eds.. Chemical
      modeling of aqueous systems II: Am. Chem. Soc. Symp,
      Washington. D. C, p. 398-413.
   Parkhurst. D. L, Thorstensen, D. C, and Plummer, N. L.,
      1980. PHREEQE—• computer program for geochemi-
      cal calculations: US. Geol. Survey Water Res. Invest,
      Rept. 80-96,193 p.
   Read, D. FabrioL R, Jamet. P. Tweed. C, and Sellia, P,
      1991. Application and validation of predictive computer
      programs  describing  the chemistry of radidnudides
      in the geosphere—CHEMVAL project- Comm.  Ear.
      Commun. RepL M0979.003, Brussels, 83 p.
   Smith. R. M., and Martell. A. E. Critical stability con-
      stants. Vol. 4: Inorganic complexes: Plenum Press, New
      York,257p.
   Sunda, W. G.. and Hanson, P. J., J979, Chcmkal speciation
      of copper in river water, In Jenne, E A., ed. Chemical
      modeling in aqueous systems: Am. Chem  Soc. Syrnp,
      Washington. D. C, p. 147-180.
   Susetyo, WM Carretra, L. An Azarnga, L V, and Grimm,
      D. M, 1991, Fluorescence techniques for metal-humic
      interactions: Fres.  ZeiL  AnaL Chem., v.339, no. 9.
      p. 624-635.
   Talibudeen, O., 1981, Cation exchange in soils, la Green'
      land.D. J, and Hayes, M. H. B, eds. The chemistry of
      soil  processes:  John  Wiley  & Sons,  Chichester.
      p. 115-177.
   Tanford. C, 1961. Physical chemistry of macromolecules:
      John Wiley & Sons, New York. 710 p.
   Tipping, E,  1993a, Modeling the competition between
      alkaline earth cations and trace metal species for binding
      by humic substances: Env. So'. Tech.. v. 27. no. 3,
      p. 520-529.
   Tipping. E, I993B. Modelling ion-binding by humfc acids:
      Coll. Surf., v. 73. p. 117-131.
   Tipping, E, I993c, Modelling the binding of europium and
      the actinides by humic substances: Radiochim. Acts,
      v. 62, p. 141-152.
   Tipping, E, and Hurley, M. A, 1992, A unifying model
      of cation binding by humic substances: Geochim. Cos-
:   .   mochim. Acta, v. 56, no. 10. p. 3627-3641.
   Tipping, E, and Woof, C, 1990. Humic substances in acid
      organic soils: modelling their release to the soil solution
      in terms of humic charge: Jour. Soil. ScL, v.4I, no. 4,
      p. 573-583.
-^Tipping, E, and Woof, C, 1991, The distribution of humk
      substances between the solid and aqueous phases of acid
      organic sous; a description based on humk heterogeneity-
      and  charge-dependent sorption equilibria: Jour.  Soil
      ScL. v.42. no. 3, p. 437-448.
 . Tipping. E. Woof. C, and Hurley. M. A, 1991. Humk
      substances in acid surface waters; modelling aluminum
      binding, contribution to ionic charge-balance, and con-
      trol  of pH: Water Res, v. 25, no. 4. p. 425-435.
                                            APPENDIX I

            ,                                    WHAM
  Database SSED, version 1.0 (complete), and database WATER, version 1.0 (first lines). The first 10 lines in SSED refer
to:
(I) database name; (2) default parameters for humic acid (see text and Table I); (3) default parameters for fulvic acid (see
text and Table 1); (4) clay cation exchange capacity (equiv g-1) and specific surface area (rh1 g'1); (5) factor/Dt in Equation
(6) thai limits the volume of the diffuse layer; (6) thcrmodynamic constants for AI(OH)j; (7) thermodynamic constants for
Fe(OH),; (8) factor K. in Equation (7) to diminish the diffuse layer at low net humic charge; (9) constants '?. /? and K, used
in the fulvic acid desorption model (see text): (10) number of following lines, each characterizing an individual chemical

-------
                            Equilibrium chemical spctiation by WHAM
         981
species. In the WATER database, only five of these lines art required, corresponding to (I), (2). (5). (8), and (10) in SSED.
In SSED a data line for individual chemical species is interpreted as follows; species number, species name, charge,
composition (up to three master species), stoichiometry (three values), log,, K and AH for the formation of the species Of
roaster species. logw*-999). pKMHA values for humic and fulvic acids, selectivity coefficient for humtc substances, and
selectivity coefficient for the day cation-exchanger. In the WATER database, lines for the individual species are the same
as in SSED, except that the selectivity coefficients are absent. The Iogn£and pK values refer to 25'C, and AH, the standard
enthalpy of reaction (kcatmol'1). is assumed to be constant.with respect to temperature.        ,7
WHAM DATABASE SSBD VERSION 1.0 (SfiEDlO.DBS)
HA paraaatara, 3.29E-3, 4.02, 8.55,  1.78,  3.43,
FA paraaatara, 4.73B-3, 3.26, 9.€4,  3.34,  5.52,
Clay paraaatara, 1E-4, 100
Doubla layar ovarlap factor,  0.25
Log  KSO{25>  t DaltaH for A1(OB)3,  9.0, -25000
tog  KSO(2S)  ft DaltaH for Fa(OH)3,  3.0, -2SOOO
Constant to  control  DDL at low ZED,  '~
Constanta  for FA datorption.l.0,1.0,:
Ho.  of^data  llnac, 197
 2, Be,
 3,Na,
 4,Kg,
 S.A1,

 7'ca,
 8,CrXXX,
 9,nn,
 lO.FoXX,
                                                        -374,  O.S, 1.72B
                                                        -103,  0.4,     8B
9,  15000
10, 1500
 12,CO,

    »Cu,  •
    ,Zn,
 16,Sr,
 17,Cd;
 18,00.
 19,Ba,
 20, Hg,
 21,Sb,
 22,002,
 23.OTV,
 24.PUXXX.
 25.PUXV.
 26.PU02,
 27,Th,

 29^NH4,
 30,Cm,
 51,OH,
 52,Cl.
 S3.N03,
 54,804,
 55,C03,
 56,F,
 57.P04,
 101.HC03,
 102.H2C03,
 104,HF,
 10S,HP04,  .
 106,H2P04,'
 107,H3P04,
 113,BaOH,
 114. Ba (OH) 2.
 115.Ba (OH) 3,
 116,Ba(OH)4,
 117,Ba5O4,
 118,BaF,
 124,KgHC03,
 125,MgCO3,
 126.MgSO4,
 127,HgHP04,
 133.A10H.
1,
,
I,
2,
3.
1,
2,
3.
2,
2,
3,
2,
2,
2,
2,
2,
2,
1,
2,
2,
2,
2,
4,
1,
4,
2,
4,
3,
1,
3,
-1,
-1,
-!,
-2,
-2,
-1.
-3,
-1,
0,
0,
-2,
-1.
0,
1,
0,
"1.
.-2,
0.
1,
1,
o,
o,
0.
2,






9,0,0,
10,0,0,



15,0,0,











28,0,0,
29,0,0,
30,0,0,
51,0,0,
52,0,0,
53,0,0,
54.0,0,
55,0,0,
56,0,0,
57,0,0,
1,55,0,
1,55,0,
1,56,0,
1,57,0,
1.57,0,
1.57,0,
2.51,0,
2,51,0,
2.51,0,
2,51,0,
2,54,0,
2,56,0,
4,1.55,
4,55,0,
4,54,0,
4,1.57,
5,51,0,






1,0,0,
1,0,0,



1,0,0,











1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,0,0,
1,1,0,
2,1,0,
1,1,0,
1,1,0,
2,1,0,
3,1,0,
1,1,0,
1,2,0,
1.3,0,
1,4.0,
1.1,0,
1.1,0.
1,1.1.
1,1,0.
1,1.0.
1,1,1,
1,1.0,
999,
999.
999 f.
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
999,
10.329,
16.681,
3.18,
12.35,
19.55,
21.70,
. 8.6,
14.35,
18.75,
18.59,
1.95,
5.2,
11.40,
2.98,
2.37,
15.26,
9.01,
ft,
0'
,
o,
0,
0,
0,
0,
0,
9,
0,
o.
o.
o.
0,
o.
0.
0,
0,
0,
0,
0,
0,
o.
0.
0..
q.
0,
0,
o,
0,
0,
0.
o,
o,
o.
0,
-3.561
-5.738
3.18,
-3.5,
-4.3,
-2.4,
0.0,
0.0,
o.o,
0.0,
0.0.
0.0.
-2.77,
2.71,
4. 55,
-O.S,
-1.54,
999, 999,
17 O A
999, 999,
3.3, 2.2,
1.3, 0.4,
999, 999,
3.2. 2.2.
O.S, 0.1.
3.4, 1.7,
2.1, 1.3,
0.8, -0.2,
'2.7, 1.7,
2.7. 1.4,
1.5, 0.8,
2.3, 1. ,
2.7. 1. ,
999, 99 ,
3.6, 2. .
0.2,-0. ,
1.7. 0. .
1.3, 0. ,
0.0, -0.7,
1.7, 0.8,
0.0. -0.7.
1.6, 0.5,
0.6, -0.4, •
1.2, 0.3,
999, 999,
2.0, 1.6,.
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999.
,999, 999,
,999, 999,
999. 999,
999, 999,
999, 999,
999, 999,
1.7, 0.4,
999, 999;
999, 999, .
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
999, 999,
1.3, 0.4.
1.
4
*»
0.25,
'0.75,
0.5.
I,
1,

1m
1,
1,
1,
1,
1,
i.
1,

1,
1.5.
i.S-
i.
i.
i.
i.
i.
i.
1.
i.
i.
i,
i,
i,
i,
i,
i,
i,
i,
i,
i.
it
i,
i,
if
i,
i,
i,
^
i,
i,
•if
i,
i,
i,
i.
0.5,
1
1
1
1
1
x

1





i
i_
i
j.
3,





1
0
0
0
0
0
0
0
0
1
1
0
0
1
1
1
0
0
1
1
1
1
1
1
•1

-------
   982

     134,Al(OH)2,
     135,A1(OH)4,
     136,US04,
     137,UF,
     138,X1F2,
     139,A1F3,
     14S,CnHC03,
     146,CaC03,
     147,CaS04,
     148,C«HP04,
     154,CrOB,
     155,Cr(OH)2,
     156,Cr(OB)3,
     157,Cr(OH)4,
    158,CrS04,
    159,CrF,
    165,HnOB,
    166,Mas04,
    167,K&C03,
    168,HaCl,
    169,KaHV04,
    170,MnHC03,
    175,F«OB,
    176,F«S04,
    177,F«C03,
    178,F«HP04,
  " 179,F«C1,
   180,F«BC03,
   185,FOOB,
   18£,Fe(OB}2,
   187,Fa(OH)3,
   186,F.(OH) 4,
   190,F«S04,
   191,F«F,
   192,F«F2,
   193,F«BK>4,
.   194, Fod,
   19S,F«C12,
   201,CoOB,
   202,Co(OH)2,
   203,CoS04,
   204,CoCO3,
   20S,COC1,
   206,CoBC03,
   211,NiOB,
  212,Hi(OH)2,
  213,R1S04,
  214,NlC03,
  215,N1C1,
  216,KlBC03,
  221,CuOH,
  222,Cu(OH)2,
  223,CU604,
  224,CuC03,
  225,Cu{C03)2,
  226,CttCl,
  227,ciiHC03,
  232,ZnOH,
  233,Zn(OH)2,
  234,ZnSO4,
  23S,ZnCQ3,
  236,ZnCl,
  237,ZnBC03,
  242,8rS04,
 243,8rC03,
 244;SrBC03,
 249,CdOB,
 250,Cd(OB)2,
 251,CdSO4,
 2S2,CdCl,
,2S3,CdC12,
 259,BaSO4,
      E. TIPPING
    1, 5,51,0,
   -1. 5,51,0,
    1. 5,54,0,
    2, 5,56,0,
    1, 5,56,0,
    0; 5,56,0,
    1, 7,1,55,
    0, 7,55,0,
    0, 7,54,0,
    0, 7,1,57,
    2, 8,51,0,
    1,  8,51,0,
    0,  8,51,0;
  -1,  8,51,0,
    1,  8,54,0,
   2,  8,56,0,
   1, 9,51,0,:
   0, 9,54,0,
   0, 9,55,0,
   1. 9,52,0,
   0, 9,1,57,
   1, 9,1,55,
   1, 10,51,0,
   0, 10,54,0,
   0,  10,55,0,
   0,  10,1,57,
  1,  10,52,0,
  1,  10,1,55,
  2,  11.51,0,
  1,  11,51,0,
  0,  11,51,0.
 -1, 11,51,0,
  1, 11,54,0,
  2, 11,56,0,
  1, 11,56,0,
  I, 11.1,57,
  2, 11,52,0,
  1, 11,52,0,
  1, 12,51,0,
  0. 12,51,0,
  0, 12,54,0,
  0, 12,55,0,
  1,  12,52,0,
  1,  12,1,55,
  1,  13,51,0,
  0,  13,51,0,
  0, 13,54.0,
  0, 13,55,0,
  1, 13,52,0,  :
 1, 13,1,55,  :
 1, 14,51,0,  ;
 0, 14,51,0,  ;
 0, 14,54,0,  :
 0, 14,55,0,  :
-2, 14,55,0,  3
 1, 14,52,0,  J
 1,  14,1,55,  3
 1.  15.51,0,  J
 0,  15,51,0,  J
 0,  15,54,0,  J
 0,  15,55,0,   1
 1,  15,52,0,   1
 1.  15,1,55,   1
 0,  16,54,0,   1
 0,  16,55,0,   1
 1.  16,1,55,   1
 1.  17,51,0,   1
 0,  17,51,0.   1
 0, 17,54,0,  l
 1, 17,52,0,  l
 0, 17,52,0,  l
 0, 19,54,0,  l
  li
   : : :
  8*
  £Xk
  JJ:!5:
         Jl.lf-
     :  a
   j ..» «
   1      •'
     -    ' :
•
   •'
 1.1..
 ttfc
 Sift
ttfc
1.1,1,
1,1,0,
 :|:J;
 ::;
feifc
1,1,1,
                  o.o,

  MI:
  «£
12.33
ll.ei,'

                          •
.
 !: !•   ! •'   »-
          '
   .
 6.0,  999  JJJ'
         '    '-
     ,
AS:
21.55,
1:1: .»»•
J; '  Jff* fff'
         '
                           »»»•

                              '
                      »**•
   «    e          '     »    .
  'o'   if*?*'   S-°'  2-3' *•»•
  'o'    Hi    ,°*°'  999, 999.
  'J'    J*?!'   X-5'  »»»• 9"'
  'S'    M6'   °-9'  »»»• »»»»
                             '
               2-OB' 999»
      *«
   '   I'll'   2'49' 999'
   '    ''   °'°'  2'7*
            '
        .,
       2.7,
             999f
1,
1,
1,
1.
1,
i.
1,
1,
I.
1,
1,
1,
1.
1,
1,
1,
1,
I.
1,
1.
1,
i.
I,
1,
1,
1,
1,
1,
1.
1,
1,
1,
1,
1,
1,
1,
1.
1,
1.
1,
1.
1.
1,
1,
1,
1,
1,
i.
1,
1,
1,
1,
1.
1,
1.
i.
1, •
1,
1,
1,
1, •
i.
i.
i.
i.
i.
i.
i.
i.
i.
i.
i.
i
0
i
i
i
i
l
i
l
i
i
i
i
0
1
1
1
1
1
1
1
• 1
- 1
1
-1
• 1
. 1
1
1
1
1 .
0
1
1-
1
• 1
1
1 .
1
• 1
1
i
i
i
i
i
i
i
i •
i
i
i '
i
i
0
'i
i
i,-
i
i
i-
l
i
i
i
i
i .
1
i
i
i
i
                                            i

-------
                        Equilibrium chemical cpccialion by WHAM                         983
                                          it*
                                                                        i.     i
                                                                        if     i
                                                                        i,     i
                                                                        i,     i  .
                                                                        if     o
                                                                        i*     i
                                                                        i.     i
                                                                        i«     i
                                                                        i,     o
                                                                        i/    o
                                                                        i,     i
                                                                        i,     i
                                                                        1,     0
                                                                        1,     1
                                                                        if     i
                                                                        i,     o
                                                                        i*     i
                                                                        i.     i
                                                                        i,     i
                                                                        i.     i
                                                                        i.     o
                                                                        1.  ;  1
                                                                        1.     1
                                                                        1*     1
                                                                        1,     i
                                                                        1.     0
                                                                        1.     1
                                                                        1,     1
                                                                        i,     i
                                                                        i.     i
                                                                        i.     o
                                                                        i.     i
                                                                        i,     i
                                                                        i>     i
                                                                        i.     i
                                                                        i,     i
                                                                        i,     i
                                                                        i.     i
                                                                        i,     o
                                                                        i,     i
                                                                        i,     i
                                                                        i,     i
                                                                        i.     i
                                                                        i,     i
                                                                        i,     i
                                                                        i,     i
                                                                        i.     i
                                                                        i,     i
                                                                        i.     i
                                                                        i.    i
                                                                        i,    l
                                                                        i.    i
                                                                        i.    o
                                                                        i.   .1
                                                                        1.    0
                                                                        if    1
                                                                        1,    1
                                                                        1,    1
                                                                        1,    1
                                                                        i,   . 1
                                                                        i,    1
                                                                        1.    1
                                                                        .if    1
                                                                        1,    i
                                                                        If.   %
                                                                        i,    1
                                                                        i,    1
                                                                        1.     1
                                                                        1,     0
                                                                        1,     1
                                                                        1,     0
260, 8aCO3,
261,BaHC03,
266,HgOH,
267,Hg(OH)2,
268, Eg (OH) 3,
269,H0SO'4,
270,BgCl,
271,BgC12,
272,HgC13,
273,H0C14,
279,PbOH,
280, Pb (OH) 2,
281, Pb (OB) 3,
282,PbS04,
283,PbCO3,
28«,Pb(C03)2,
285,PbCl,
286,PbCl2,
909 irnOAit.
A7*f ww»*»»f
293,U02(OH)2,
294, 002 (OH) 3,
0,
.if
1,
0,
-1,
Of
1,
0,
-If
-2,
i.
o.
-if
o.
o.
-2,
i.
o.
1-
^f
o.
-i.
295,«J02)2(OK)2,2,
296, (002) 3 (OH) 5,1,
297,002504,
298,002003,
299,U02(C03)2,
a os UOH.
d V9 jf W
-------
   984
                                     ,~E. TIFFING
     WHAM DATABASE HATER VERSION 1.0  (HATERIO.DBS)
     HA parameters, 3.29E-3, 4.02, 8.55, 1.78, 3.43, -374,  O.S,  1.72E-9    isooo
     FA parameters, 4.73E-3, 3.2«, 9.64, 3.34, 5.52^-103*  0.4*     ai-io   «SS
     TXMlKlo Invar nvar-lun ^.^f-n*-. n,5IC             *     *  "**»     BB-10,  1500
     Double layer ovarlap factor, 0.25
     Constant to control DDL at low ZED, IBS
     No. .of data linai, 214
     1,H,            1,  1,0,0,   1,0,0,   999,
     2,Ba,           2,  2,0,0,   1,0,0,   999,
     3,Ha,           1,  3,0,0,   1,0,0,   999,
                                           0,
                                           0,
                                           0,
                                            999, 999
                                            1.7, 0.4
                                            999, 999
                                  APPENDIX!
                               WHAU-W Program Ltsllng

 'l"^-"'^"?1*!!"*11*1"""**"""**"
 'if Computer coda in TORBO BASIC                                            it
 'if Speciation of natural waters containing bunic aubatancea                if
 'if This version produced on 13 April 1993 by E.Tipping,  IFB, Hindemere     if
 CLS

 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
 PRINT
•••••*•**«***«**««•«•*•
•**   W indarnara   ••
"•*   H umlc  .     ' •*
•'**   A  uaoua      •*
••*
*••
•**
  quaoua
Modal
-      '
W atara
                    **
                    **•
                    •*
Equilibrlua apaclatiou of vatara*
Variabla or fixad pH"
Variable or fixad pCO2"
HS praaant or abaant*

E.Tipping  IPE  April 1993r
•**
   .   Varaion 1.0    •*«
 **»•****••*«•«••••••••
PRINTtlNPOT "NAME OF SOURCE FILE IOMIT .DAT TRAILER]
PRINT
OPEN SF$ + ".DAT" FOR INPUT AS fl

INPUT fl,8$,8$
INPUT fl,S$,DBS$
INPUT fl,8
                                            fila haadara
                                            databaaa idantifiar
                                            array aica raguirad
' OIKanaion and claar arraya,  than raad in cbamical conataata

GOSOB WHSR1         .
«*«**•*****•**•«•****«*•••**••*
' Continao to raad in data fila

INPUT f1,8$,PRECISION
INPOT .fl,8$,TENPK     .
IN9OT tI,S$,TH(l),TH<2)
INPUT,f1,8$,FHFIX$
INPUT tl,6$,PH8TART
INPUT fl,8$,PC02
INPUT fl,S$,NCOMPV

FOR I\ « 1 TO NCOMP*
INPUT *l,S$,Xt,TCONC
»$CX*> - 8$
T(X%)  - TCONC
NEXT 1%
                                            praciaion
                                            taap K
                                            total conca (HA, FA)
                                            pH fixad / vaxiabla
                                            atarting or fixad pH
                                            PC02
                                            no. of naatar specie*
                                          *  total conca
                                                                              1
CLOSE SI

-------
                         Equilibrium chemical spcciation by WHAM
                                                              985
f**«******************
' Perform calculations

COSDB KWSR2



' calculations finished ; signal and make output file

SOUND 400,5  s  SOUND 500*5  t SOUND 700,6       * sound output


OPEN SF$ + ".OUT" FOR OUTPUT AS 12               ,                  '

COSUB WHSRB

CLOSE «2                                               -          "     ' ,

PRINT
PRXHT "CALCULATION FINISHED f CONSULT *.* ";SF$;".OUT ** FOR DETAILED OUTPOT"


STOP
END
 KWSRlt
•ft
'if
«ii
called by main program
DIMensions arraya
reads in data from data base
sets up Model V
                                                                             i*
                                                                             fff
                                                                             ft
                                                                             it
* DIHa&sion arrays t inorganic chemistry
DIM H*(NSP*>
DIM N$(NSP\)
DIU T(HSP\)
DIM TCALC{NSP\)
DIM A(NfiPX)
DIM C(NSPX)
DIM CHMNSPM
DIM M1MNSP*}
DIM M2«t(NSP*>
DIM M3MNSPX)
DIM 6»(N8P«)
DIM S2%(HSP\)
DIM 83«{N8P%)
DIM UC(NSPIc)
DIM DH(NSPX)
DIM CAMMllfS}
 ' DIHension humic arrays

 DIM 8P(2,6)
 DIM SH(2,12)
 DIM KH(2,8)
 DIM PKH(2,B)
 DIM KMH(2f8,ilSPX)
 DIM SITEM12,2)
 DIM PKHKA(2,HSP*)
 DIM PKKHB(2,NSP\)
 DIM FP(8)
 DIM BIBTERM(NSPIt)
 DIM MONBTERMCNSP*)
 DIM BISNO(12,NSP*)
 DIM MONSNO(8,NSP%)
 DIM BIN0(NSP*)
 DIM MONNU(NSP\)
 DIM NU(2,NSP*)
 DIM CHC(2,NSP%)
 DIU CDDL{2,NSP%)
                  numerical identification of species
                  nominal identification of species
                  total concentrations of components
                  total calculated concentrations of components
                  activities of solution species
                  concentations of solution species
                  charges on solution, species
                  master species identifiers
                  master species identifiers    '           .
                  master.species identifiers
                  stoichiometries
                  stolchiometries
                  stoichiometries
                  equilibrium constants
                  enthalpies
                  activity coefficients
                  BA,  2 » FA

                  proton-binding sites not forming bidentate sites
                  bidentate  metal-binding sites .
                  K's  for proton-binding
                  PR's for proton-binding               .
                  K's  for metal-proton exchange
                  proton sites making bidentate sites
                  metal-proton exchange  constants
                  metal-proton exchange  constants
                  dissociation factors for proton-binding
                  terms calc'd in finding bidentate theta's .
                  terms calc'd in finding monodentate theta's
                  values of  HU at each bidentate site
                  values of  NU at each monodentate site
                  overall. HO for each metal/bidentate sites
                  overall NU for each metal/monodentate sitea
                  overall NO for each complexed species
                  cone of complexed species per litre
                  cones in KS DDL's                            .

-------
986
                                    E. TlPNNG
 DIM BIZ(12)
 DIM HONZ(B)
 DIM BXNDTEST*(NSP*)
                                net charge at each bidentate site
                                not charge at each monodentate  sit*
                                allows binding of a species  to  bo tasted
  Input of constants for inorganic  speciation and HS coaplexation

OPEN DBS$ + -.DBS" FOR INPUT A8 f 2

INPUT *2,S$                                .      ' database identifier
INPUT f2,S$.NCCKJH(l),Pia»(l),I>KHB(l),DPKHA(l),_
      DPJCHB(i)f9(l),FPR(l),UU>ias(l),KOLMT(l)   ' HA properties
INPUT t2,S$,NCOOH(2),PKHM2),PKHB(2>,DPKHM2),_
      DPKHB(2),P(2),FPR(2),RM>IUS(2),MOLWT(2)
INPOT «2,S$,DLF
INPOT i2,S$,KZBD
INPOT t2,S$,NODATA*            .   '     '
                                                  ', FA properties
                                                  ' double layer overlap factor
                                                  ' Dt vol at low Z
                                                  ' nunber of •peciea
FOR I* «  1 TO NODATA*

INPOT f2,r
IF PKKB&(l,Xk) < 999 THEN BINDTEST*(X*>  •  1

NEXT I«6
                            ^

CIiOSB «2
                                                  ' identifies species that
                                                  ' don't bind to HS
«**••****«****•**•*******•***•*****
' Set derived constants for Model V

FOR X* • l.TO 2
PKH(X*,1)
            PKHMX*) -
            PKHA(XM +
                     f (DPKBA(X%)/2)
                     - (DPKHB(5(1)
SITE*(9,1)
SITE%(12,1)
                    SITE*(1,2)
                    8ITB\(2,2)
                    8ITE\{3,2)
                    8ITE\(4«2)
                    8ITB%(7,2)
                    8ITB\(8,2)
                    8ITEV(9,2)
                    8ITB\(10,2)
                    SITB*(11,2)
' Set site concentrations

FOR I\ • 1 TO 4
SP(l,Ht) « (1 - FPR(l)) * NCOOB(l)  /  4<
SP(2,IM •> (1 - FPR(2» * NCOOU(2)  /  4
NEXT I*

-------
                         Equilibrium chemical specialion by WHAM
                                                                           . 937
 FOR 1% « 5 tO 8                     -
 SP(1,I\)  ~ (1 - FPR(l))  * NCOOH(l) / 6
 SP(2,IX)  - (1 - FPR(2»  * NCOOK(2) / 8
 NEXT X*

 FOR IX «  1 TO 12
 SM(1,X*)  m FPR(l)  * NCOOH(l)  /IS
 01(2,1%)  - FPR(2)  * NCOOH(2)  / 1«           .      .
 NEXT 1%


 * Sat natal-proton exchange constant*  for individual sites
                          i
 FOR I*  «  1 TO K8F*

 IF  N$(I*)       • ""  THEN jniBMM                 •  species not defined
 XF  PKKB&(1,X\)  • 999 THEN
FOR fl* « .1 TO 4
               PKMRAUfX*)
PJMH
KMH(2
NEXT
 FOR Jit « S  TO 6
 PKKH         - (3«PKMH*.(1,H())  - 3
PK«a
NEXT J\
GOTO WWSRU.2
                3.96*PKMHA.{2,Ht)
                10*(-PKMH)
                                                  '  3*A -  3 conversion >  RA.

                                                  '  3.96*A  conversion i  FA
' if eoaa to hore, than no binding of species l% con occur
tnrsRiLit

FOR J\ - 1 TO 8
KKB(1,C\,Z«) - 0
KHH(2, J\«X\) - 0
NEXT A            •
 •k

HWSRIUt
NEXT X*

RETORN                            •     •
 MH8R2I
'••
'••
'«

'••
'*•
'ft
             colled froa uoin program ••-• calls KMSR3  and MT8R4
             sets initial trial valuas  o£ master species'
             initialises all activities and concns by calling WMSR*
             controls pH laprovement  If pB  not  fixed
             controls level of precision
             calls MH8R3 to cole speciation at  a given pH
             tests fox correct pH with  CERATXO  if pH not  fixed
             «ats up screen to report on progress of calculation
                                                                             ft
                                                                             ft
                                                                             ft
                                                                             ft
                                                                             it
                                                                             ft
                                                                             ft
                                                                             ft
'ii»«Hi*ftHII«l«iff»»ifHIIHt»*li»HHH«MHt»»StH»fltiit»lft«ftilit

' set initial trial values         -
XS « 0.1  '"

ZED(1>. * -1E-4
ZEO(2) - -1E-4

RATIO (1) • 10
RATIO (2) - 10

PH « PHSTART
Ml) - 10M-BH)

FOR X* « 2 TO SO
A(X»s) i T{X*)
                                                  • ionic strength

                                                  «. c fof HA
                                                  ' Z f or FA

                                                  » RATIO for BA
                                                  • BATXO for FA

                                                  * fixod or starting pH
                                                  ' activity of  H+

                                                  . apecies il is  R+
                                                  . catiohic master species

-------
988

NEXT Xfc

A<52)
A * TH(2)) > 0     THEN PRINT "HOMIC SUBSTANCES
IF PBFIX$ « "NO"           THEN PRINT •              PH
IF PHFIX$ • "YES*          THEN PRINT •              PH
IF PC02 • 999   .           THEN PRINT "            PCO2
IF PC02 < 999          .    THEN PRINT «            PCO2
                                    SOURCE FILE
                                  PRECISION (\)
                                  USING -f.fi*** "|1.0001«PRECISION
                                                  ABSENT-
                                                  PRESENT"
                                                  VARIABLE"
                                                  FIXED-
                                                  VARIABLE"
                                                  FIXED-
PRINT   /                 .

PRINT "ITER*;TAB(15);"PH";TAB(30)»"IS";TAB{4S);"CHRJITIO«;TAB(62>j«DPH"
' Begin calculations

NOMIT* » 0  .

CRIT » 1E-4 s  COSUB WKSR4

IF PHP1X$ • "NO" THEN HWSR2M.
                                         ' itaration countar

                                         ' initialization

                                         ' routine if pH variable

                                         ' routina for
IF FHFIX$ m "YES" THEN CRTEST$ • "PHASE2"
IF PHFIX^ - "YES" THEN CRIT • (PRECISION/100) *2  ' fixed pH calen
IF PHFIXJ • "YES" THEN GOSUB WWSR3
IF PHFIX$'« "TES" THEN WHSR2L4
                                         ' fixed pH calc* done; return
«****•**•**•*****•********•**•*«*****•*****«*•«
' D»» pR adjuating routine, solving for each pH

HWSR2L1I

CRTE8T( • "PHASE1"

CRIT. « 1E-4 t GOSUB WWBR3

DPHP m-Q.2

IF CHRATIO < 1 TREK WWSR2I.2
IF CHRATIO > 1 THEN WWSR2L3
                                         ' find initial CHRATIO etc

                                         ' initial pH adjust factor
' Coae to here If CHRATIO is < 1

VWSR2L2:
DPH •> DPHF / CHRATIO
IF DPH < 0.01 THEN CRIT
                          (PRECISION/100)*2
                                         ' pH increment
                                         ' adjust precision

-------
                        Equilibrium chemical spctiation by WHAM
                         989
IP DPH < 0.0001 THEN CRTEST$ • "PBASE2"
PH - PH - DPH
GOSUB WWSR3
IP CRTESTS » •PHASB2* AND_
ABS(CHRATIO - 1) < SQR(CRIT) THEN WWSR2L4
IP CHRATIO > 1 THEN DPHP • DPHP/3
IP CHHXTIO > 1 THEN WWSR2L3
GOTO KWSR2L2
  refining
  increment pH           '
  calculate speciation

  finished
  change pH adjust factor
  go to > routine
  eontinua to ineramant pH
«++++++*++**++++++++++++++++++++
• Coma to hara if CHKXTIO i> > 1

WWSR2L3I

DPH • DPHP * CHRATZO
ZP DPH < 0.01 THEM CRIT • (PRECISION/100)*2
IP DPH < 0.0001 THEN CRTEST$ • "PKASB2"
PH » PH 4 DPH
COSUB KWSR3
IP CRTEST^ * "PHXSE2- AND_
ABS(CHRJLTIO - 1) < SQR(CRIT) THEN WWSR2L4
ZP CHRATIO < 1 THEN DPHP - DPHP/3
IP CHRATIO < 1 THEM WWSR2M
GOTO WWSR2L3

WWSR2L4»
RETURN
  pR ineramant
  adjuit praciaion
  rafiaing
  ineramant pH
  calculata cpaciation

  finlahaa
  ehanga pH adjuat factor
  go to < routina
  eontinua to ineramant pH
'•ffiftitiittftHi»titlftttffit»fi«»tifft*ff»»ffftff««Mftl»itliSfffflf»ft«ill«i
 WWSR3t    ' eallad by KW8R3 - call« WWSR4                        .          it
•if          control* improvamant of activitia* (not H+)  and ZFA,ZHA       ' it
•iff          call* MMSR4 to do tuuiB and charga calculations                . ii
•ii-         -in PHASE 1 ratuxna'Khan CHRATZO ha.a bacoma nearly-eoiuitattt     ii
'ii       >   in PHASE 2 return* when aua and charga balancaa O.K.          ii
•ii          raporta prograB* of  calculation to acraan                      ii
•iiiiiiiiiiiiiiiiiiiiiiiiiiiitiiiiittfiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiitiiifititf
ACll » 10*C-PH)

liASTCHRATZO « 0 I LAST1CHRATIO « 0
' activity of H*

' raaatting vmluaa
• Begin iterative cycle ; cone, back to hare' until convergence is achieved

WKSR3Llt

       •> NOMIT* + t                              ' update iteration counter
 **************************
 ' Calculate ionic strength

 IS - 0                        s

 POR X* « 1 TO MSP\
 IP C 0 THEN IS • IS + (0.5 *  100 THEN IS • 100
' preparing to sum


   • OK(X*))>


' avoids initial high IS
 «•«««••**•*•**«*•*•**•
 *  Improve trial values
 FOR X% •  2 TO 50
 IP T(X*>  » 0  THEN WWSR3L2
 CONCFACTOR «  T(X\)/TCALC(X\)

-------
990
                                    E, TIPPING
A(Jt*)  .
HHSR3L2I
NEXT XV-
                SQR (CONCFACTOR)
' cation activities
 FOR X* » 52 TO  100
 IF T(X\) • 0 THEN WWSR3L3
 CONCFACTOR • T(») /TCALC 0 AND TESTITER • TESTXTBR* THBN_
                   ZBD(1> . ZED(l) +  ((ZCALC(1)-ZED(1))/S)
 IF TH(2) > 0 AND TESTITBR « TESTITER* THBN_
                   ZBD(2)- ZED{2> +  ((ZCAM{2)-ZED(2))/5)
IAST1CBRATIO
LASTCHRATIO
               LAETCHRATIO
               CBRATIO
* Calculate total oonens of mastar apaeiai
COSUB HWSR4  .
' racoxd valuQv to avoid
' finding complete aola. in
' firat phaaa - saa balow
*********************************             •     •   •               -
' Output currant status to screen

LOCATE 20
PRINT NDHZT%;TAB(13)|USINO •«.*W"|5H;TAB(28)|TOINC  «*.*f*****"|IS;TAB(41);
                      USIKO "«.8«i«|CHRATIO;TAB{fiO)7DPH                   ~

'••*«*••*•*•*•****«*••***•******••*•*•••«*•••«*••*«•«•*••«««*«••»*«*•«•«*••«
' Teat whether CBRATIO nas been found to. acceptable, precision in first phase

IF CRTEST$ • "PHASE1" AMD_
(LASTCHRATIO / CHRATZO) >'0.9  AND
{LASTCHRATIO / CBRATIO) < 1.1  AHD_
(IAST1CHRATXO / CBRATIO} > 0.9 AND
(IAST1CHRATIO / CBRATIO) < 1.1 THEN WWSR3L6
IF CRTEST$ « "PHASE1- AND.
LASTICBRATIO • CBRATIO THEN WN8R3L6
                                                   return
                                                 •' oscillating - return
«***««*•*****«*••••••***«•«•***«*•*••••**»•«•••**•**«••••••••*««*«
' Test cations, anions, Z's for convergence - note that final test
' is CHRATIO, done in pH-adjusting routine
FOR X\ • 2 TO 50
IF T(X\) - 0 THEN HHSR3L4
CONCERR • (2«(T{X\) - TCALC(X«())/(T(X%)
IF CONCERR > GRIT THEN KHSR3X.1
WWSR3L4«             .
NEXT »
                                          TCAM(»)))*2
               ' cations

               ' error term
               ' re-iterate
FOR » m 52 TO 100
IF T(X\) m 0 THEN KK8R3L5
CONCERR - <2*(T GRIT THEN WWSR3ML
HWSR3LSI
NEXT X%
                                                                  sinions
IF (TH(1) + TH(2)) • 0 THEN WWSR3L6
IF ZBRR(l) > GRIT THEN WWSR3L1
IF ZERR(2) > GRIT THEN WWSR3L1
                                                                 ' error tern
                                                                 ' re-iterate
                                                 ' no humics ; return
                                                 ' re-iterate
                                                 ' re-iterate
WWSR3L6:
RETURN

-------
                           Equilibrium chemical spea'alion by WHAM
                                                                              991
 WWSR4*
•ft
'if
*fi
•ft
              called by NKSR2 and HWSR3 - calls WWfiRS
              calls WWSR5 to calc activities,. cones of complexes and
                   amount*, bound Toy FA and BA                       •
              suns to gat total ealc'd concna
             . calcsr+ve and -ve charge and CHRATIO
                                                                               ftf
                                                                               ••
                                                                               JS
  * Calc activities and concns of inorganic complexes, annte bound to FA & HA

  GOSOB WWSRS


  ••••••*•«**<
                                           — — —       — *,—— *,mt*mmmmmmmm9m
  • DO  summations to obtain total calculated eoncns of aaster species

  FOR X* « 1 TO 100

  IF A(X*> - 0  THEM TCALC(XM  •  0
  IF A(») - 0  THEN WWSR8L1

  IF X*   - 1  THEN WWSR8L1
 IF X*   -51 THEN KWSRBL1
FOR
           - 0

          1 TO NSP*
 IF Kl*(rt> - X* THEN TCALC(X%) •  TCAXiC(X*)_
           4 (VOLSOL '   * C(t\)
           .4 (CHC(l,Vb)
           4 (CHC(2,W      ,  .,_
             	     CDDL(1,T\)
           4 (DVOL(2)

  IF M2*fl%) - X% THEN
           4 (VOLSOL
           4 (CHC(l.Tdt)
           4 fDVOL(l)
           4 (DVOL(2)
              > 0 TKEH POSCH « POSCH +  (C(»J *
           < 0 THEN NEOCH . NECCH -  (C(») *
         . (POSCH/NECCH)
CBRATIO

RETURN
                                                       '  preparing to sun

-------
 992
                                    E TIPPING
             called by WWSR4 - calls MWSR6 and WSR7
             calcs OH- activity, act coeffs, COS, 2- if pC02 fixed
             calcv activities, cone* of inorganic conplaxes
             every 2nd itertn, calls WWSR6 to g«t tnj'm and Z's fox BS
             call* WWSR7 to gat binding by DI* accumulation (FA,HA)
 HHSRSt
'II
'II
'II

'iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiliiiiiiiiiiiiiiiiiiiii

' Calculate activity of OH- fron A(l),  temp,  deltaH                      •
                                                                              II
                                                                              f|
                                                                              §§
                                                                              ||
                                                                              |§
UCtf   « - 14 +
A(51> -
                (2935*  (0.003354 -  (1/TBMPK)))
 >***•**••*«***********•**•**********•*•*****•••***•*«•**«*•*•
 ' Calculate activity coefficiehts using extended Debye-Huckel
GMOIA(O)
ACTA
ACTB
ACTC
ACTD
GAKKMl)
GAHHA(2)
GUat&(3)
GAtOi&(4}
1
0.270 +
0.330
- ACTA
ACTB
10* (
10* ( 4
10* ( 9
10* (16

(8E-4*TBMF

* SQR (IS)
* SQR (IS)
ACTC/d +
* ACTC/U +
* ACTC/(1 +
* ACTC/(1 4

K)



( 3
( €
( 9
(12



x

• ACTD)))
• ACTD}))
* ACTD}))
*• ACTD} ) )
                                                      ' const 'A in D-H
                                                      ' const B in D-H
                                                      ' act eo«ff M+/-
                                                      ' act eoaff K2+/-
                                                      ' act co«fC M3+/-
                                                      ' act coaff K4+/-
»•**«••««•••••*«•«*•*••*•***••*****«•«**•**••••
• calculata concantrations of inorganic •p«el«*

' Firat do tba carbonate aystam

IF PC02 « 999 THEN MWSRSL1

ENTHTERM « 220 *  (•0.962) *  (0.003354 - (1/TEHPK))
A(55) « FC02 / 10A(18.149 * ENTBTERU) / A(l) / A(l)
                                                      ' pCO2 not apaciflad
                                                      ' activity of C032-
WWSR5I.ll


FOR »( • 101 TO HSP*

IF S1*(X*) • 0 THEN VWSR5L2
           (

IF S1«(X«)' > 0 AND A(K1\(XK}) •  0 THEN KMBRSU
IF 62MX*) > 0 AND A(K2\(X«c)) -  0 THEN WH8RSL2
IF 83%(X\) > 0 AND A(U3\(X\» •  0 THEN KWSRSL2


EKTHTZRM - 220 *  DB(X*)  *  (0.003354  -  (1/TEHFX))
LOOACT • LK(Zk)  +  EKTHTERM
LOOACT • LOQACT  +  (8l«(Xk)
LOQXCT - LOGACT  +           *
                                                      * coaplax not dofin«d
                                                                    i
                                                      ' no calculation if
                                                      ' contributing cpvciaa
                                                      ' absent
 IF S3\(X«c) >  0 THEN.
 LOGACT • IAGACT  + (E3\(»)  *  LOO10(A(M3\(3C*J ) ) )

•A(A)  « 10* (LOGACT)

 VWSRSL2I
 NEXT X*               .
 '  Calculate concns from activities

 FOR X* m i TO HSP*
 CHARGE* - ABS(CH\(XV))
 C(X*)
 NEXT X*

-------
                          Equilibrium chemical spcdation by WHAM               .        993
  ' Calculate concentrations of hunlc -bound coaponantv

  * Chock if th« iteration BO. (NOXXt*) is a aultiple of 2 j if not. ratum
  ' Note that TBSTXTBR, TBSTXTEK* are alto ui«d in Bubroutin* WWSR3

  TESTITBR    « NDKXT*/2
  TESTZTBRV .  • INT{NUMITV2)
  IF TESTXTER > TBSTXTERfc TEEN NMSR5L5             »  no esale tnia lt«r; return


  FOR HS* • 1 TO 2

  IF THCHS\)  . 0 THEN WWSRSL3                      '  no calcn if HS ab«nt

  GOSUB WWSR6                                      .  calculat** ZC&LC'c and WJ'


  •  Caleulata eoncn.  of Bptcifioally bound spaciaa  par litra from HO and [HS]

  FOR XV «  1 TO NSP\
                             TH(HS\)
 NEXT    .

 *WSR5L3»
' MEXT HSV                          -  •  •              '           '...'•


 ' Calculata aaximua voluman of BH and FA diffu** layarv

 FOR HS\ m i TO 2

 IF TOiHS%| . 0 CHEN WWSRSM                           'no cojcn if HS mbBent

 VTERM1       - RADIUS (HSX) 4 < 3. CUB- 10 / 8QR(XS)}
 VTERH2
 VTERM3       « 4.19 * VTEKH2
              . 6E23 * VTKEM3 • (1000 /
              . DDLVOL(ES\) • TH{HS\)
 * Adjust diffuse l«y«r voluaa for low ZED
 ZTBRH        . KZBD • ABS(ZED(BSk))
 DTOI«KCSS%)  . DVOUOUC(BS%) • ZTEKH / (1 + ZTBRH)     ' muc Vol. litras/litr.

 WKSRSMt
 NEXT BS%        -


 '  Caleulata tha actual dlf fu*a layer voluaac, u«ing DLF

 OENOIC • 1 + {(OVOUOX(l) * DVOUOX(2)) /DLF)

 FOR HS* - 1 TO 2                               '

 DVOL(HS\) - OVDUOX(HS\) / DKNOM                -
 NEXT HS\
      /      t             .                  •

VOLSOL «  1 -  DVOLC1)  - OVOZr(2)

 FOR HS* « 1 TO 2                                                .
            >  0 THEN GOSUB WHSR7                       . cftlc DDL comsn.
WWSRSLBi
RETURN

-------
 994
                                    E. TIPPING
 WWSR6:    • called by HWSRS        '.'     M   '                              **
«ff          calcs NO (specific binding) and z for FA, BX                    §§
'ifti*««iftlltKifff»f*«fftti*tfliliftl»H»t»«ftilttftttfitlitttiiii«if«tilf«*
K • P(HS*> * LOGIC(IS)
                                                   •'static interaction factor
TEHPVM. » KB(HSl
FP(X*>  - 1/d + TEMPVMi)
NEXT X*
                                                 ' protonation factor*
* Binding at bidentate sites

FOR 3ftt .m l TO 12   .
SYTB1* -
SYTE2* -
                                                 ' do «acb >it« in turn

                                                 ' idontifiai proton *it«i that
                                                        up th« bid*ntat«
SUMBITERM • 1

FOR K* - 1 TO HSJP*
IF BIHDTEST\ 2
BXZ(OX) « BXNBTCH *

NEXT J*t
                                                              2 bound fit
                                                              H4 bound at site 1
                                                              H+ bound at «ite 2
                                                              net charge
                                                              thi» cite'* charge
 '  How calculate total uaounti bound, and total net charge (bidentate cites)

 FOR Kit • 1 TO MSPX          .   '                                     ,
 XF BXNDTESTMKfc) " 0 THEN WHSR6L3
 BZNO(K*t) - 0

 FOR n m 1 TO 12
 BXN(7(1C*) -
 NEXT J*
 NEXT

-------
                          Equilibrium chemical ipedalion by WHAM                       995

 BXZCALC «.0
 FOR J* - 1 TO 12  ,
 BZZCALC - BIZCALC 4 BIZ(J\)
 NEXT «J%        •


 »*•**••«******•**«•*•****•«**«          •   •    •     '           •
 '  Binding at monodantata «itac

 FOR J\ - 1 TO 6

 EUHMONTERM « 1                             .

 FOR K* • 1 to NSP*
 IF BINDTESTMK*)  • 0 THEN WH8R6X4
 TEHPVAb      - 2*H*ZED(HS*)*(1 - GB«(Xk))
 TEKPV3O.      • EXP(TKMPVM.)
 TBMPVAL      - nat(BS%,J
 MONBTBRM(K«t) • TEHPVAL*F
 STJMMONTERM   - SOMMONTERM + MOKBTKRM{K*)
 MHSRfiXiis
 NEXT Kk                                                                ''


 SOKMOHTHETA • 0                                  ' prftparing to Sim
 HONMETCH    • 0                                  ' thttta'* and chargos

 FOR K% « 1 TO HSPUt
 IF BXKDTESTMK*)  • 0 THEN HHSR6I.5
 MOHTHBTA      « HOHBTBRH(KV)/SUUHONTSRH
 SUMMONTHETA.   • 80MKONTBETA + KOHTEBTA
 HONSmT(0^,K%)  * MOOTHETA'SPCES*, J»)
 HONHETCH    ' • HOlOtETCH + {MONTHETA*CB\(1CK} )
 KWSR6L51   .
 NEXT K*


 PROT1CH  - FP(Olt) *  (1 - SOKMOHTHETA)            * H+ bound
 KOHHSTCB • KOSDOCTCH 4 PROT1CH - 1                * nat eharga
 KOHZ(OX) - MONNETCH • 8P(B8%,0%)

 NEXT* Jit        .            '    ,        ' '      ,                 •


 *  Now calculate total amount* bound, and total nat charg*  (taonodantata *ita*)

 FOR K* « 1 TO NSF%
 IF BINDTEST*(K%)  -" 0 THEN HHSR6L«
             0
FOR 0* «  1 TO 9
MONHa(Kfc)  . MONNa(KHc) 4- XONSNa<3*,K*)
NEXT J%                             •

«HSR6XiCs                •       .
NEXT K\        '           ~
KONZCAIiC • 0

FOR J\ - 1 TO 8
KONZCALC • MONZCALC 4 MONZ(0%)
NEXT J%
_***********•**•*«*****•*•**•****«**»•**••«***
 •  overall  summation j bidantate 4 aonodantata

 FOR K* • 1 TO NSP*
 IF BINDTESTMK*)  * 0 THEN WWSR6L7

 WWSR6L7:
 NEXT K%

-------
  996                                , E TIPPING

  ' Calculated value of Z, and Z error tarn

  ZCUiC(HS*) « BIZCAZiC + MONZCALC
  ZBRR(HS*)  m (2MZBD(HS*)~ZCAX£(HS*))/(ZSD(KS*)*ZCALC(KS*))JA2

 RETURN                                              •
  KWSR7t   '* called by MfSRS
 'it          calca binding by DL accumulation, for PA and HA                 *•
 '*tiiHt*»iHff««HHfHfffHHifMIHIIHiHtlimHIIHIfiHHIif«|iffff{«f»t

• ' First clear the DDL array for this HS*

 FOR J% m 1 TO NSP\
 CDDL(HS»,J*) • 0
 NEXT JV                       '      .
 TOTCHH *  ZED(HSU)  • TH(B6\)
 TDCONC • - TOTCHH / DVOL(ES\)
 IF TDCONC < 0 THEN WWSR7L1
                                                  '  total charge to ba
                                                  •  n«utraiia«d, par  Utra

                                                  '  total cone of countarionc
                                                  '  par litre of diffuae layer

                                                  '  aniona attracted
 •  Coma  to here  if humic* have a net negative charge (eationi attracted)

 HHSR7L2s

 TDCONCCALC - 0

 FOR X\  -  1 TO NSP*
 IF CHfc(Xt)   <  0  THEN CDDI>(BS\,X^)  • 0
 IF CBV(Xtk)   <  0  THEN WWSR7W
 CDDL(HS\,») «  C{»)  *  (RATIO (Hfl1k)*(CH\{»))>
 TDCOHCCAIiC  «  TDCOMCCALC >  (CDDIi(HS»,X\)*CB\(X\) )
                                     '    '
NEXT X*
TDCONCBRR -  (2 •  (TDCONC - TDCONCCALC)  /  (TDCONC +
IF TDCONCERR < CRIT THEN WWSR7L5
 ' Adjust ratio and re-try

 RATIO (HS\) •  ( (RATIO (HS\)  • TDCONC / TDCONCCALC) + RATIO (Hfl*})/2
 GOTO 1W8R7L2               .
 ••••*****•**•**•*******•••****«••«•**••**•***•••**•••««•••••**«*****«•
 • Como to bar* if huaics have a net poaitive charge  (aniona attracted)
TDCONCCALC • 0
FOR X* m 1 TO KSP\
IF CU>((X\)   > 0 THEN CDDXf(HS*,X*)  • 0
IF CKMX*)   > 0 THEN WWSR7M
TDCONCCALC
WWSR7L4I
NEXT X%  ...
               TDCONCCALC +  (CDDL(HS*,X\)*CHMX*) )
.TDCONCERR «  (2 *  (TDCONC - TDCONCCALC)  /  (TDCONC 4 TDCONCCALC) )A2
IF TDCONCERR ,< CRIT THEN MWSR7L5

-------
                         Equilibrium .chemical spetiation by WHAM
                                                                           997
' Adjust' ratio and re-try    .

RATIO (ES*) - ( (RATIO (HS\) * TDCONC / TDCONCCALC) + RATIO(HSX) )/2
GOTO WWSR7L1

WHSR7L5I
RETURN  •        '               .       '
 tMSRBi    * called by »»ln program - call* HWSR9
'ft          maJwB output file
                                                                             It
                                                                             If
DIVIDER? * _
FF « 1.00001

PRINT 12, DIVIDER?
PRINT «2, _
                                         format factor j avoid* tt««0y output
**********
PRINT ff2, DIVIDER?
PRINT §2,
PRINT §2,
PRINT §2,
IF PEFIX?
IF PEFIX?
IF PC02 •
IF PC02 -
PRINT §2,
PRINT 82,
PRINT 82,
PRINT 12,
PRINT 82,
PRINT 82, "INPUT DATA"
PRINT 82,
PRINT 82,
PRINT §2,
PRINT *2,
PRINT 82,
                  OOTPUT FILE FROM HHAM-H  VERSION 1.0
                                                               *.********ii
          •SOURCE FILE ";TAB(25);SF?
          •DATABASE    "|TAB(25);DBS?
          . -NO"  TEEN PRINT 82,"PH";TAB(25);"VARIABLE"
          « "TBS" TEEN PRINT 82,"PB"/TAB(25);"FIXED-
          999     TEEN PRINT 82,"PCO2";TAB(25)j"VARIABLE"
          999     THEN PRINT *2,"PC02"|TAB(25);«FIXED-
          •PRECISION * ";TAB(25);USINO "8.«f AAAA"jFF*PRBCISION
          •STARTING PE ";TAB(25};USING "f.•«****"JFF*FHSTART

          DIVIDER?         •                             '
          "TEMPK-f TAB(25)/USING) "f.fl8****";PF*TEMPK
          -TOT HA-|TAB(25)|USINO ««.t«***A«*FP*ta(i>
          "TOT FA-;TAB(25)fUSINO "f.f«****"|FF*TH(2)
          "PC02"|  TAB(25)|U8INO "i.itfA*AA-;FF»PC02
PRINT 82,
PRINT 82, "MASTER SPECIES"/TAB(25)j"TOTAL CONG"
FOR X* « 1 TO 100
IF XV « 1 TEEN KWSR8L2        S
IF X* • 51 TEEN WNSR8L2
IF T(XX) •'0 TEEN WWSR8L2
PRINT t2,»;TAB(5)|N?(X1t>|TAB(25)|USING "f.t«AAAA"|FF*T(X1ll)
MHSR8L2*                                                 .
NEXT X\
PRINT 12,•                       -
PRINT 82, DIVIDER?
PRINT 82,
PRINT f2, •RESULTS'
PRINT 82.
PRINT 12,
PRINT 82,
PRINT 82,
PRINT «2,
          •NO. OF ITERATIONS";TAB(22)fUSING
          •PB-l               TAB(25)»USIBO •|.t«AAAA"|FF*PB
          "IONIC STRENGTE-J   TAB(25)jUSING •§.•«***»"|FF*IS
          "CHARGE RATIO-|     TAB(25};USIHG «f.||tA*AA">FF*POSCS/HEGCH
PRINT 12, •CHARGE DIFFERENCE-|TAB(24);U8ING "+t.f8lAA*A-»FF*POSCH-NEOCH
,IF TH{1> >
IF TH(2) >
IF TH(1) >
IF TH(2) >
PRINT 82,
PRINT 82,"HATER VOLUMES"
PRINT 82,"FRACTION EA-DDL "
PRINT 82,"FRACTION FA-DDL
           0 TEEN PRINT t2,
           0 TEEN PRINT 82,
           0 TEEN PRINT t2,
           0 TEEN PRINT t2.
•ZBD-HA"|TAB(24)|U8ING «+8.t«AAAA"lFF*ZED(l> .
•ZED-FA-;TAB(24))USINO «+8.f«AAAA"»FF*ZBD(2)
•RATIO-BA*|TAB(2S)7USINO «i.f»|AAAA»;FF*RATIO(l)
-RATIO-FA«|TAB(25)jUSINO -8.i88AAAA";FF*RATIO(2)
                              I TAB (251 BUSING "«.-t8lAAAA";FF*DVOL(D
                              ;TAB(25)BUSING ••.«!**A*"jFF*DYOL(2)
 PRINT #2,"FRACTION SOLUTION «;TAB(25)|USINC •«.««J""«;FF*VOLSOL
 PRINT 82,                                 i
 PRINT 82, "CARBONATE ALKALINITY  ";TAB(24)>USING "•»8.«8tAAAA<>}FF*_
                           (C(101) + (2*C(55))  + C(51) - C(D) * VOLSOL

-------
998
                                   E. TIPPING
PRINT «2,
PRINT #2, DIVIDER?
PRINT *2,
PRINT #2, "FRACTIONAL DISTRIBUTIONS OF MASTER SPECIES"
PRINT #2,                  '.'•.'•
PRINT «, "MASTER SPECIES";TAB{23);"S"|TAB(31);«DHA"»TAB(40);«DPA"j
                                         TAB(49);"HA»;TAB(S8);"FA" ~
PRINT #2,                                          ,
FOR XV - 1 TO 100
IF XV - 1 THEN MMSR8L3
IF XV • 51 THEN HWSR8L3
IF T(XV) « 0 THEN WWSR8L3
COSUB «WSR9                                   .
PRINT *2,XV;TAB(7);N$(XV);TAB{21);USINO "t.«f";FRACS;TAB(301/FRACDHAj
            TAB(39);FRACDFA;TAB(48);FRACHA;TAB{57);FRACFA
WWSR8L3I
NEXT XV
PRINT t2,
PRINT *2,
PRINT #2,
PRINT «2,
PRINT 12,
PRINT #2,
          DIVIDER*

          •CONCENTRATIONS AND ACTIVITIES"
          "VALUES OF Ha (KOL BOUND / O HUKIC SUBSTANCES}-
          "NO IS SPECIFIC BINDING f DNU IS DIFFUSE LATER"
          TAB(25)|"TOTAL MATER";_
          TAB{54);"SOLUTION PHASE"
PRINT ff2,"SPECIES";TAB(22)I•[FREE]«;TAB(33);"[OROANIC]";
                   TAB(S2) « «CONC<>;TAB(e3) | •ACTIVITY-
PRINT #2,
TOR XV « 1 TO NSPV
IF C(XV) - 0 THEN VWSR8L4
PRINT 82,XV;TAB(7};N$(XV);TAB(21);USINO "§.««t****";FF«VOLSOL*C(XV};TAB(33)j
FF*((DVOL(1)*CDDL(1,XV)) + (DVOL(2)*CDDL(2,XV) > + CHC(1,XV)
.  TAB(50);FF*C(XV)fTAB(62>jFF*A(XV}
VWSRBL4*
NEXT XV  '.           '•
PRINT §2,
PRINT *2, DIVIDER*
PRINT *2,
PRINT «2,
PRINT t2,
PRINT 12,
PRINT f2, TAB(22)i«DNU-HA-;TAB(34};"DNU-FA"|TAB(46)|"NU-HA";TAB(58)}"NU-FA"
PRiffj: v2y             •
FOR XV « 1 TO NSPV
IF TH{1) « 0 THEN WHSR8L5
DNU1 m DVOL(1)*CDDL(1/XV)/TH(1)
WWSRBLBf                                   .
IF TH(2) « 0 THEN HHSR8L6
DND2 • DVOL(2)*CDDL(2,XV)/TH(2)                   '           '      ,
KWSR8L€»
IF DOT1 « 0 AND DN02 « 0 AND NU(1,XV) - 0 AND NU(2,XV)  . 0 THEN WWSR8L7
PRINT i2,XV;TAB(7);N$(XV);TAB(21);USINO "*.«|***»«;FF*DNUl;TAB(33)i
                FF*DN02;TAB(45);FF«NU{1,XV);TAB(57);FF«NU(2,XV)

WWSRBL7«           .          '
NEXT XV
PRINT i2,                                  N
PRINT 92, DIVIDER*                        •

RETURN.        •     ' ;                          '                    > .
'™sS"Si"*!S2"?™*w5S^^
'**          ealcs distribution of chosen master eaaciaB                     ««
TOTALS   * 0
TOTALDKA - 0
TOTALDFA •> 0
TOTALKA  B 0
TOTALFA  •> 0

-------
                         Equilibrium chemical speciation by WHAM
                                                                     999
FOR Y* - 1 TO NSP*
IF M1MYX)
IF HIM**)
IF MIMY*)
IF M1MW
IF HIM**)

IF M2MYX)
IF K2MY*)
IF M2*(Yfc)
IF H2M**)
IF H2V(Y\)

IF M3*{y*>
IF H3*(Y*)
IF M3*(Y*}
IF K3MY*)
IF H3MYM

NEXT Y\
       X* THEN TOTALS
       XV THEN TOTALDHA
       X* THEN TOTALDFA
       X* THEN TOTALKA
       X* THEN TOTALPA

       X* THEN TOTALS
       X* THEN TOTALDHA
       Xfc THEN TOTALDFA
       X* THEN TOTALS*.
       X* THEN TOTALFA

       X* THEN TOTALS
       X* THEN TOTALDHA
       XV THEN TOTALDFA
       X* THEN TOTALHX
       X% THEN TOTALFA
FRAGS   * TOTALS   / TCALC(XX)
FRACDHA « TOTALDHA / TCALC(X*)
FRACDPA « TOTALDFA / TCALC(X*}
FRACHA  - TOTALHA  / TCALC(X*)
FRACFA  - TOTALFA  / TCALC(X*>

RETURN
TOTALS   4
TOTALDHA 4
TOTALDFA 4
TOTALHA  4
TOTALFA  4

TOTALS   4
TOTALDHA 4
TOTALDFA 4
TOTALHA  4
TOTALFA  4
(VOLSOL
(DVOL(l)
(DVOL(2)
(CHC(l,Y«t)}
(CHC(2,YK)

(VOLSOL
(DVOL(l)
(DVOL(2)
(CHC(1,TT\)
(CHC(2,Y%)
TOTALS   4 (VOLSOL
TOTALDHA 4 (DVOL(l)
TOTALDFA 4 (DVOL(2)
TOTALHA  4 (CHC(1,YX)
TOTALFA  4 (CHC(2,YX)
                        C(YX)
                        CODL(l(y%)
£l»t(YX)}
                        CDDL(1,Y%) *
                        CDDL(2,TX) *
                        CDDL(2,Y%)
                                 APPENDIX 3
                               WHAM'S Program Uatag

•«ffftfiftiffKftiftlft«ft»fffftftffft»ifftfHff»t***ftttt*«ft*«*ftttft«ff«iftC«H«ttff*lffft*»
'« WHAM-S  VERSION 1.0                ~                                      ft
'iff Computar coda in TDRBO BASIC                               'ft
'•I Sp«ci»tion of humlc-rleh ••din«nt> and coil*                             tt
••• This veriion producod on 13 April 1993 by B.Tipping, IFE, Wind«r»»r»    . ••
'•ffffffftffffftffffffttftlftffftffft«ff*ft»ff«fH»««t»ffffff*ffHtMifft»*ti*tff«*ft*Hfftfft*li»ifft
CLS

PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
•***•*••*•**«*••*****••
"«•   W ind«rn«r«   **
"•*   H umio        ••
•**   A quaottB      ••
»**   H odol        *•
•**   «.             **
"**   S oil*        ••
«*•     •diaenti    ••
*** •           •     **
«**   Version 1.0   *.*
•***•*******«**«*••**•
   Equilibrium apaciation of aoilB/aadimanta"
   Hamiea (FA and/or HA) praaant"
   Variabla pH"
   Fixad or Tariabla pCO2a
   Precipitation of Al(OH)3 t F«(OH)3«
   Clay oation-axchangar*
   Salaeti-vlty for ion-aacehangaa
   PA adaoxptioa/da«orptioBM •

   E.Tipping  IFB  April 1993"
PRINT:INPUT "NAME OF SOURCE FILE  [OMIT
PRINT

OPEN SF$ 4 ".DAT* FOR INPUT AS «1

INPUT #1,S$,S$   /
INPUT #1,S$,DBS$
INPUT ftl,S$,NSP\
                                  .DAT TRAILER]  t ",EF$
                                             fila baadar
                                             database  identifier
                                             array oize required

-------
 1000
                                    E. TIPPING
 ' DIHension and clear arrays, than read in chemical constant*

 GOSUB WSSR1   .
 I******************************
 ' Continue to read in data file

INPUT f1,8$,PRECISION
INPUT tl,S$,TBHPK
INPUT tl,S$,C60LID
INPUT fl,S$,TH(l),TH(2)
INPUT tl,8$,TCIAY
INPUT ftl,8$,FAAQST
INPUT ftl,8$,PCO2
INPUT fl,S$,NCOKP*

FOR I* - 1 TO NCOMPV
INPUT S1,8$,X*,TCONC
N$(XK) .8$
T(») • TCONC
NEXT I\

CLOSE *1
                                                   precision  -
                                                   temp K
                                                   •oil cone g/1
                                                   total' cone*  (8A,FA)
                                                   total clay cone  g/1
                                                   [FA] in aqueou* phase
                                                   PCO2
                                                   no. off aaxtar >p«ci«c
                                                   total cone*
»*****•******••••*•***
' Porfota calculation*

GOSUB WSSR2

IF FA&QST * 999 CHEN GOSUB W88R9
IF F&AQ8T < 999 THEN FA&Q • FA&Q8T
                                                 '• calculatA FA aorption
                                                 •' *q IFA] providad
                                                 ' *ound output
 to***************************************************
 ' Calculations finished j signal and make output

 SOUND 400,5   s   SOUND 500,5   t SOUND 700,8

 OPEN SF$ •+ ".OUT-  FOR OUTPUT AS 12

 GOSUB NSSK10

 CLOSE f2                 •

/PRINT ,                	
 PRINT •CALCULATION FINISHED  t CONSULT  *.*• •;SF$i".OUT ** FOR DETAILED OUTPUT*

 STOP
 END   .               .                              .
'«Bfftff«ff*fiitiHfHCtf*tilt»l»tftMttt*itf«»«i»ffMtftiMHfftiti*lf«fft*milftt
 WSSRlj    • call«d fey main program                   '                       «•
'iff          DIHanaion* array*                                               tl
'•ft          road* in data fron data ba*a                                    ft
'it          cat* up Kodal V                                                 «
'fti«llft«»fti«»»tMlftl»«»ltif»»tf««»fftff«i»i«*ff*«liff||fftiffifft»»ftft«t»ftft

* DlKonsion array* ; inorganic chaaiitry   ' '
DIM NMNsp*)
DIM N$(NSP\)
DIM T(NSP\)
DIM TCALCfNSP%)
DIM A(NSP\)
DIM C(NSP*)
DIM CH^(HSPX)
DIM M1\(NSP%)
DIM K2\(NSP\)
DIM M3%(NSP\)
                               nuBorical identification of  specie*
                               nominal  identification of specie*
                               total  concentration* of components
                               total  calculated concentration* of component*
                               activities of solution species
                               concentations of solution specie*
                               charge*  on solution species
                               master specie* identifier*
                               master species identifiers
                               master species identifiers

-------
                         Equilibrium chemical speciation by WHAM
                                                                          1001
DIM SlMNSFfc)
DIM 52MNSP*)
DIM S3*(NSP\)
DIM LK(NSP*)
DIM DH(NSP*)
DIM GAMMAU)
                               stoichio&etries
                               stoichioBetries
                               stoiehionetrieB
                               equilibrium constants
                               enthalpies
                               activity coefficients
' DXHension humic arraya ; 1 • HA, 2 • FA
BUS SF{2,8)
DIM 6H<2,12)
DIM KH<2,8)
DIM FKH(2,8)
Dm KMH(2,8,NSP*)
DIM SITE\(12,2>
DIM FXKSA<2,NSP*)
DXH PiatHB(2,NSP«S)
DIM BXTHBTA(12,NSP%)
DXH HONTHETA{8,NSP*}
DIM FP<8)
DIM BXBTERH(NSP*)
DIM MONBTERM(NSP»0
DIM BXSNU(12,NSP*)
DIM KONSNU{6,NSP*)
DXH BXNU(NSP*)
DXH HONNU(NSP*)
DXH NU(2,NSP*)
DXH CHC(2,NSP*)
DXH CDDM2,HSP%)
DXH BtZ{12)
DXH KONZ<8)
DXH BINDTESTMNSPM
DXH KSBLHS(NSP*)
DXH FA(10)
                               proton-binding sites not forming bidantata vitas
                               bidantata matal-bindings  •
                               K's for proton-binding              '  •
                               PK'c for proton-binding
                               K's for natal-proton axchanga
                               proton sitas aakihg bidantata sitas
                               natal-proton aicchanga constants
                               natal-proton axebanga constants
                               tbata's for bidantata sitas
                               tbata's for nonodantata sitas
                               dissociation factors for proton-binding
                               tarns calc'd in finding bidantata tbata's
                               tarns calc'd in finding nonodantata thata's
                               values of Ha at aach bidantata slta
                               valuas of NU at aach nonodantata sita
                               ovarall KO for aach natal/bidantata sitas*
                               ovarall HO for aach matal/monodantata sites
                               ovarall MO for aach conplaxad species
                               cone of conplazad species par litre
                               cones in HS DDL's
                               net charge at each bidantata site
                               net charge at each nonodantata site.
                               allows binding of a species to be tested
                               selectivity coefficients for HA/FA DDL's
                               FA fractions for sorption nodal
* DXHansion bu&ic arrays for clay ninaral ion-axchangar
DIM KOAY(HSP*)
DXH OATtDDL(HSP\)
                             ' selectivity coefficient
                             ' cone of species in clay DDL, per litre
'Input of constants for inorganic .speciation and BS complexation

OPEN DBS? 4- ".DBS" FOR INPUT AS 12

                                                 ' database  identifier

                                                 ' HA properties
XHPUT 92, B^
INPUT ft2,S(,NCOOH(l),P13lMl),PKHB(l>,DPKHA(l),_
      DPKHB(l),F(l),FPR(l),RADiaS(l>,KO&HT(l) '
INPUT ft2,S$,NCOOH(2>,PKHM2),PlCHB(2),DPKHM2),_
      DPKHB(2>,P(2),FPR<2),RADIUS(2},HOI»T(2)

INPUT t2^fi$^DLF   '   •
INPUT t2,S$,USOAL25,DBAL
INPUT *2,S$,UCSOFE25,DHF8
INPUT «2,S$,KZED
INPUT *2,S$,ti*MMA,KO,BETA
INPUT i2,S$,NODATA\
FOR I* «  1 TO NODATA*
INPUT i2,XX,N$ (Xlt) ,CHX(X\) ,
                                                   FA properties
                                                   clay properties
                                                   double layer overlap factor
                                                   A1(OB)3 ppt
                                                   Fe(OB)3 ppt
                                                   DL vol at low Z
                                                   FA sorption constants
                                                   number of species
                                  l,H2V(rH)f
                                  •*»» I A^S 11 un (   	
         PKMHA(l,X%),PKHHM2,XX)«l&ELRS(X\),KCIAY(X'k)

IF PXKBA(l.Xk) < 999 THEN BXNDTEST\(X«() • 1

NEXT X*

CLOSE #2
                                                  ' identifies species that
                                                  ' don't bind to HS

-------
 1002
                                   E. TIPPING
 »••***••*•••»•*•*•*•**•*•******•**•*
 '  Sat derived constant* for Modal V

 FOR XX • 1 TO 2
PKH(XX,1)
PKH(XX,2)
PKH(XX,3)
PKH(XX,4)
PKH(XX,5)
PKH(XX,C)
PKH(XX,7)
PKH(XX,8).
PKHMXX) - 
PKHMXX) +  (1 - FPR(2)) * NCOOB(2) / 4
NEXT IX

FOR IX - 5 TO 8
SP(1,XX) « (1. - FPR(l)} * RCOOB(l) / B
SP(2,IX) « {1 - FPR(2)) * NCOOH(2> / 8
NEXT IX

FOR XX • 1 TO 12
5H(1,XX) • FPR{1) • NCOOH(l) / 16
SM(2,IX) • FPR(2) * NCOOH{2) / 1C
NEXT XX        -
' Set metal-proton .exchange conttanti for individual «it««

FOR IX - 1 TO NSPX
IF N((XX)      * "»  THEN WS8R1L1
IF PXHHM1,XX) m 999 THEN H8SR1X.1

FOR JX • 1 TO 4
PKMH         - PKKHA(1,IX)
                                                  '  specie* not defined
PKKH         - PKMHA(2,XX)
KMH(2,JX,XX) •> 10*(-PKMH> .
NEXT JX

-------
                          Equilibrium chemical xpecUlion by WHAM
                                                               1003
 FOR Jit - 5 TO 8
 PKHH         - (3*PKHHA(1,IX)} - 3
 PXHH         m 3.96*PKHHA(2,I\)
 KKH(2,J%,I«) - 10M-PKMB)
 NEXT J%
 GOTO WSSR1L2

 * If cone to hero, than no binding of species XX can occur
 WSSR1L1*                                        •...".

 FOR JX - 1 TO 8
 KKB(1,3X,IX) • 0                          /   -
 KMH(2,JX,IX) « 0
 NEXT «JX

 WSSR1L21                  .
 NEXT IX ,               .               -            :

 RETURN
                                     ' 3*A - 3 conversion ; HA

                                     * 3.96*A  conversion ; FA
 HSSR2s
'Iff
'ft
'ffff
'«
'••
'•ff
called from main program - call* WS8R3 and HSSR4
Beta initial trial value* of aaeter 0peci«a
initialises all activities and concns by calling WSSR4
control! level of precision               '
calls WSSR3 to eale speciation
checks for pptn of Al(OH)? and Fe(OB>3
«ets up screen to report on progress of calculation
                                                                              f §
                                                                              it
                                                                              ft
                                                                              ft
                                                                              §§
                                                                              f§
                                                                              ft
• Set initial trial value*

XS « 0.1

ZED(l) - -1B-4
ZED(2) • -1E-4

RATIO(l)  - 10
RXTZO(2)  -10
RXTIOCLXY « 10

     - 10* (-4 J
FOR XX • 2 TO SO
A (XV) - 1E-6 «
NEXT »
A<52) - T(S2»
A(S3) - T(53)
ACS4) « T<54)
A(56) - T(5€) * 1B-4
A(57J « *(S7) « IB-IS

IF PCO2 - 999 THEN A(S5) m 1E-6  * T(S5)
                                    ' ionic strength

                                    ' Z for HA
                                    ' Z for FA

                                    ' RATIO for HA
                                    ' RATIO for FA
                                    ' RATIO for clay

                                    ' starting.K+ activity

                                    * species il is H
                                    ' cationic master species
                                    • Cl
                                    ' N03
                                    ' S04
                                    ' F
                                    ' P04

                                    ' C03.2-
' Summary output to screen, and headers

LOCATE 14          _                                    .
PRINT "
LOCATE 14          '
PRINT *  SOURCE FILE J  ";SF$
PRINT -PRECISION (X) t  -fVSIKO «#.«***»«; 1.0001*PRECISION
IF PC02 m 999              THEN PRINT  «   *      PC02  I VARIABLE-
IP PC02 < 999              THEN PRINT  «         PCO2  j FIXED"

-------
 1004
                                    E TIFFING
PRINT
PRINT «ITER";TAB<12);"PH";TAB(24);-IS«;TAB{36},"CHRATIO-;TABU8};
                        «PE PPT"JTAB<60};-AL PPT«;TAB(72)j-ADJEXP"
 »•**•»•**»***»******
 ' Begin calculations

NOHITt « 0

ALPPTS - "HO"
FEPPT5 • "HO"

CRIT • 1B-4 I COSUB HS8R4

CHIT - (PRECISION/100)»2 s COSUB WSSR3
                                                 ' iteration counter

                                                 ' first speciate with
                                                 ' no pptn axiom*

                                                 ' initialization

                                                 ' full speoiation
«***•*«***•**••***************•********•***•*•*•***             , ,•
• Check for precipitation of Fa {OB) 3 and/or A1(OH)3

UCSOFE — LKSOFE25 * (0.219 * UHFE • ({1/298} - {1/TEMPK))}
IF t(il) .« 0 THEN WS8R2L1
fclAPFB « L0010(JL{11)) + (3*FH)
IF LIAPFE >m IKBQtt THEN FEPFT< • "TTBS-
IF LIXPFB >m LKSOFE THEN COSUB WSSR3             • f«calculat« with Fa (OH) 3

WSSR2I.lt

LKSOAL . 1KSOM.2S + (0.219 • DHAL * ((1/298) - (1/TEKPK111
IF T(5) - 0 THEN WSSR2L2                       »/*«wr«.|||
LIAPAL • LOC10(A(S)) + (3«PH)
IF LIXPAJ. >- LTSOAL THEN ALPPT$ - "YES-
                    **** °°SDB WSSRS             ' recaloulat. with A1(OH>3
RBTORH
 KSSR3t
'Iff
'•ft
'Iff
•tf
             callad by M8SR3 - call* WS8R4               .
             control* iaprovoaant of activitiac and ZFA,ZRA.
             call* W88R4 to do mai* and eharga calculation*
             t«>t> for nan and charge balaneai
             tasta for eonvargane* of ZFA, ZHA, ZCIAT
             raportB progran of calculation to icraan
                                                                             It
                                                                             ••
                                                                             ••
                                                                             ft
                                                                             it
                                                                             ••
' Begin iterative cycle t return to here until gat convergence

HSSR3Llt   '

       • NUMIT* 41                              • update iteration counter
' Calculate ionic strength  .

IS • 0
                                                 f preparing to turn
FOR x*- 1 TO NflP\                                     -
IF C(3CTt) > 0 THEN IS . IB +  (O.S •  (C(») * CHV{») * CBfc(X*}))
NEXT Xk
IF IS > 100 THEN IS • 100
                                                 ' avoids initial high IS
' Adjustment tern ADJEXP; decreaaes with NOMIT* 6 with ionic  strength
IADJEXP • ADJEXP
ADJT1 « SQR(IS)
ADJT2 » NOMIT* / (3E4
ADJEXP » O.S - ADJT2
                                s IF LXDJEXP • 0 THEN IJUJJEXP  •  0.5
                                s IF ADJT1  < 0.03 THEN ADJT1 - 0.03
                        ADJT1)  J IF ADJT2  > .0.48 THEN ADOT2 « 0.48
                           .     t IF ADJEXP > IADJEXP THEN ADJEXP •  IAOOEXP

-------
                         Equilibrium chemical ipcciation by WHAM
                                                  1005
 ,«•*«••••«*••••«•***«**•••*****•*«***•****••••*•••••*•*•
 '  Improve A(l). (K+ activity)  - begin after 10 iterations

 A1PACTOR m (POSCH / ((POSCH + NEGCH)/2))

 IP HOMIT* > 10 THEN A{1)  - A<1) / (A1FACTOR*ADJEXP)

 PH - -LOGIC (A(l))
 «*********••*««•**•••*******
 *  Improve catlona and aniona

 FOR X* « 2 TO 50
 ZF T(X*} • 0 THEN MSSR3L2

 IP » « 5 AND ALPPT*  - "YBS"_
 THEN A(5) - (10MLKSOAL))  *  (A(l)*3)
 IF X* • 5 AND ALPPT$  • •YES" THEN WSSR3L2
 IF X* • 11 AND FEPPT$ * «TES-_
 THEHA(U)  - C10*(LKSOPE)) « (A(l}*3)
 ZF X* « 11 AND FEPFT* • "YES" THEN W8SR3L2

 COHCFACXOR • T (»t)/TCALC (XX)
A(X\)  " A(X%)  * COHCFACTOR / (CHRATIOAADJEXF)
WS8R3L2i
«2XT X*
                          cationi.
                        ' chack for Al(OH)3 pptn   .
                        ' Al controllad by A1(OH)3

                        ' chack for F«(oa)3 pptn
                        ' Fa controllad by Fa(OH)3
FOR X* • S2 TO 100
ZF T(») « 0 THEN WSSR3L3
CONCPACTOR . T(X%)/TCAI«(X%)
A(X«t) • A(3Pt) * CONCPACTOR
WSSR3L3I
NEXT X%
                          aniona
<*«•*********•***•**
* I»prov« ZED valna*
IF TH(1) > 0 THEN 2ED(1) • ZBD(l) 4  ((ZCALC(l)-ZED(l))/30)
ZF TH(2) > 0 THEN ZBD{2) » ZED(2) +  «ZCALC(2)-ZED(2})/30)
*************•••*•*•********•*•*••**••«••«
* Calculate total concns of mastar apacia*  '
OOSOB HSSR4
**•*«*•**••••«*********«•••••****
' Output currant ntatue to acraan

ZF FEPPT$ « "NO"  THEN FEPPrt « 0
ZP FEPPT$ « "YES" THEN FEPPTH • 1
ZF ALPPT$ « "NO"  THEN ALPPT* • 0
IF ALPPT$ « "YES- THEN ALPPTk « 1
LOCATE 19
PRINT
O "S«.«B«-;PH;TAB(22);OSINO «t.lll***";IS;TAB(35);_
  Tast cations, anions, charge-balance and Z'e for convergence
FOR XV - 2 TO. SO
IF X* « 5  AND ALPPT$ m "YES" THEN WSSR3L4
IF X* « 11 AND FEPPT$ - "YES" THEN WSSR3L4
IF T(X\) - 0 THEN HSSR3L4
CONC8RR ~ <2*
-------
 1006
                                    E TIPPING
 IF CONCERR > CHIT THEN HSSR3L1
 WSSR3L4S
 NEXT X%
                                                    re-iterate
 FOR X% « 52 TO 100
 IF T(»)  - 0 THEN WSSR3LS
 CONCERR - (2*tT(X%)  - TCALC(X*))/(T(X*)  4 TCALC(X*)})*2
 IF CONCERR > CRIT THEN WSSR3L1
 WSSR3L5J
 NEXT X%             .
                                                  ' anions


                                                  ' xe-iterate
IF  (CHRATIO -  1)*2  > CRZT TEEN WSSR3L1
IF ZERR(l) > GRIT THEN WS8R3L1
                                                 ' re-itezate


                                                 ' re-iterate
IF ZBRR(2) > CRIT THEN HSSR3L1
                                                 ' ra-it«rat«
RETORN
 HSSR4s
'fft
•I*
'*!
 ' called by KSSR2 and WSSR3 - calla WSSRS
   calls WSSRS to calc actlvltlai, cones off complaxoa and'
        anounti bound by FA, BA and CUT
   BUM to get total eale/d eoneni         .         ,
   calca +v« and -v« charge and:CBRATXO
                                                                             0§
                                                                             8§
                                                                             fit
' Cale act* ft conena of Inorganic complex**, amnt* bound to FA, HA, clay '
GOSDB WSSRS                   .                                  ,

'•*•***••«•••**••*•*••••**•***«*•**•«•••*•*•**••••••««••••*••*••••*
' Do suamationa to obtain total calculated eoncns of naatar specie*   •

FOR X* - 1 TO 100                      ,
IF A(X*> * 0  THEN TCALC(XV) - 0
IF A(X\) « 0  THEN WSSR4L1
IF Xt    « 1  THEN WSSR4L1
IF X*    m 51 THEN HSSR4L1
TCAIiC(X%) « 0

FOR VK « 1 TO NSP\
IF K1MY*)
           • X* THEN
            (VOLSOL
            (CHC(1,Y*)
            (CHC(2,YM
            (DVOL(l)
            (DVOL(2)
            (DVOLCLAY
 IF M2MTOO
          4
  • X* THEN
  (VOLSOL
4 (CHC(l,n()
4 (CHC(2,Y%)
4 (DVOL(l)
  (DVOL(2)
          4  _
            *
                         CDDL(1,Y%)  .* 82MYM).
               CL&YDDLCVIt)
                                            ' preparing to SUB
                                            '" solution
                                            ' HS bound (1)
                                            ' HS bound (2)
                                            ' DDL (1)
                                            ' DDL (2)
                                            ' DDL clay
                                                      ' solution
                                                      ' HS bound CD
                                                      ' KS bound (2)
                                                      ' DDL (1)
                                                      ' DDL |2)
                                                      ' DDL clay

-------
                          Equilibrium chemical cpeciation by WHAM
                                                                            1007
IF M3*(**)
        ""4
         .4
         4
         4
NEXT Y%

HSSR4L1I
NEXT X*
             « X* THEN TCALC(XX)
             (VOLSOL
             (CHC(l,y\)
             (DVOIt(l)
             (DVOli(2)
             (DVOLCIAY
                                    TCJVLC(X*)
                                      *
                         CDDL(1,T\)
                         CDDL(2,Y*)
                         CIAYDDL(YX)
                                    * 83M**))
                                    * S3M**))
solution
HS bound
HS bound
DDL (1)
DDL (2)
DDL cloy
                                                                  fl)
                                                                  (2)
*'Calculate 4ve and -v« charges and CHRATIO


POSCH » 0    I  NEGCH • 0

FOR XX - 1 TO NSP\
IF CH*(X\) > 0 THEN POSCH « POSCH 4  (C(») * CHX(XM)
IF CHH(X\) < 0 THEN KEGCH * MBOCH -  (C(X\) * CH\(XX))
NEXT X*

CHRATIO - (POSCH/NSCCH)



RETXnU)
                                                       ' preparing to aum
            called by H88R4 - call* HSSR6, HSSR7 and H88R8
            calc« OH- activity, act coeffi, CO3;2> if pcO2  fixed
            calcB activitiev, conc» of inorganic ccnaplexea
            call* WSSR6 to get m't and Z't for Bfl
            call* WSSR7 to get binding by DL accumulation  (FA,H&)
            ealli W88R8 to get binding by Mi accumulation  (CZAT)
      - - 14 4  (2935*  (0.003354  -  (1/TEHPK)))
A(51) - 10*(LKW>
 WSSRSi
'«»
•••
*«f
•iff
'ffff
'ffffftl«ffff«ftR«tlftffltlilfilCII»HHItttlHfftt»ff*«S»ftflHfftiHftWfffftlH«tlii*H

• calculate activity of OH- from A(l) , temp, deltaH
                                                                              It
                                                                              ti
                                                                              if
                                                                              ft
                                                                              if
  Calculate activity coefficient* u*ing extended Debye-Ruclcel
GAMMA(O)
ACTA
ACTS
ACTC
ACTD
CAMHA(l)
CAMHA(2)
CAKHAO)
CAKHA(4)
1
0.270 4
0.330
- ACTA
ACTB
10* (
10* ( 4
10* (9
.10* (IS

(8E-4*TEMPK)

SQR (IS)
8QR (18)
ACTC/d 4(3
ACTC/(1 ,4 ( C
ACTC/ (1 4 f S
ACTC/U * (12





* ACTD)))
• ACTD)))
* ACTD)))
• ACTD)))
                                                         const A. in D-B
                                                         const B in D-B'
                                                         act coeff M4/-
                                                         act coeff K24/-
                                                         act coeff K34/-
                                                         act coeff M44/-
* Calculate concentration*  of inorganic species

' First 'do the  carbonate  system

IF PCO2 « 999 THEN HS8R5L1

ENTHTERM • 220  *  (-0.962) * (0.003354 - (1/TEUPK))
A(5S) -" PC02 /  10*(18.149 4 ENTHTERM) / A(l)  / A(l)
                                                       ' pCO2 not specified
                                                         activity of C032-
CACCOMA-M

-------
  1008                            ,   E.TIPHKO

  WSSRSLlt

  FOR X* - 101 TO NSP*

  IF S1MX*)  • 0 THEN WSSR5L2

  IF S1MX*)  > 0 AND A(M1*(XM)  « 0 THEN KS8RSL2
  IF S2*(X*}  > 0 AND A(H2*(X*n  • 0 THEN H88R5X.2
  IF S3*(X%)  > 0 AND A(H3*(X*))  • 0 THEN NS8R5L2

  BNTHTERM -  220 * DB(X*) *  (0.003354 -  (1/TEKFK))

  LOGACT  • LK{») +  ENTHTERM .
  LOCACT  - LOCACT 4  (S1*(X*}  * LOO10 (A(K1*(X\) 1))
  tOQACT  - LOQACT +  (S2\(IX)  * LOQ10(A(H2\(XX))))

  IF S3\{»)  > 0 TOB1L
  LOGACT  - LOCACT +  (S3*(X*}  • LOO10(A(M3H(H))))

 A(Z%) » 10*(LOGACT)
                                      /       ,
 WSSR5L2*
 NEXT X*
                                                       ' co&pl«x not dof inad

                                                       ' no calculation if
                                                       ' contributing ipttei«B
                                                       ' abiant
 ' Calculattt concne from activltia*

 FOR X* • 1 TO NSP%
 CHARGED « ABS(CH*(X*))
 C(XV}   - A(X\) / GAIOIA{CBARGB%)
 NEXT X%
* Calculate concentration* of bunic-bound


FOR HS* m 1 TO 2

IF TB(HS%) '• 0 THEN WSSR5W

GOSOB HSSR6
                                           component*
                                                       ' no calcn if HS absont

                                                               ZCALC'0 and KO'm
' calculate ooncn. of .pacifically bound .p.cle. per litr. from NO and [HS]


                         » TH(H5\)              .
FOR X* « 1 TO NSP%
CHC(HS%,X\)  .
NEXT »
WSSRSI»3»
NEXT B8%
 • calculate aaxiaum volunai of HA and FA diffu.e layer.

FOR HS% • 1 TO 2
IF TH(HS\) • 0 THEN HSSR5M
                           +  0.04B-10 / SQR(ISJ)
             « (VTERH1*3) - { (RADIUS 
-------
                         Equilibrium chemical speciation by WHAM
                         1009
HSSRSL4S
NEXT HS*


' Calculate maximum volume of clay mineral diffuse layer

ZF TCLAY - 0 THEN WSSR5L5            "/    '

DVDLMAXCLAY • 1E3 * TCIAY  * SACIAY *  (3.04E-10/SQR{IS))     '  litres/litre

WSSRSXiSt

* Calculate  the actual diffuse layer volumes, ueing DLF

DENOM - 1  +  ((DVOLMAX(l) +  DVOUOX(2) + DVOLHAXdAY) / DLP)

FOR HS* -  1  TO 2
DVOL(HS*)  -  DVOLMAX(HS\) /  DENOH
NEXT  HS*            ,
OVOLCtAX -
                       / DBNOH
VDLSOL - 1 - DVOL(l) - OVOL(2) - DVDLCTAY

FOR HSX « 1 TO 2
IP TH(H5*) > 0 THEN COSOB WS8R7
NEXT HS\

IF TCIAY > 0 THEN GOSUB WSSR8

RETORN  '                                :
'  calculate HS DDL concns


'  calculate clay DDL concn
 KSSRCt  •  ' called by WSSR5                     v                            ft
'••          calcc NO (specific binding) and Z for FA, HX            .       if
•fffftftfffitfffffffflfftffffffffffffffffifffffffffffffffffffffffffffffffffffffffft
W * P(HS*) * LOClO(Zfi)

FOR 1% • 1 TO 8
TEMPVM. • KH(HS\,I\)*EXP(2*W*ZED EXP(TEMPVAL)
TEMPVAIi     - XUH(H8\,SYTEl
BIBTERM(Kt) - TEMPVAL*PP(fiYTElXJ*FP(SYTE21«)/(A(l)*2)
STMBITERM   - SOMBITERK + BIBTERH(Kfc) '
WSSR6L1:
NEXT K*    -   ' .   v  .
SUMBITHETA -  0
BIMETCH    *t  0
' preparing to sum
' theta'o and charges

-------
 1010
                                    E TIFFING
 FOR K* • 1 TO NSPV
 IF BINDTESTMK*) - 0 THEN HSSR6L2
 BITHETA          - BIBTBRK{K*)/SOMBITERM
 SUMBITBETA       • EUMBITHBTA + BZTUKTA
 8ZSN0(d*,X*}     • BITHETA*SM(HSX,J*)
 BIMBTCH          • B1MBTCH + (BXTHETA*CB*(X*})
 KSSR6L2*
 NEXT 1C*
 PROT2CH  « 2*FP(SYTB1*)*FP(SYTE2*) • (1 - BUKBITHETA)
 PROT1CH1 . FP(SCTR1*)*{1 - FP(fiYTE2\))  • (1 - SUMBITHBTA)
 PROT1CH2 - FP(SWE2%)*(1 - FPt *ita 1
                                                              H* bound *t .it. 2
 BIKBTCH
           BZHETCB + PROT2CB 4 PROT1CB1
           BINETCH * 6H(HS\,JM
                                        + PROT1CH2 - 2
                                                              net
                                                              thlf
 NEXT Sflt
 «  Now calculate total amount* bound,  and total nat chaxga (bidantata

 FOR K* m 1 JO KSPV
 IF BIHDTEST*(K*) - 0 THEN WSSR6L3
FOR SJ* • 1 K> 12
BI«CT(K»|  • BZHD(ICH)  f
NEXT A

HSSR6li3t
NEXT K*       •
BIZCALC m  0
FOR OX . 1 TO 12
BIZCAU: •  BIZCALC 4 BIZ(JX)
NEXT J%
 '*••**«****«*•*••*•*•**•••**•*
 ' Binding at nonodantata aitai

FOR J\ - 1 TO 6

STOHONTERM • 1
FOR K* • 1 to NSF*
IF BINDTEfiTV(KX) • 0 THEN WSSR€M
             • 2*W*ZED(BS*)*U - CBX(K%})
             • EXP(TKHFVAtt)
TEUFVAL
TBUFVAti
TEUFVAL
MOHBTBRH(Kfc)
SOMHONTERM
KSSR6L4I
NEXT K\
             • TEMFVA1.*FP(J\)/A(1)
             • SUHMONTERM 4 KONBTERH(R\)
SOMHONTHBTA • 0'
MONMETCH    • 0                        *

FOR K* - 1 TO NSP*
IF BINDTB8TV(K%} « 0 THEN HSSRfiLS
MONTHETA      • MONBTERM(K%)/fiOMMONTERM
SUMMONTHETA   « SOMMONTHETA + MONTHETA
KON8Na(J\,K«(} - MONTHETA*SP(HS\,0\J
HONHETCH     ,« KONMBTCH + (KOKTBETA*CR\(K%))
WSSR6LS:
NEXT K%
                                                  ' praparing to BUB
                                                  ' thata'« and chargai

-------
                         equilibrium chemical spedalion by WHAM
                         10! I
 PROT1CH  - FP(J%) • (1 - BOMMONTHBTA.)     >        • H+ bound
 MONNBTCH.« KOKKETCH + PROT1CH - 1                ' net charge
 HONZ{0*) « MONNBTCH * 8P(HSX,JV)
                                                 • • i
 NEXT y\    .

 • Now calculate total amounts bound, and total net charge (monodentate sites)

 FOR K* - 1 TO NSP*
 IF  BINDTBST\(K*) • 0 THEN WS8RCLC                         .
           . 0
FOR J\ - 1 TO 8
W3SR6l.6t
NEXT K%
MONZCALC •  0         '

FOR 0% - 1  TO 8
MONZCALC «  HONZCUiC * HOHZ(J\)
MSXT J*
,»'*•**«*****•*•**••****«••••••••*•*••*•**«•««*
 •Overall  automation t bidantata + aonodantata

 FOR me • 1 TO NSP*     \                  s
 IF  BXNDTBSTMK*) • 0 THEN HS8R6L7
 KU(H8\rK«c) «
 W8SR€L7t
 NEXT K*
' Calculated value of 2, and Z error term

ZCALC(HS*) - BXZCALC + MOHZCUiC
ZERR(HS\}  • (

RETORK
**tff*ltiiMttillMlttffffitt»ltltHit»ilttfffSI»ftt*iff»tti*ttf»flftfiltfftfit
 HS8R7i     ' called by M88RS                                                  It
'•••         caloc binding by DL aoounulation, for FX and H\        •         ft
•eeiftiiitfiititiitiii»«iiitittiiiiififfifiiittitiititistiiiitiiHi«fitfstitiiH

' Flnt clear  the DDL array for thii HflX

FOR 0* -  1  TO  NSF*                          -          ,         '
CDPtL(B8\,J1t| • 0                        .
NEXT OK               '      •
TOTCHN «  ZED(HS*)  * TH(B8%)
TDCONC « -  TOTCHH /  DVOL(R8%)
IF TDCONC <  0 THEN WSSR7ltl
* total charge to be
* neutralised, par litre

* total cone of counteriona
• par litre of diffuse layer

' anions attracted
 «*«***«««««***•**»••*•*•**•*••*••««***•*••********•**•****•******•*****
 ' cone  to here if humics have a net negative  charge'(cations attracted)

 WSSR7L2I                               .

 TDCONCCALC  - 0

-------
 1012
                                    E TIPPING
FOR XV -  1 TO NSF*                                           •
IF  CH*(X*)   <  0 THEN CDDXi(KS*,X%) • 0
IF  CB*(X*)   <  0 THEN WS8R7&3
CDDL(BS*,X%} «  C(X%)  • K8BLRS(X*} *  (RATIO (HS*)«(CB* (X*)H
TDCOKCCALC   •  t DCONCCALC 4  (CDDI,  0 THEN CDDXi{HS*,X*) " 0
IF CHMXV)   >  0 THEN KSSR7I4
CDDI.(E8%,X%) -  C(XX)  * K8ELHS(X«) * (RATIO (HSV)
TDCONCCALC   •  TOCONCCAbC  / (TDCONC 4 TDCONCCALC))*2
 IF TDCONCERR < CRIT THEN WSSR8L2

-------
                          Equilibrium chemical speoation by WHAM

         ratio and r«-try

 RATXOCLAT - ((RATIOCLAY • TDCONC / TDCONCCAtC)  + RATIOCIAT) /2
 GOTO WSSR8L1           .            '  -

 WSSRBWt
 RETURN                                     •
                                                                           |(j|3
 *fti*ftiiiftlift«ll«Hlffiifllfft»ftfffifitHftlfttt«iitt»«H«fifHffHii|«|fti|
  HSSR9S    '  call«d by Bain program                                         ••
 'if          calcs distrbtn of FA batman solid end aqnaeus phases'          ft
 'ftiftlftl*ifftffllltftMf»lttiftWlfflKK»filMtfffitiffif|«ffft|fl|i|t«fiitflf||
 CFA  - TO (2) / CSOLID

 I***********************************
 * Calculate distribution term for FA

 DISTSICUA -0
 FOR X% • 1 TO 10
 PISTSICMA « OZ8TSZGKA + (X**aUOOJ
 NEXT I\                 '•    ,

 <****•**•****************•************«*
 'C*lcul«ta agu«ou» cone* of FA fraction*

 FOR X* «  1 TO 10                        ,

 ZFA « {0.1 * HCOOHCt)  * 1%)
 ZFA « BETA * (XPA - ABB(IBD(2)))
 XFA «1* (KO * EZP(ZFA)  •  GflOUED)

 YFA - (X**GAHMA)  * CSOLID *  CFA /  DISTfiXOtO.

 FA(Z%} - TFA / JCFA-."

 NEXT Kt     .          ' •  ,  •         . .

 «**•*********••**••*••**••••••
 *  Sum to find total aguoouc  FA

 FAAQ m 0

 FOR I* • 1 TO 10           -
 FAAQ - FAAQ  4 FA(Z%)
 NEXT X%
                                                  ' gFA par gsoil
RETURN
'ft
             -called by mala program - call* WSSR11 and K8SR12
             aakoi output flla    '
                                                                             ••
DIVIDERS " —
FF - 1.00001
                                       ' format factor ; avoids massy output
PRINT §2, DIVIOER$
PRINT f2,_
"****•****        OOTFDT FILE FROM WHAM-S  VERSION 1.0
PRINT f 2. 0IVIDBR$
PRINT #2, "
PRINT §2, -SOURCE FI&E «;TAB(25) ;SP$
PRINT 82, "DATABASE    *;TAB<25) ;DBS$
                                                               «*•****.••

-------
1014
                                   E. TIPPING
IF FC02 m 999     THEN PRINT *2,«PC02-;TAB(25);"VARIABLE-
IP PC02 » 999     THEN PRINT ft2,«PCO2";TAB(2S);"FIXED-
PRINT 82, "PRECISION % «;TAB(2S);0SING «8.888****«|FF*PRBCISION
PRINT 82,
PRINT 82, DIVIDERS
PRINT #2,                                    '
PRINT #2, "INPUT DATA"
PRINT 82,
PRINT #2,
PRINT 82,
          "TEHPK";
          "CSOLID";
PRINT 82, -TOT HA";
PRINT 82, "TOT FA";
PRINT 82,
IF FAAQST
                     TAB(2S);USING  «8.lf*****";FF«TEMPK
                     TAB(25);USING  "8.*ft****";FF«CSOLID-
                     TAB(25);0SING  "8.8t8*AA*«jFF*TH(l)         .
                     TAB(25);0SINO  «8.888*A*A«iFF*TH{2)
           "TOT CLAY«;TAB(25);USING  "8.888*AAA";FF*TCLAY
            999 THEN PRINT 82, •FAAQ";TAB(2S);" UNKNOWN"
IF FAAQST  < 999 THEN PRINT 82, •FAAQ«;TAB(25);17SINO "8.888*AAA";FF*FAAQST
PRINT 82,  "PC02";      TAB(25);0SINO "i.888AAAA«;FF*PCO2
PRINT 82,  '
PRINT 82,  "MASTER SPECIES";TAB(25);"TOTAL CONC"   •.
FOR XX « 1 TO 100
IF XX • 1  THEN WSSR10L1
IF XX - 51 THEN WSSR10L1
IF T(XX) • 0 THEN WSSR10L1
PRINT 82,XX;TAB(5);N$(XX);TAB(2S);UBINO "8.8*8AAAA";FF*T(XX)
WSSRlOLl*                                               .
NEXT 3Pb       '                    •
PRINT 82,
PRINT 82,DIVIDERS
PRINT 82,
PRINT 82,  "RESULTS-
PRINT 82,                                      :
PRINT 82,  "NO. OF XTERATION&";TAB(23};USING «8«8if";NDMITX
PRINT. 82,  «PH";               TAB(25)/USING  "8.888A**A";FF*PH
PRINT 82,  "IONIC STRENGTH";   TAB(25);USING  "8.888*AAA«;FF*IS
PRINT 82,  "CHARGE RATIO";     TAB(25)|VSINO  «8.888A*AA«;FF*POSCH/NBOCH
PRINT 82,  "CHARGE r>IFFKRENCB";TAB(24);0SIRO "+8.888A*AA";FF«POSCH-NEGCH
XF PAAQST  « 999 THEN PRINT 82, "FAAQ (CMC)" ;TAB(25);0SINO «8.888AA*A";FF«FAAQ
IF FAAQST  < 999 THEN PRINT 82, "FAAQ (FIXED)-;TAB(25);USING "8.888*AAA";FF*FAAQ
IF TH(1) > 0 THEN PRINT 82, "ZED-HA";TAB(24);USIMO •*8.888****"|FF*ZEO(1|
IF TH(2) > 0 THEN PRINT 82, "IBD-FA";TAB{24) ;USING "+8.888*AAA*;FF*ZED(2)
IF TH(1) > 0 THEN PRINT 82, "RATIO-HA"  ;TAB(25);VSINO "8.888A*AA";FF*BATIO(1)
XF TH(2) > 0 THEN PRINT 82, "RATIO-FA"  ;TAB(25);USINO •8.888AAA*";FF*RATIO(2)
IF TCLAY > 0 THEN PRINT 82, «RATIO-CLAY";TAB(25);OSINO "8.888AAA*"fFF*RATIOdAY
PRINT 82,
PRINT 82, "MATER VOLUMES",         '
PRINT 82,"FRACTION HA-DDL, •  ;TAB(25);USING. "8.888AAAA«;FF*DTft)L(l)
PRINT 82,"FRACTION FA-DDL "  ;TAB(2S)/USING "8.888AAAA";FF*DVPL(2)
PRINT 82,"FRACTION CLAY-DDL ";TAB(25);USING -8.888AAAA";FF*DVOLCIAY
PRINT 82,"FRACTION SOLUTION ";TAB(25);USING «8.888AAAA-;FF*VOLSOL
PRINT 82,
IF FBPPT$  « "YES" THEN PRINT 82,"[FE(OH)3 PPT]-;TAB(2S);
             USING «8.888***»-;T(ll) - TCALC(ll)        ~
IF FEPPT$  - -YES" THEN PRINT 82, "X TOT FE PPTD«;TAB(24);USING "888.88";
             100*(T(11) - TCALC(11))/T(11)
IF T(ll) > 0 AND FEPPT« • "HO-  THEN PRINT 82,-PE(OH)3 SATN ISDBX«*TAB(24)>-
             USING "+8.888**A*";LIAPFE - LKSOFE                   .
IF ALPPT$  « "YES" THEN PRINT 82,"[AL(OH)3PPTJ";TAB(25);
             USING «8.888****"|T(5) - TCALC(S)
IF ALPPT$  • "YES" THEN PRINT 82,"X TOT AL PPTD";TAB(24);0SINO "888.88";.
             100*(T(5) - TCALC(5))/T(5)  '
IF T(5) >  0 AND ALPPT$ • "NO"  THEN PRINT 82,«AL(OH)3 SATN INDEX";TAB(24);_
             USING "+8.888AAAA";LIAPAL - LKSOAL
PRINT 82,                                                             .
PRINT 82,  DIVIDER^
PRINT 82,
PRINT 82,  "FRACTIONAL DISTRIBUTION  OF MASTER SPECIES"
PRINT 82,
PRINT 82,  "MASTER SPECIBS";TAB(23);"S«;TAB(31);"DHA";TAB(40);*DFA";
                     TAB(49);-H&«;TAB(5e);"FA";TAB(66);"CLAY"      *"
PRINT 82,
FOR XX « 1 TO 100
IF XX - 1  THEN HSSR10L2
IF XX - 51 THEM WSSR10L2

-------
                         Equilibrium chemical speciation by WHAM
1015
 XF T(XX)  « 0 THEN WSSR10L2
 COSOB HSSE11                                  .
 PRINT «2,XX;TAB<7)>N$(XX);TAB(21);X«ING "*.B68";FRACS;TAB(30);FRACDHA;_
       V      TAB(39) |FRACDFA;TAB(48);FRACIIA;TAB($7);FRACFAjTAB(66}fFRACtCLAY
 WSSRIOMt                            .
 NEXT XX
 PRINT *2»
 PRINT #2, DIVIDER?                                                   .  -
 PRINT *2,                                                       .
 PRINT #2, "CONCENTRATIONS AMD ACTIVITIES OF SPECIES XN SOLUTION PHASE'
 PRINT t2,                                                    .
 PRINT 82,«SPECIES-»TAB{23)f"C08C«;TAB(38)|"ACTIVITY"
 PRINT 12,                   .                 .
 FOR XX •  i TO SSPX
 XF C(XX)  « 0 THEN WSSR10L3                                  .
 PRINT «2,XX;TAB<7)|N$TOINO «t.*«»***«jC         •                             f|
 *ti  .       . calca dlBtrlbution of cho>«& ma«t«r *p«ei«B          .           if
 'lllifft»lffff»fftf«f»l8fff8HKIt8lt««t«»i»ft|lttHf»iHSi*»ftftftliftt»tftHI
TOTALS
TOTALDHA
TOTALDFA
TOTALDC
TOTALHA
TOTALFA
FOR TtX • 1 TO NSP%
XF KlX(YX)*'
XF MlX(YX)
XF HIX(YX)
IF MIX(YX)
IF HIX(YX)
IF MIX(YX)
IF M2X(TX)
IF M2X(YX)
IF K2X(YX)
IF K2X(YX)
IF K2XIYX)
IF H2X(YX)
XX THEN TOTALS
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALDC
XX THEN TOTALHA
XX THEN TOTALFA
XX THEN TOTALS
XX THEN TOTALDHA
XX THEN TOTALDFA
XX THEN TOTALDC
XX THEN TOTALHA
XX THEN TOTALFA
TOTALS 4 (VOLSOL «
TOTALDBA 4 (DVOL(l)
TOTALDFA 4 (DVOL(2)
TOTALDC 4 (DVOLCLAY
TOTALHA 4 (CHC(1,YX)
TOTALFA 4 (CHC(2,YX)
TOTALS 4 {VOLSOL
TOTALDHA 4 (DVOL(i)
TOTALDFA 4 (DVOL(2)
TOTALDC 4 (DVOLCLAY
TOTALHA 4 (CHC(1,YX)
TOTALFA 4 (CKC(2,YX)
> C(YX) * SIX(YX))
CDDL(1(YX) * SIX(YX))'
CDDL(2,YX) * £1X(YX})
CLAYDDL(YX) * SlX(YXJ)
SIX(YX))
SIX(YX))
C(YX) * S2X(YX))
CDDL(l.YX) * S2X(YX))
CDDL(2,YX) * S2X(YX)}
CLAYDDL(YX) * S2X(YX))
S2X(YX))

-------
 1016
                                       E TIPWNO
IF H3*(Y*)
IF 1(3% Oft)
IF H3*(Y*}
IF M3*(X*)
IF H3*(X%)
IF H3*(Y*)
X* THEN TOTALS
X* THEN TOTALDHA
X* THEN TOTALDFA
X* THEM TOTALDC
X* THEN TOTALHA
» THEN TOTALFA
TOTALS 4 (VOL80L
TOTALDHA 4 (DVOL(l)
TOTALDFA 4 (DVOL(2)
TOTALDC 4 (DVOLCLAY
TOTALHA > (CHC(1,*\)
TOTALFA 4 (CHC(2,X*)
NEXT
FRACS
FRACDHA
FRACOPA
PRACDCLAY
FRACHA
FRACFA
TOTALS / TCALC(X%)
TOTALDHA / TCALC(Xfc)
TOTALDPA / TCALC(X\)
TOTALOC / TCALC(X%)
TOTALHA / TCALC(tt)
TOTALFA / TCALC(X«)
RETORN
 WSSRl2i
'ft
'ft
' called by WSSR10
 -calc«  total. *qu0ou« coacni of na«t«r *p«el««
  cmlca  Kd valua*                 .
                                                                                     iff
                                                                                     *«
                                                                                     »§
' Calculattt «guaouB voluna and total noloi pr«a«nt

IF TH(2) > 0 THEN VOLAQ  « VOLSOL /  (1 - (DVOL(2)*PAAQ/TH(2» )
IF TH(2) - 0 THEN VOLAQ  - VOLSOL •                                      .

IF TH(2> > 0 THEN TOTAQ  « TOTALS +  ((TOTALFA  4 TOTALDFAJ*FAAQ*VOLAO/TH(2J I
IF TH(2) - 0 THEN TOTAQ  .TOTALS                               V    «'"1*"

AQCONC - TOTAQ  / VOLAQ

TOTSOLID « T(«) - TOTAQ

KD * (TOTSOLID  / CSOLID) / AQCONC

                                                      ;   . •

' Apparent KD ; calculated for ALL water being in 8 (DDL ignored)

TOTSOLIDAPP • T(rt) - AQCONC

KDAPP • (TOTSOLZDAPP / CSOLID) /AQCONC

RETURN                            •                                              ,
fltffiffHfffftt**tti«ifflili««ttMtiifft**t»i*fti*Mt*flitftf«tfiil*tff|«liii||
'i*fffiffffftff*fifift«fftf|lt«t*tliH*l«fflf*ttftft**i«fI*»tfWHtf|*!HiHIH«fti
                                    APPENDIX 4

  WHAM-W input and output files, water example. In the input file the master species' concentrations ate in moles per
liter of total water in the system. «nd the concentrations of humic and fulvic acids are in g I*'. The partial pressure of CO,
(pCX)2) is in atmospheres. In the output file, ooneenlrstions 'of individual species, ionic strength, charge difference, and
carbonate alkalinity, are given b moll*1, concentrations of humic and fulvic adds in gl*1. The terras ZED-HA and
ZED-FA refer to die net charges on the fulvic and humic acid in cquivg*1. The terms RATIO-HA and RATIO-FA refer
to the variable K in Equation (8). The water volumes are in liters,

   MATER          ,WATERT1
   Database       ,KATER10
   Database eizo  ,400
   Precision*     ,0.01
   TeapK          ,283
   THA TFA g/i    ,lE-3,9E-3                                    •

-------
                      Equilibrium chemical spctiation by WHAM
                                                         1017
pH fixed
PHSTART
pco2
Ho. jbast.
Ha
Kg
Ca
Hi
cu
Cl
S04
          •p.
,7
,0.00035
,7
,3,0.0002
,4,0.0001
,7,0.0003
,13,18-6
,14,1H-C
,52,0.0005
,54,0.0002
«••*•*«*•        OUTPUT PILE FROM HBAM-W VERSION 1.0          •**•••**•
««•••»«*****«••**«*«**•*•****•****•**•*<>**••••**'******************•••"
SOURCE PILE
DATABASE
PH
PRECISION *
STARTING PH
          KATERT1
         - HATBR10
          PIXBD  •
          l.OOOB-02
          7.000E+00
*•**•*«••*****••****••*••***•*•****«*•**********•«*•*•*•*•*********•*»*
INPUT DMA

TEUPX
TOT HA.
TOT PA
PCO2.

WU5TBR SPECIES
~\3  Na
 4  Ktr
' 7  Ca
 13 Hi
 14 Cu
 52 Cl
 54 S04
          2.fl30E+02
          l.OOOE-03
          9.000E-03
          3.SOOE-04

          TOTAL CONC
          2.000B-04
          l.OOOE-04
          3.000E-04
          l.OOOE-06
          l.OOOE-OC
          5.000E-04
          2.000E-04
 ***•«**••»***••**«•••***«****•**•*•*•••*•*••••**••*••*••**•**•••••****•
RSSUUIS

NO. OP ITERATIONS
PH     .
IONIC STRENGTH
CHARGE RATIO
CHARGE DIPPERBNCE
ZED-HA
ZED-PA
RATIO-HA
RATIO-PA

HATER VOLUMES
FRACTION HA-DDb
FRACTION PA-DDL
FRACTION SOLUTION
          48
          7.000E400
          1.518E-03
          0.985E4-00
         -1.401E-OS
         -1.235B-03
         -2.746E-03
          4.495E400
          2.071E400
          7.731B-05
          6.841B-03
          0.993E+00
 CARBONATE ALKALINITT   4C.954B-05
FRACTIONAL- DISTRIBUTIONS OP MASTER SPECIES

MASTER SPECIES        S       DHA      DPA
3    Na
4    Mg
^    Ca
0.986
0.947
0.947.
0.000
0.001
0.001
                                      0.014
                                      0.026
                                      0.027
                                   HA

                                  0.000
                                  0.002
                                  0.002
                                                          PA

                                                         0.000
                                                         0.022
                                                         0.022

-------
1018
                                   E. TIPPING
    13
    14
    52
    54
Hi
cu
Cl
S04
0.538
0.002
1.000
1.000
0.001
, 0.000
0.000
0.000
0.015
0.000
0.000
0.000
0.010
0.429
0.000
0.000
0.436
0.569
0.000
0.000
   CONCENTRATIONS AND ACTIVITIES
   SPECIES
                •  TOTAL HATER
                [FREE]     [ORGANIC]
1
3
4
7
13
14
51
52
54
55
101
102
124
125
126
145
146
147
"••-' 51 )'•'•'
- -<6ii'ftj
•212
213
214
215
216
221
222
223
224
225
226
227
B
Ha
Kg
Ca
Hi
Cu
OB
Cl
B04
C03
HC03
H2C03
KgHCO3
MflC03
KgS04
CaBCO3
CaC03
CaSO4
^HiOH'TsiV^
%{OH)2
N1S04
NiCOS
NiCl
N1HC03
CUOH
CU(OH)2
CUS04
CUC03
CU(C03)2
CuCl
CUBC03
1.0378-07
1.971E-04 .
9.2688-05
2.7738-04
4.9388-07
1.070E-09
3.0808-08
5.000E-04
' 1.914B-04
2.666E-08
6.9S68-05
1.8278-05
6.030E-08
1.334E-09
1. 9888-06
. 1.6648-07
6.426E-09
6.591B-06
1.792E-10
3.696E-13
1.246E-08
5.739E-09
5.058B-10
2.555E-08
8.495B-11
4.825B-13
2.803E-11
1.1618-10
,4.6088-15
9.9178-13
8.980E-10
1.S15B-09
2.6818-06
5.253B-06
1.5898-05
4.6128-07
9.8488-07
O.OOOB+00
O.OOOE+00
O.OOOB+00
O.OOOB+00
O.OOOE+00
1.2738-07
8.8138-10
9.2938-12
1.38SE-08
2.4328-09
4.476E-11
4.592E-08
3.925E-11
2.5748-15
8. 6808-11
3.9988-11
7.3938-12
3.7358-10
1.3028-08
3.3618-15
1.9538-13
8.0848-13
O.OOOE+00
1.4498-14
1.3128-11
      ,  SOLUTION PHASE
      CONG       ACTIVITY
                                                    1.044E-07
                                                    1.9858-04
                                                    9.333B-05
                                                    2.7928-04
                                                    4.973B-07
                                                    1.0788-09
                                                    3.1028-08
                                                    5.0352-04
                                                    1.927B-04
                                                    2.6858-08
                                                    7.004E-05
                                                    1.840E.-05
                                                    6.072B-08
                                                    1.343E-09
                                                    2.0028-06
                                                    1.676B-07
                                                    £.4718-09
                                                    6.6378-06
                                                    l.eoSB-10
                                                    3.7218-13
                                                    1.255B-08
                                                    5.779E-09
                                                    5.093B-10
                                                    2.573E-08
                                                    8.554B-11
                                                    4.859E-13
                                                    2.8238-11
                                                    1.169E-10
                                                    4.640E-15
                                                    9.98CB-13
                                                    9.0438-10
                                                       l.OOOE-07
                                                       1.9028-04
                                                       7.9108-05
                                                       2.3678-04
                                                       4.2158-07
                                                       9.1338-10
                                                       2.971E-08
                                                       4.824E-04
                                                       1.633E-04
                                                       2.275B-08
                                                       6.7108-05
                                                       1.840B-OS
                                                       5.817E-08
                                                       1.3438-09
                                                       2.002B-06
                                                       1.6058-07
                                                       6.4718-09
                                                       6.6378-06
                                                       1.729E-10
                                                       3.721E-13
                                                       1.255B-08
                                                       S.779B-09
                                                       4.8808-10
                                                       2.4658-08
                                                       8.195E-11
                                                       4.8598-13
                                                       2.623E-11
                                                       1.1698-10
                                                       3.933B-15
                                                       '9.567B-13
                                                       8.6638-10
   VALVES OF 00 (MOL BOUND / O ROHIC SUBSTANCES)
   NO IS SPECIFIC BINDING ; DNO 18 DIFFUSE LAYER
    1    H
    3    Ha
    4    Kg
    7    C»
    13   Hi
    14   Cu
    102  H2C03
    124  H0RC03
    125  MsfCOS
    12C  MgS04
    145  CaHC03
    146  CaC03
    147  C«S04
    211  NiOH
    212  Hi(OH)2
    213  N1SO4
    214  N1C03
    215  NiCl
    216  N1HC03
    221  CUOB
    222  Cu(OH)2
                        DNU-HA
                           DNU-FA
RU-HA
KU-FA
.627B-08 1
.8988-05
.4588-04
.362E-04
.7688-07
.6838-09 .
1.422B-06
2.1108-08
1.0398-10
1.5488-07 ]
5.8238-08 ;
5.002E-10 4
5.1318-07 !
6.271B-11 ;
2.877E-14 :
9.700E-10 i
4. 4 678-10 4
1.770E-10 I
8.942E-09 4
2.973B-11
3.756E-14
L.643B-07
.125E-04
.0438-04
.1038-04
.6218-06
.5138-09
.3988-05
.5S8B-08
.021E-09
L. 5228-06
i. 6388-07
I. 9188-09
>. 0458-0 6
2.8418-10
2.8298-13
». 5378-09
1.392E-09
9.018E-10
l.OSOB-08
L.347E-10
J.693E-13
O.OOOE+00
O.OOOB+00
1.606E-04
6.S79B-04
9.5038-06
4.2588-04
O.OOOB+00
O.OOOE+00
O.OOOB+00
O.OOOE+00
O.OOOE+00
O.OOOB+00
O.OOOB+00
2.886B-10
O.OOOB+00
O.OOOE+00
O.OOOE+00
O.OOOB+00
O.OOOE+00
2.829E-06
O.OOOE+00
O.OOOB+00
O.OOOB+00
2.454E-04
7.3418-04
4.8488-05
6. 2118-95
O.OOOB+00
0.0008+00
O.OOOB+00
0.0008+00
O.OOOE+00
O.OOOB+00
O.OOOE+00
4. 0388-09
OJOOOE+OO
O.OOOB+00
O.OOOE+00
O.OOOE+00
O.OOOB+00
1.132E-06
O.OOOE+00

-------
     223  CUS04
     224  CUCO3
     226  CUC1
    .227  CuHCOS
                          Equilibrium chemical spcciation by WHAM
                                                                           1019
2.182B-12
9.034E-12
3.470E-13
3.142B-10
2.145B-11
8.882E-11
1.S72E-12
1.423E-09
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
O.OOOB+00
O.OOOE+00
O.OOOE+00
O.OOOE+00
                    files., ccdin
                                APPENDIX 5

                          nt example. See explanation of Appendix 4. In addition, note that CSOLJD and
  WHAM-S input and output
TCLAY are the concentrations of total wilds and total cation-exchanger day in g|-'.

SEDIMENT
Databaaa
Array •!••
PraclcionX
TanpK
CSOLID
THA TPA
TCIAT
FAXQ
pCO2
No. mast «p.
Na
Mg
Ca
FaZIX
Hi
Cu
Cl
604
,8EOTl
,88B010
,400
,0.1
,278
,100
,5,0.1
,20
,0.020
,0.001
,8
,3,«E-4
,4,4B-3
,7,5E-3
,U,5E-3
,13, IK -5
,14,lE-5
,S2,5E-4
,54,lE-4
                                                                         I***
   *********        OUTPUT FILE FROM WHUi-8  VERSIOH 1.0         *********
   •A**********************************************************************
                            8BDT1
                            SSED10
                            0.100B+00.
SOURCE FILE

PRECISION *•

•••••ft*****************************************************************
   INPUT DATA

   TEMPK
   CSOLID
   tOT HA.
   TOT FA
   TOT CLAY
   FAAQ
   PC02

   MASTER SPECIES
    3  Ha
    4  Kg
    7  Ca
    11 FalXI
    13 Hi
    14 Cu
    52 Cl
    54 S04 /
                         2.780E+02
                         l.OOOE+02
                         S.OOOE+00
                         0.100E+00
                         2.000E+01
                         2.000B-02
                         l.OOOE-03

                         TOTAL COHC
                         fi.OOOE-04
                         4.000E-03
                         5.000B-03
                         5.000B-03
                         1.00OB-OS
                         l.OOOB-05
                         5.000E-04
                         l.OOOE-04
   RESULTS   '

   NO. OF ITERATIONS
   PH
   IONIC STRENGTH
                         367
                         7.016E+00
                         1.475E-03

-------
1020
                                   E TIPPING
   CHARGE RATIO
   CHARGE DIFFERENCE
   FAAQ (FIXED)
   ZED-HA   •>
   ZED-FA
   RATIO-BA
   RATIO-FA
   RATIO-CLAY

   HATER VOWJMBS  .

   FRACTION BA-DDIi.
   FRACTION FA-DDIi
   FRACTION CLAY-DDI,
   FRACTION SOLUTION

   tFE{OH)3 JPPTJ
   * TOT FE PFTD
                  0.999B+00
                 -1.0448-06
                  2.000B-02
                 -1.3208-03
                 -2.668B-03
                  9.4648+00
                  4.3058+00
                  2.5098+01
                  0.1388+00
                  2.6178-02
                  5.13SB-03
                  0.8318+00

                  3.6628-03
                  73.25
   FRACTIONAL DISTRIBUTION OF MASTER SPECIES
   MASTER SPECIES
    3
    4
    7
    11
    13
    14
    52
    54
N*
Kg
CA
Fftlll
Nl
ca
cl
604
0.633
0.032
0.025
0.000
0.005
0.000
1.000
0.994
 DBA

0.248
0.354
0.363
0.000
0.065
0.000
0.000
0*005
 DFA

0.021
0.014
0.014
0.000
0.003
0.000
0.000
0.001
 HA

0.000
0.465
0.494
0.973
0.882
1.000
0.000
0.000
 FA

0.000
0.011
0.009
0.027
0.028
0.000
0.000
0.000
CLAT

0.098
0.124
0.095
0.000
0.017
0.000
0.000
0.000
   *•**•**•***••*•****«««***••****•*•*••**««*****«•«*««*«,•«*«»,,**,««„«<

   CONCENTRATIONS AND ACTIVITIES OF SPECIES IN COLDTION PHASE

   SPECIES              COSC           ACTIVITY     •
                                                          i
    1
    3
    4
    7
    11
    13
    14
    51
    52
    54
    55
    101
    102
    124
    125
    126
    145
    146
    147
    185
    186
    187
    188
    190
    194
    195
    211
    212
    213
    214
    215
    216
    221
H
Km
Kg
Ca
Nl
Cu      ,
OB
Cl
S04
COS
BC03
B2C03
H0BC03
KgC03
M0S04
CABC03
C«C03
C*S04
F«OB
Fa (OB) 2
F«(OB)3
Fa (OB) 4
F«S04
F«C1
F«C12
NlOB
Ni(OE}2
N1SO4
N1CO3
NiCl
N1BCO3
CuOH
1.006E-07
4.S68B-04
1.531B-04
1.470E-04
2.674E-17
S.1998-08
1.982B-12
2.091E-08
6.016B-04
1.1588-04
7.968E-08
2.2548-04
6.130E-05
3.1368-07
6.0368-09
1.717E-06
2.614B-07
9.086E-09
2.007E-06
4.124E-13
5.5168-10
2.7478-10
1.133B-12
1.310E-17
1.9338-19
8.808E-22
1.276B-11
1.778B-14
7.S67E-10
1.8068-09
6.283B-11
7.747E-09
1.065E-13
        9.6438-08
        4.3808-04
        1.302E-04
        1.251B-04
        1.881B-17
        4.422B-08
        1.6868-12
        2.005B-08
        5.7698-04
        9.8538-05
        6.778B-08
        2.162E-04
        €.1308-05
        3.0078-07
        6.036B-09
        1.7178-06
        2.506B-07
        .9.0868-09
        2.0078-06
        3.5078-13
        5.289B-10
        2.7478-10
        1.0868-12
        1.2568-17
        1.645B-19
        8.4478-22
        1.224E-11
        1.778E-14
        7.567E-10
        1.806E-09
        6.0258-11
        7.4298-09
        1.0218-13

-------
                        Equilibrium chemical specialion by WHAM

   222   Cu(OH)2        4.0858-16        4.0858-16
   223   CUS04          2.9388-14        2.9388-14
   224   CUC03          €.4258-13        6.4258-13
   225   CU(C03)2       7.5728-17        6.4418-17
   226   CttCl        '  2.0922-15        2.0068-15
   227   CUHC03         4.7908-12        4.5938-12
                                                          1021
  *•*•**•••*•••**•«*******•**•*****•*•***••******«***«**•**********•****•
  TOTAL AQOBOXfS-PHASB CONCBHTRATXOKS AND KD VALUES

  SPECIES '  .          AQCOHC              KD  '
   3.    Na
   4     Kg
   7     C*
   11   F«III
   13   Hi
   14   Cu
   52   Cl
   54   804
      4.570E-04
      1.746E-04
      1.716E-04
      7.20BB-06.
      1.228E-07
      4.144B-10
      5.985E-04
      1.190B-04
44.775E4-00
42.207B402
42.630E4-02
48.0COB402
42.413E405
-2.918E-OC
+5.'151E-02
  KD-AFP

43.1298*00
+2.1918+02
+2.B13E+02
+C.927E+03
+8.043E+02
+2.4138+05
-l.«458+00
-1.594B+00
                               APPENDIX 6

WHAM-S input «nd output files, soil enmple. See oqduution to Appendices 4 ud 5.
 SOIL
 Data
 Array «lr«
 TompK
 CSOLZD
 TEA TPX
 TCIAY
 FAIQ
 PCO2
 No. »a«t vp.
 Ha
 Kfl
 Al
 Ca
 FaXXZ
 Co
 Sr
 Ca
 Am
 Cl
 604
,SOZLT1
,888010
,400.
,0.1
,278
,300.
.25,5
,0
,999
,0.01
,11
,3,28-4
,4,18-3
,5,38-2

'11,28-3
,12,18-6
,16,18-6
,18,18-6
,28,18-6
,52,28-4
,54,48-5
  «***•**•*
                   OOTPUT FILE FROM WHAM-S  VERSION 1.0
                                                                «•*•*••**
  SOURCE FItS
  DATABASE
  PRECISION \
          SOILT1
          SSED10
          0.1008+00
  INPUT DATA

  TBHPK
  CSOLID
  TOT HA
  TOT PA
           2.780E+02
           3.000E+02
           2.SOOE+01
           5.000E+00

-------