EPA/600/R-94/138
                                             August 1994
   FISH PHYSIOLOGY, TOXICOLOGY
AND WATER QUALITY MANAGEMENT
          Proceedings of 3rd Biennial International
       Symposium, Nanjing, PRC, November 3-5,1992
                       Edited by

     David J. Randall1, Hong Xiang1, and Robert V. Thurston2
                  'Department of Zoology
             The University of British Columbia
             Vancouver, B.C. Canada V6T 1Z4

               fisheries Bioassay Laboratory
                 Montana State University
              Bozeman, Montana USA 59717
       ENVIRONMENTAL RESEARCH LABORATORY
        OFFICE OF RESEARCH AND DEVELOPMENT
       U.S. ENVIRONMENTAL PROTECTION AGENCY
                ATHENS, GEORGIA 30605
                                            Printed on Recycled Paper

-------
                                    DISCLAIMER
       The information in this document has been funded in part by the United States
Environmental Protection Agency. Papers describing EPA-sponsored research have been subject
to the Agency's peer and administrative review, and have been approved for publication.
Mention of trade names or commercial products does not constitute endorsement or           !
recommendation for use by the United States Environmental Protection Agency.
                                         ii

-------
                                     FOREWORD

       Joint ecological research involving scientists from every country in the world is essential
if global environmental problems are to be solved. Recognition of this international aspect of
environmental pollution control is reflected in the joint activities undertaken under Annex 3, Item
4 of the USA-PRC Protocol for Environmental Protection. This component of the protocol
provides for cooperative research on the environmental processes and effects of pollution on
freshwater organisms, soils, and groundwater, and on the application of transport and
transformation models.

       Specific areas of cooperation in environmental research include inorganic chemical
characterization and measurement; inorganic chemical transport and transformation process
characterization; biological degradation process characterization;  oxidation/reduction process
characterization; field evaluation of selected transport, exposure and risk models; and application
of models for environmental decision-making concerning organic pollution in semi-arid
conditions, heavy metal pollution, and permissible loading of conventional and toxic pollutants
in Chinese rivers. Activities include seminars, workshops, symposia, training programs, joint
research, and publications exchange.

       The third biennial symposium presented under the protocol was held on the campus of
Nanjing University, Nanjing, Jiangsu Province, People's Republic of China, on November 3-5,
1992.  Scientists from four countries presented papers at the symposium, which was sponsored
by the U.S. Environmental Protection Agency, the PRC Environmental Protection Agency, the
PRC National Science Foundation, the EPA of Jiangsu Province,  the University of British
Columbia, and the Nanjing University. The two earlier symposia were held in Guangzhou, PRC,
on September 14-16, 1988, and Sacramento,  Calif, USA, on September 18-19, 1990.

       Symposia are an effective means of fostering cooperation  among scientists from different
countries, as environmental organizations seek to gain the information necessary to predict the
effects of pollutants on ecosystems and apply the results on a global scale. The symposia provide
a forum through which distinguished scientists from laboratories  and institutes from several
countries can exchange scientific expertise on environmental problems of concern to EPA and
the international environmental community.
                                               Rosemarie C. Russo, Ph.D.
                                               Director
                                               Environmental Research Laboratory
                                               Athens, Georgia
                                           111

-------
                                      ABSTRACT
       Scientists from four countries presented papers at the Third Biennial International
Symposium on Fish Physiology, Toxicology and Water Quality Management, which was held on
the campus of Nanjing University, Nanjing, Jiangsu Province, People's Republic of China.  This
proceedings includes 20 papers presented in sessions convened over 3 days. Papers address the
regulation of growth hormones in fish, the reproductive effects of anthropogenic chemicals in
fish, the effects of pollutants on physiological functions in fish, and the physiological responses
of rainbow trout to copper and ammonia. Descriptions are provided concerning research on the
effects of fenpropathrin on aquatic organisms, the bioaccumulation of organic chemicals in fish,
the relationship between uptake of chemicals and oxygen consumption in fish, the toxic effects of
meothrin on gill ultrastructure in grass carp, the bioaccumulation of hydrophobic organics in
foodwebs, the monitoring of industrial effluents in Nanjing, the investigations offish deaths in
Lake Xuanwu, and the rehabilitation of streams degraded by mining wastes. Presentations also
covered nonpoint sources and water quality, conversion of municipal wastewater into resources,
modeling of toxic chemicals in the Yangtze River, regional water pollution control in rural
China, adsorption of organic pollutants on soils and sediments, heavy metal speciation modeling
on the Le An River, and uncertainty analysis using the QUAL2E model.

-------
                                    CONTENTS
                                                                              Page
FOREWORD	iii
ABSTRACT		iv
ACKNOWLEDGMENT	vii

SESSION 1. FISH PHYSIOLOGY
      Hao-Ren Lin and David J. Randall—Chairpersons

      The Regulation of Growth Hormone Secretion in Carp  .................  1
             H.R. Lin, X.W. Lin, Q. Zhang and R.E. Peter
      Evaluationof Reproductive Effects of Anthropogenic Chemicals in Fish	14
             W.H. Benson, T. Dillon, and B. Suedel
      Sublethal Responses of Two Larval Fishes to Rice Field Pesticides	21
             A.G. Heath and J.J. Cech, Jr.
      The Effects of Water Pollutants on Some Physiological Functions of Fish  . .  .  . .  .  .31
             Y.M. Huang and W.H. Lin
      The Physiological Responses of Rainbow Trout to Copper and
      Ammonia: Mechanisms of Toxicity and Ammonia Excretion	36
             E.W. Taylor and R.W. Wilson

SESSION 2. FISH TOXICOLOGY
      Hongjun Jin and Robert V. Thurston—Chairpersons

      Ecotoxicological Study of Fenpropathrin on Some Common Chinese Aquatic
      Organisms and Evaluation of Its Effect on the Aquatic Ecosystem	47
             Y.W. Yin, Z.H. Wang, Y.Y. Zhang, Y. Xu, L.H. Xu, W.Z. Wu and G.S. Chen
      Bioaccumulation of Organic Chemicals in Tissues of Fishes	61
             R,V. Thurston, J.F. Neuman, C.J. Brauner and D.J. Randall
      The Relationship Between the Uptake of Organic Chemicals
      and Oxygen Consumption in Fish	71
             C.J. Brauner, J.F. Neuman, R.V. Thurston and D.J. Randall
      Toxic Effects of Meothrin on the Gill Ultrastructure in Grass
      Carp, Ctenoparyngodon idellus	84
             B.S. Xhou, Y.Y. Zhang and Y. Xu
      Chemical and Biological Controls on the Bioaccumulation
      of Hydrophobic Organic Compounds in Foodwebs	92
             D.L. Swackhamer
      Ecotoxicological Monitoring of Major Industrial Effluents in Nanjing, China	99
             H.J. Jin, X. Lou, Z.H. Zhang and G.X. Wang

-------
                                                                               Page

      Investigation into the Cause of the Death of Fish Occurring in Lake Xuanwu (IV)    108
             G.X. Wang, D. Fang, G.Q. Shi, X. Lou, Z.L. Xia and HJ. Jin
      Floodplain Rehabilitation Along Streams Degraded by Mining
      and Milling Wastes	112
             F.F. Munshower, D.R. Neuman and D.J. Dollhopf

SESSIONS. Water Quality Management
      Rosemarie C. Russo and Yuhuan Lin—Chairpersons

      Nonpoint Sources and Water Quality	133
             L.A. Mulkey, R.R. Swank, Jr., and R.C. Russo
      Ecological Engineering System for the Conversion of Municipal
      Wastewater into Resources in Zhenjiang	    139
             Y.H. Zhuo, X. Shen and X.L. Geng
      Water Quality Modeling of Toxic Organic Chemicals in the
      Yangtze River and Its Applications	151
             O. Y. Xu, G. Y. Sheng and H.X. Zou
      A Plan of Regional Water Pollution Control in Rural Areas of China .  .             160
             Y.C. Zhang
      Adsorption of Neutral and Ionic Organic Pollutants on Soils and Sediments  . .       169
             J.C. Westall
      A Study of Heavy Metal Speciation Modelling on the Le An River	183
             Y.H. Lin and Q. Li
      Uncertainty Analysis with Correlated Inputs Using QUAL2E .	196
             L.C. Brown
                                        VI

-------
                               ACKNOWLEDGMENTS
       Organizing and presenting a symposium and preparing a proceedings is frequently a
complex task, particularly when participants represent organizations in several countries.
The cooperative work of those involved is gratefully acknowledged.  Sponsoring organizations
for this symposium include the U.S. Environmental Protection Agency, the Environmental
Protection Agency of the People's Republic of China, the National Science Foundation of the
People's Republic of China, the Environmental Protection Agency of Jiangsu Province, the
University of British Columbia, and Nanjing University,- The scientists, engineers, and
environmental managers who participated in the symposium, of course, are deserving of primary
recognition. In particular, recognition is accorded to the organizing committee composed of
Dr. Hongjun Jin, Nanjing University; Dr. Haoren Lin, Zhongshan University; Dr. Yuhuan Lin,
Academia Sinica; Dr. Rosemarie Russo, U.S. Environmental Protection Agency; Dr. David
Randall, University of British Columbia, and Dr. Robert V. Thurston, Fisheries Bioassay
Laboratory. As session chairpersons, they also assured the efficient functioning of the
symposium.
                                          vu

-------

-------
 THE REGULATION OF GROWTH HORMONE SECRETION IN CARP

                        H.R. Lin,1 X.W. Lin,1 Q. Zhang1 and R.E. Peter2

                                     INTRODUCTION

       Recent progress has been made towards increasing our understanding of the processes involved
in the neuroendocrine regulation of growth hormone (GH) secretion in goldfish (Chang et al.,  1985;
Marchant et al., 1989a,b,c).  However, additional research is required before the role of the hypothalamus
in the regulation of growth hormone secretion in teleosts, especially the cultured carps, is understood to
the same extent that it is in homeothermic vertebrates.  Such research is essential to provide a foundation
for any subsequent studies on the stimulation of growth rates in these cultured fish.

       In the present study, both in vitro and in vivo experimental approaches were used to examine the
effects of hypothalamic peptides (gonadotropin-releasing hormone  (GnRH) and thyrotropin-releasing
hormone  (TRH)),   and   several  catecholaminergic  drugs   (apomorphine   (APO),   L-dopa,
oc-methyl-para-tyrosine (MPT), 6-hydroxydopamine (6-OHDA), epinephrine (E) and norepinephrine (ME)),
on the regulation of growth hormone secretion in the Common carp (Cyprinus carpio) and Grass carp
(Ctenopharyngodon idellus).


                              MATERIALS AND METHODS

       The influence of hypothalamic peptides on GH release from the pituitary of mature common carp
was examined in vitro using fragments of the pars distalis (adenohypophyses).  Common carp (350-600
g body weight) were obtained from local suppliers (Guangzhou, P.R. China).  Fish were maintained in
recirculating,  indoor, 250  L aquaria at room temperature  (18-28 °C) and a natural photoperiod.
Experiments were conducted using a column perfusion system, similar to that previously used for goldfish
pituitary fragments (Mackenzie et al., 1984; Habibi et al., 1989), but with minor modifications. Pituitaries
were dissected out from common carp anesthetized in tricaine methanesulfonate (TMS; Sigma, St. Louis,
MO)  and killed  by decapitation.   Pituitaries  were immersed in iced  Hank's balanced salt solution
supplemented with 25 mM Hepes (Sigma) and 0.1% bovine serum albumin (HBSS;  Sigma).  The
pituitaries were then washed twice with HBSS, and diced into fragments  (smaller than 1 mm3) with
surgical scissors.  The fragments were pooled and again washed twice with HBSS. Groups of fragments,
each equivalent  to  one-half of a  complete pituitary, were placed between  two layers  of  Cytodex
microcarrier beads (Sigma) in a 0.3 ml perfusion chamber.  The fragments were  perfused overnight (8-10
hr) with Medium 199 containing Hank's salts, L-glutamine (Gibco Laboratories, Grand Island, N.Y.), 25
mM Hepes and 56 U/ml Nystatin (Sigma) at a flow rate of 5 ml/hr. Two hours prior to experimentation,
the perfusion medium was switched to  HBSS at 19±1 °C  and the flow  rate increased to 15  ml/hr.
Fractions were collected at 5-minute intervals using an automated fraction collector and stored at -25 °C
until analysis. Synthetic (D-Ala6, Pro9-NEt)-LHRH was provided by Ningbo Fish Hormone Factory,
(Zhejing Province, China).  (D-Trp6, Pro9-NEt)-LHRH, chicken GnRH-H,  and sGnRH and its analogs
(D-Arg6, Pro9-NEt)-sGnRH and (D-Ala6, Pro9-NEt)-sGnRH, were kindly provided by Drs. J.E.  Rivier
and W.W. Vale (The Clayton Foundation Laboratories  for Peptide Biology, The Salk Institute, La Jolla,
    Department of Biology, Zhongshari University, Guangzhou, P.R. China

    Department of Zoology, University of Alberta, Edmonton, Alberta, Canada

-------
 CA). TRH was provided by the Laboratory of Reproductive Biology, Institute of Zoology, Academia
 Siruca (China).  All peptides were diluted immediately before use with HBSS from a stock solution (10
 pM for GnRH peptides, 100 uM for TRH).

        The influence of (D-Ala6, Pro9-NEt)-LHRH and several catecholaminergic drugs on serum GH
 levels was examined in vivo in Grass carp fmgerling.  Grass carp fingerling (20-40 g body weight) were
 obtained  from local suppliers (Guangzhou, P.R. China) and maintained in recirculating, indoor  250 L
 aquaria at room temperature (19-24 °C) and a  natural photoperiod. APO, L-dopa, a-MPT, 6-OHDA
 epinephrine and norepinephrine (all purchased from Sigma, St. Louis, MO) were dissolved in 0.7% NaCl
 with 0.1 % sodium metabisulphite. Drugs were injected into the caudal vasculature with a 25-gauge needle
 attached to a 1.0 ml disposable syringe, and then blood was serially sampled at 2, 6 and 24 hr post
 injection. Blood samples were allowed to clot on ice for several hours before the serum was separated
 by centrifugation and stored at -25 °C until hormone measurement.

        GtH concentrations in the perfusates were  determined by radioimmunoassay (RIA) (Peter et al
 1984) specific for Common carp GtH-H (Van Der Kraak et  al., 1992).  GH  concentrations in the
 perfusates and serum samples of Grass carp were determined by RIA specific for Common carp GH
 (Marchant et al.,  1989c).  Serial dilutions of perfusates and serum samples from Grass carp resulted in
 displacement curves parallel to the Common carp GtH and GH standard curves in the respective RIAs
 The hormone responses to each pulse of GnRH and TRH peptide were quantified according to Habibi et
 al. (1989), and expressed as a percentage of the average basal  hormone levels (prepulse) in the three
 fractions collected immediately preceding each pulse (i.e. % of prepulse).  Comparisons of hormone
 responses between different treatments were made by Student's t-test. Comparisons of hormone responses
 to different pulses of GnRH or TRH within an experiment were made using one-way analysis of variance
 followed by Duncan's multiple range test.  The half-maximal  effective dose (ED50) levels of GnRH-A
 and TRH in stimulating GH release were calculated according to Karber (1931).

                               RESULTS AND DISCUSSION

 THE EFFECTS OF GnRH PEPTIDES ON GH RELEASE

       All of the peptides tested (LHRH, (D-Ala6, Pro9-NEt)-LHRH, (D-Trp6, Pro9-NEt)-LHRH, chicken
 GnRH-n, sGnRH and its analogs (D-Arg6, Pro9-NEt)-sGnRH and (D-Ala6, Pro9-NEt)-sGnRH) (1,10 and
 100 nM), were found to stimulate dose-dependent increases in GtH and GH release from in vitro perfused
 pituitary  fragments of Common  carp  (Fig.  1).   Among  the GnRH  peptides  tested, (D-Arg6,
 Pro9-NEt)-sGnRH (sGnRH-A) was the most effective stimulator of GH and GtH release.  However
 sGnRH-A stimulated GtH release more  effectively than GH release.   Chicken GnRH-II,  sGnRH and
 (D-Ala6, Pro9-NEt)-sGnRH  caused comparable stimulation of both GH and GtH release. Compared to
 sGnRH and its analogs, LHRH and its analogs were consistently less potent in stimulating GH and GtH
 secretion in  this  in  vitro perfusion  system  (Table  1).   Intraperitoneal injection of (D-Ala6
Pro9-NEt)-LHRH (LHRH-A) (0.01, 0.001 ug/g body weight) also significantly increased the circulating
 GH levels  of Grass carp fingerling (Fig. 2).

-------
                      b age d e f
                4001-
m a

II200
v fc.
e. B.
                        JL
                           L.
                                    .
                             b egad f e


                              JL
                                         JL
                                                          a eGnRH (Trp"?Leu8-LHRH)

                                                          b sOnRH-A (D-Arg6Fro9Net-sGnRH)

                                                          c B-Ala6Pro9NEt-6CnRH

                                                          e LHRH

                                                          d LHHH-A (»-Ala6Pro9NEt-LHRH)

                                                          f D-Trp6Pro9KEt-LHRH

                                                          g cGnRK-II (H
_L



c  d|e|f  g 	la b 7|d[rhHg|
                          lOOnM
                                               lOnM

                                        GnRH Peptides
                                                                  InK
                          c a f d e
400

ST3oo
C 0
llaoo
«J O
c: t-
li100

beg
J_

. pl-

•

2.



b
T
JL

c
f pi-
Tjf
-1-

e


c


b

J_^
c

a f d e

b c a g f d e


r^—i T
. .rHj _JU,
d e f g (T|b c idTS-fffgl
i i i 	 1
lOOnH lOnH ' InH
GnRH Peptides
Figure 1.  GH and GtH release from perfused pituitary fragments of common carp in response to different
GnRH peptides administered in 2-min. pulse manner. Each value represents meanj+SE of 4 observations
from 4 separate experiments. The average basal GH and GtH levels of the first prepulses were 191.4+88.0
ng/ml and 42.8+.16.7 ng/ml, respectively. The letters in the column represent each treatment group in the
range test, and  groups with common underscoring are  not  significantly different (P>0.05, Duncan's
multiple range test).

-------
                        60
                        50
                      E
T
                        30
                         ol
Figure 2. Effects of intraperitoneal injections of (D-Ala6,Pro9-NEt)-LHRH(LHRH-A) on serum GH levels
of grass carp fmgerling. A. Controls; B. LHRH-A 0.001 ug/g body weight/week for six weeks; C. LHRH-
A 0.01 pg/g body weight/week for six weeks. *Signifcantly higher than controls (P<0.05).

-------
                   Table 1.  Comparison of the relative GH and GtH releasing
                                 potencies of GnRH peptides.
       sGnRH
       (D-Arg6, Pro9-NEt)-sGnRH
       (D-Ala6, Pro9-NEt)-sGnRh
       LHRH
       (D-Ala6, Pro-NEt)-LHRH
       (D-Trp6, Pro9-NEt)-LHRH
       chicken GnRH-II
Relative GH-
releasing potency
       1.00
       4.30
       0.96
       0.31
       0.60
       0.33
       0.89
Relative GtH-
releasing potency
       1.00
       9.24
       1.15
       0.28
       0.35
       0.51
       1.27
       These results suggest that the structural modifications (i.e. GnRH analog A) that increased the
potency of GnRH on GtH release also increased the potency of GnRH on GH release. The results agree
with the findings of Marchant et al. (1989b, c), who found that GnRH peptides function as GH-releasing
factors in goldfish.  Such increases in circulating GH levels may be sufficient to accelerate body growth
in Grass carp fingerling.

THE EFFECTS OF TRH ON GH RELEASE

       Exposure of Common carp pituitary fragments to 5 minute pulses of increasing concentrations of
TRH (0.1, 1.0,  10, 100 and 1000 nM) at 60 minutes intervals resulted in a rapid and dose-dependent
stimulation of GH secretion so that GH levels were increased (upper  panel, Fig. 3).  The ED50 was
estimated  to be 9.7±2.3 nM by using  10 doses of TRH in  order of  increasing  and decreasing
concentrations.  TRH did not effect GtH release (lower panel, Fig 3). These results indicated that the GH
response to TRH was specific, and that TRH may act as a GH-releasing factor in carp.

THE EFFECTS OF CATECHOLAMINERGIC DRUGS ON GH RELEASE

       Actions of Apomorphine (APO)

       APO, a dopamine agonist, significantly increased basal  GH release but decreased basal GtH
secretion from perfused pituitary fragments of Common carp in a dose-dependent manner (10, 100 and
1000 nM) (Fig. 4). Intraperitoneal injection of APO in Grass carp fingerling significantly increased serum
GH levels after 2 and 6 hours, while the GtH level remained unchanged (Fig. 5).
APO acts similarly to dopamine, which inhibits basal and GnRH-induced GtH release in goldfish and carp.
The results of this study therefore, provide further evidence that the release-inhibitory factors regulating
GH and GtH secretion in goldfish and carp appear to be separate  and distinct.

       Actions of 6-hydroxydopamine (6-OHDA)

       Intraperitoneal injection of 6-OHDA (20 ug/g body weight), a neurotoxin for catecholaminergic
neurons (Zambrano et al., 1975), significantly lowered serum GH concentrations of Grass carp fingerling
6 and 24 hours  after injection, compared with control fish (Fig. 6).

-------
                                                                  TRH
                          90
                                             Fractions
 
-------
                        ol
                        	 a>
250


200




100


 SO
                                    abed
                                                   J-.
                                                           I
                                              10     100    1000
                                             APO(nM)
Figure 4.  Influence of APO (10, 100, 1000 nM) on basal GH release from perfused pituitary fragments
of common carp.  The basal GH levels v/ere expressed as the average hormone levels over the initial 15-
min. period before exposure of the fragments to the peptides.  Each value is mean + SE of 4 observations
from 4 individual perfusion runs.  The average pretreatment  GH level was 94.0+18.1 ng/ml. The letter
in the column represents  each experiment group in  the  range  test,  and groups without common
underscoring are significantly different (P<0.05,  Duncan's multiple range test).
        Actions of L-8-dihydroxyphenylalanine (L-dopa)

        Serum  GH levels of Grass carp fingerling were significantly higher 2 hours after intraperitoneal
injection of L-dopa, the metabolic precursor of dopamine, (10 and 100 ug/g body weight) compared with
control fish.  However, at 6 hours after injection, serum GH secretion returned to the basal level (Fig. 7).

        Actions of a-methy-para-tyrosine (a-MPT)

        Injection of 50 ug/g body weight a-MPT, an inhibitor of L-dopa  and catecholamine synthesis,
significantly decreased serum GH levels in Grass carp fingerling 24 hours post injection (Fig. 8).

        Actions of epinephrine (E)

        Injection of 100 ug/g body weight epinephrine significantly decreased serum GH levels in Grass
carp fingerling 2 and 6 hours post injection. Serum GH levels in Grass carp fingerling injected with a
low dose of epinephrine (10 ug/g body weight) were also significantly depressed six hours after injection
compared with control fish (Pig. 9).

        Actions of norepinephrine (NE)

        Serum GH levels in Grass carp, fingerling injected with NE at either a low dose (10 ug/g body
 weight) or a high dose (100 ug/g body weight) were significantly decreased at 2 and 6 hours post injection
 compared with control fish (Fig.  10).

-------
                                               controls
            OH
            c
            E
                50
                40
                30
                20
APO,  20 ug/g  b.w. •-
                 0 :£	L
                                     Hours postinjection
                                                                                  24
 Figure 5. Effects of intraperitoneal injections of APO on serum GH levels of fmgerling of grass carp
 *Sigmficant higher than controls (P<0.05).
               30
          E


          C
               20
         CO
                                 -o controls
                                 "• 6-OHDA, 20  ug/g b.w.
                                      hours postinjection
Figure 6.  Effects of intraperitoneal injections of 6-hydroxydopamine (6-OHDA) on serum GH levels of
lingering of grass carp.  *Significanfly lower than controls (P<0.05).
                                          8

-------
         85
         75
         65  -
         55  _
     E
    CO
45  -
         35  -
                                                         controls
                                                       o L-dopa, 10  ug/g b.w.

                                                         L-dopa, 100 ug/g b.w.
                                                                              I
                                 hours postinjection
Figure 7. Effects of intraperitoneal injections of L-6-dihydroxyphenylalanine (L-dopa) on serum GH levels
of fmgerling of grass carp. ^Significantly higher than controls (P<0.05).

-------
                            -o controls
         s
        c
        e


        M
        o

       OT
            30 t-
            25
            20
15  -
             ot_
                            -•o^-MPT 50 ug/g b.w.
                                   Hours postinjection
                                                                               24
Figure 8. Effects of intraperitoneal injections of a-methyl-para-tyrosine (oc-MPT) on serum GH levels of

fingerling of grass carp.  *Significantly lower than controls (P<0.05).
                                          10

-------
   E
   (-1
   (U
   co
105

 95

 85

 75


 65

 55

 45

 35

 25
  0
                                Hours  nostinjection
  controls
o E,  10 ug/p. b,w.
• E,  100 uR/g b.w.
                                                                             24
Figure 9.  Effects of intraperitoneal injections of epinephrine (E) on serum GH levels of fingering of grass
carp. *Significantly lower than controls (P<0.05).
                                            11

-------
         70
         60
   00

   c
  e

  r-i
  
-------
       The results of the present study support the findings using goldfish (Chang et al., 1985), which
indicated that catecholamines were involved in the neuroendocrine regulation of GH release in teleosts.
Intraperitoneal injection of 6-OHDA, a catecholaminergic neurotoxin, or a-MPT, a catecholamine synthesis
inhibitor, decreased GH levels.  Intraperitoneal injection of L-dopa increased serum GtH concentrations,
while intraperitoneal injection of APO, a dopamine agonist that crosses the blood-brain barrier, increased
serum GH levels.  Finally, serum  GH concentrations were decreased by intraperitoneal injection of
norepinephrine (NE) and epinephrine (E).

                                  ACKNOWLEDGEMENTS

       This work was supported by a grant 3-P-87-1028 from the International Development Research
Centre of Canada to H.R. Lin and R.E. Peter.

                                        REFERENCES
Chang,  J.P.,  T.A. Marchant,  A.F.  Cook,  C.S.  Nahorniak and  R.E. Peter.   1985.  Influences of
       catecholamine  on  growth  hormone  release  in  female   goldfish,   Carassius   auratus.
       Neuroendocrinol. 40: 463-470.
Habibi, H.R., T.A. Marchant, C.S. Nahirniak, H. Van Der Loo, R.E. Peter, J.E. Rivier and W.W. Vale.
       1989.  Functional relationship between receptor binding  and biological activity for analogs of
       mammalian and salmon gonadotropin-releasing hormones in the pituitary of goldfish (Carassius
       auratus).  Biol. Reprod. 40:  1152-1161.
Karber, G. 1931.  Beitrag zur kollektiven behandlung pharmakologischer reihenversuche.  Arch. Exp.
       Path. Pharmak.  162:480-484.
Mackenzie, O.S., R.R. Gould,  R.E. Peter, J.E. Rivier and W.W. Vale.  1984. Response of supervised
       goldfish pituitary fragments  to mammalian and salmon gonadotropin-releasing hormones. Life
       Sci. 35: 2019-2026.
Marchant, T.A., LG. Dulka and R.E. Peter.  1989a.  Relationship between serum growth hormone levels
       and the brain and pituitary content  of immunoreactive somatostatin in the goldfish,  Carassius
       auratus.  Gen. Comp. Endocrinol. 73: 458-468.
Marchant, T.A., J.P. Chang, C.S. Nahorniak and R.E. Peter. 1989b. Evidence that gonadotropin-releasing
       hormone also functions as a growth hormone-releasing factor in the goldfish. Endocrinology 124:
       2509-2518.
Marchant, T.A. and R.E. Peter.  1989c.  Hypothalamic peptides influencing growth hormone secretion in
       the goldfish, Carassius auratus.  Fish Physiol.  Biochem. 7: 133-139.
Peter, R.E., C.S. Nahorniak, J.P. Chang and L.W.  Crim.  1984.   Gonadotropin  release from the pars
       distalis of goldfish, Carassius auratus, transplanted beside the brain or into the brain ventricles:
       Additional evidence for gonadotropin-release-inhibitory  factor.  Gen.  Comp. Endocrinol. 55:
       337-346.
Van Der Kraak, G., K. Suzuki, R.E.  Peter, H. Itoh and H. Kawauchi.  1992.  Properties of common carp
       gonadotropin I and gonadotropin II. Gen. Comp. Endocrinol. (in press).
Zambrano, D. 1975.  The ultrastructural catecholamine and prolactin contents of the rostral pars distalis
       of the fish Mugil platanus after reserpine or 6-hydroxydopamine administration.  Cell  Tiss. Res.
        162: 551-563.
                                             13

-------
                 EVALUATION OF REPRODUCTIVE EFFECTS
                  OF ANTHROPOGENIC CHEMICALS IN FISH

                     William H. Benson1, Thomas Dillon2 and Burton Suedel3

                                      INTRODUCTION

        Exposure of fish to toxic chemicals may yield effects varying from acute lethality to sublethal
 changes in reproduction, growth and development (Murty, 1986).  A sublethal effect that warrants
 more detailed study is the influence of environmental exposure to anthropogenic agents on
 reproduction and offspring growth (Weis and Weis, 1989). Reproductive effects may arise in fish
 when some aspect of the environment alters the maternal contribution to yolk quality or quantity.
 With regard to anthropogenic agents, biologically important effects could arise through their direct
 incorporation into the yolk or through alteration in the egg yolking process. The manifestation of
 these subtle influences on reproductive biology could range from altered life histories of the  affected
 generation to teratogenic effects.

        There is evidence that maternal environment influences reproduction and offspring quality.
 Classic experiments documenting maternal effects have been conducted with cattle, rabbits and pigs
 (Pirchner, 1983) snails, fruit flies, rye grass, and tobacco (Mather and Jinks, 1982).  Maternal effects
 are implicated in all reproductive modes, from live-bearing with extensive maternal-fetal exchange to
 egg or seed production. Maternal effects arise through contributions that the mother makes to her
 offspring's phenotype beyond her nuclear genes.  Routes for these contributions are: cytoplasmic
 inheritance, pre-and post-natal nutrition, pre- and post-natal transmission of pathogens by feeding,
 imitative behavior, and interaction of sibs directly with each other or through the mother (Mather and
 Jinks, 1982). Maternal effects may arise from non-genetic or genetic sources. For example, non-
 genetic effects may arise from the environment in which the mother finds herself and a genetic source
 could be genetically based variation in maternal nourishment of offspring.  Males can influence the
 maternal genetic effects of their grandchildren through their daughters, despite having no impact on
 their own offspring through this route.  "Grandfather effects" such as these have been documented in
 species as diverse as horses, cattle (Pirchner, 1983), snails (Boycott and Diver, 1923; Sturtevant, 1923;
 Diver et al., 1925)  and mosquitofish (Reznick, 1981).

           Maternal environmental effects in fish on the size and number of eggs or offspring
 produced have been documented in several cases (e.g., Aim, 1949; Reznick, 1981; Cheong et al.,
 1984; Tanasichuk and Ware, 1987). Also, many  cases are known where chemicals of environmental
 interest induce reproductive  effects (Niimi, 1983; Heath, 1987; Weis and Weis, 1989). Such research
 has primarily focused on the reproductive effects of chemicals on fecundity and fertility though other,
 possibly subtle, effects of sublethal concentrations are likely.
    'Department of Pharmacology and Research Institute of Pharmaceutical Sciences, School of Pharmacy,
The University of Mississippi, University, MS 38677, U.S.A.

    2U.S. Army Corps of Engineers Environmental Laboratory, 3909 Halls Ferry Road, Vicksburg, MS
39180-6199, U.S.A.

    Department of Biology, The University of Mississippi, University, MS 38677, U.S.A.

                                           14

-------
       The purpose of this investigation was to evaluate the relationship between reproductive effects
and polychlorinated bipheriyl (PCB) tissue concentrations resulting from exposure to contaminated
sediments from an inland waterway.  Such "residues-effects" data are infrequently (ca., 6%) reported in
the literature for fish or other aquatic organisms (Dillon, 1984). However, these types of data may be
valuable when interpreting the biological significance of contaminant tissue concentrations in field-
collected animals.

                               MATERIALS AND METHODS

SEDIMENTS

       The PCB  contaminated sediments were obtained from existing sediment samples previously
collected from Sheboygan Harbor, Wisconsin and stored in 15-L glass jars at 4°C.   Sediments were
mixed to produce treatment concentrations identified as low, medium and high which corresponded to
PCB concentrations, as Aroclor 1254 equivalents, of 0.82, 14.0, and 27.0 |ag/gm (dry weight),
respectively.

EXPOSURES

       Adult fathead minnows (Pimephales promelas) were obtained from Northeastern Biologists
(Rhinebeck, NY).  Females averaged 25 to 40 mm in length, and males averaged 40 to 50 mm in
length.  Following a 30 day acclimation to all experimental conditions except sediment, fish were
placed in 40 L glass aquaria containing a 2 to 4 cm layer of PCB contaminated sediment. Aquaria
containing only water  served as controls.  The aquaria were supplied with moderate aeration and a
continuous flow of aged, charcoal filtered tap water having the following mean water quality
characteristics: pH 7.8, hardness 59 mg CaCO3/L and 8.1 mg/L dissolved oxygen.   Flow rates (80
ml/min) were set to provide three volume additions every 24 h.  Temperature and flow rates were
monitored daily while total dissolved solids (TSS) was monitored on a weekly basis.

        During the initial experimental phase, 7 male and 21 female fish per aquarium were exposed
for 5 weeks with water temperature maintained at 20° + 1°C and photoperiod maintained at 12-h
light: 12-h dark.  Fish were fed 3% freeze-dried brine shrimp in the morning and 3% commercial diet
(Tetra-Min) in the evening, on an estimated body weight basis. Following the initial exposure phase,
water temperature was increased  1°C per day to 26° + 1°C.  In addition, the photoperiod was
lengthened 1  h every other day to 16-h light:8-h dark to induce gametogenesis, sexual dimorphism and
reproductive  activities in the test organisms. During this 10-d induction phase, spawning substrates
were introduced into each of the  aquarium.  The use of frozen brine shrimp (3%) replaced freeze-dried
brine shrimp  to further induce spawning.  Following induction, 1 male and 6  female fish were selected
from each aquarium to evaluate treatment effects on reproduction. Fecundity and clutch size were
monitored for an additional 9 weeks.

CHEMICAL ANALYSIS

        Chemical analysis was conducted by the Tennessee Valley Authority (Chattanooga, TN).
Briefly, fish  were thawed, homogenized in petroleum ether and cleaned with concentrated sulfuric
acid. Water  was removed during repeated extractions with sodium  sulfate and an aliquot of the
homogenized fish tissue was removed for gravimetric determination of percent lipid prior to acid clean
up.  Extracts were run on a Hewlett-Packard gas chromatograph equipped with a 30 rneter DB-5 fused
capillary column and electron capture detector.  All standards were  purchased from Ultra Scientific
(Hope, RI).  The detection limit for Aroclor 1254 was 0.10 jjg/g and for PCB congeners, 0.01

                                            15

-------
 DATA ANALYSIS

        Treatment effects on all parameters were analyzed by means of one-way analysis of variance.
 Differences were considered statistically significant at p < 0.05. Square root or log 10 transformations
 were used when data sets were not homogenous.  Arc sine transformations were used for non-
 homogeneous data which were expressed as percentages.  Mean separation for homogeneous data was
 achieved by means of Waller-Duncan k-ratio t test.  When transformations were unsuccessful in
 achieving homogeneity a Proc Rank nonparametric procedure was used for mean separation (SAS
 Institute, 1985). To facilitate statistical analysis and presentation of the PCB analytical information,
 data reported  to be less than the detection limit were considered to be equal to the detection limit.

                                           RESULTS

        Table 1 presents the major findings of the investigation. Survival was high in all PCB-
 contaminated sediment treatments and ranged from 80 to  100%. Regarding reproductive effects, there
 was a very clear separation between the control and low treatments and the medium and high
 treatments; with fish reproduction in the latter treatments being almost completely inhibited.  This is
 substantiated by examination of fecundity and frequency of egg production.  The mean number of eggs
 produced by fish in the control, low, medium and high treatments were 5,166,  3,346, 1,172 and 97,
 respectively, while mean number of clutches were 22.4, 14.4, 3.2 and 0.6.

               Table  1.  Survival, fecundity and frequency of egg production in fathead
                    minnows exposed to  PCB-contaminated sediment treatments
Survival (%)
Treatment
Control
Low
Medium
High
male
100
100
80
80
female
100
100
95
90
Fecundity
(total eggs)
5,166 + 1,152 A
3,346+ 417 A
1,172 + 1,114 B
97+ 9B
Frequency
(total clutches)
22.4 + 3.3 A
14.4 + 2.8 A
3.2 + 3.0 B
0.6 + 0.6 B
               Values are reported as mean + SE. Values within columns followed by
               the same letters are not significantly different (p < 0.05).

       Tissue concentrations in fish from all experimental treatments were significantly different from
one another after 7 and 16 week of exposure (Fig. 1).  After 7 weeks of exposure, respective mean
tissue concentrations in the control, low, medium and high treatments were 0.10, 5.25, 13.7 and 18.4
ug/g wet weight, expressed as Aroclor 1254 equivalents.  PCB residues were also determined in fish at
termination of the  experiment. These residues, therefore, reflect 16 weeks of continuous exposure to
the experimental treatments with the last 9 weeks under conditions to induce gametogenesis and
spawning activities.  Mean values after 16 weeks of exposure generally reflected patterns observed
after 7 weeks except that the absolute concentrations were elevated over 2 fold.  Significantly different
PCB concentrations in the control, low, medium and high treatments  were 0.10,  11.6, 36.0 and 47.2
pg/g wet weight, respectively.

       Mean percent lipid values for all samples from this experiment are shown in Figure 2. At 16
weeks of exposure there is a significant increase in percent lipid in the medium and high treatments.

                                            16

-------
                                                                      I
                                                                      w
 
                   CO
                                                                       IP
                                                                       § <=>
                                                                       Iv
                                                                       2 &
            e

            1

            1
                                                                         ••a
            •of
            £ §
            c« o
            C2 ;Pi

            1 M
            « '«

            ?i
            a
            2
            II
            v—.  rt
            c- f
            \* -o
            O 
-------
              ca
 3   w
 to   O
 o   a,
 cu  x

 M   W
     tn
     I
a
                                                       o
                                                       HH
                                                       ffi

o   u
g   «
s   s
o   ^
      CO
                                                                   I
                                                                   C/D

                                                                   g,
                                                                   vj

                                                                   £
                                                                   X3
                                                                   W

                                                                   W >O

                                                                   § o
                                                                   c
                                                                   cd cj
                                                                   .s «
                                                                   § -a
                                                                   rt W)
                                                                   G 
                                                                   O +J

                                                                   " §
                                                                   2 g
                                                           Q
     r^(     -O >
     HH     S> ,2
w  cS

|  I

a -a"'
'5  o
E -c
•o  v


ii
eg  55
T]  o

o  &
                                                                   C
                                                                   8
                                                                  « ^
               awn
                              18

-------
                                        DISCUSSION
                                                    ?
       One of the objectives of this investigation was to evaluate the biological consequences of
bioaccumulation.  To optimize this comparison one must have graded responses in both the
consequences and the bioaccumulation components.  In this investigation, both were observed.

       The experimental treatments, represented by control, low, medium and high PCB-contaminated
sediments, had a significant impact on reproduction fathead minnows. This is based on the
observation that both fecundity and fertility were affected by these compounds in a dose-related
fashion.  Interestingly, fish exposed to PCBs at medium and high concentrations had a higher percent
lipid content than those not exposed or exposed to low concentrations.  This brings up the interesting
prospect that the energy allocation pattern of exposed fishes may have been altered from reproduction
to storage.  PCBs are lipophilic, suggesting that lipid stores sequester concentrated levels of these
compounds. Thus, even if environmental concentrations were to decrease, when and if these stores are
metabolized, they could release high levels of PCBs into the organism. One characteristic of PCB
activity is estrogenicity (Hansen, 1987).  This property makes  it likely that these compounds influence
reproductive function. However, in addition to reproductive and hormonal effects, PCBs have a
multiplicity of other toxicological effects including neurotoxicity, immunotoxicity, hepatotoxicity, as
well as mutagenic and carcinogenic effects (Safe, 1984). Furthermore, commercial PCB products,
such as Aroclor 1254, are mixtures of individual biphenyl molecules; or congeners.  There are 209
theoretically possible congeners and recently, much attention has been given to their individual
toxicological effects. For example, Dillon et al. (1989, 1990)  examined the effects of individual PCB
congeners on survival, growth and reproduction in the freshwater cladoceran, Daphnia magna  (IUPAC
numbers 52, 77, 101, 118, 138,  153, 180) and freshwater fish, Pimephales promelas (IUPAC numbers
52, 101, 138, 153, 180). These species accumulated substantial amounts  of each congener examined.
However, there were little to no detectable effects on survival, growth and reproduction. These
investigators indicate that because the concept of PCB toxicity has been developed primarily through
technical findings with mammals, caution should be exercised when applying those concepts to fish
and other aquatic organisms.

                                  ACKNOWLEDGEMENTS

        Funding for this work was provided by the U.S. Army Corps of Engineers Long-Term Effects
of Dredging Operations Program (LEDO).

                                        REFERENCES

Aim, G. 1949. Influence of heredity and environment on various forms of trout. Instit. Freshw. Res.
        Drottningholm, Fisheries Board of Sweden, Report No. 29:29-34.
Boycott, A. E. , and C. Diver.  1923. On the inheritance of sinistrality in Limnaea peregra. Proceedings
        of the Royal Society of London(B) 95:207-213.
Cheong, R. T., S. Henrich, J. A. Fair, and J. Travis. 1984. Variation in fecundity and its relationship
        to  body size in a population of the least killifish, Heterandria formosa (Pisces: Poeciliidae).
        Copeia 1984:720-726.
Dillon, T.M. 1984. Biological consequences of bioaccumulation in aquatic animals: An assessment of
        the current literature. Technical Report D-84-2. U.S. Army Engineers Waterways Experiment
        Station, Vicksburg, MS.
                                             19

-------
 Dillon, T.M., W.H. Benson, B.S. Suedel, and W.D. Burton. 1989. Acute and chronic toxicity of PCB
        congeners to aquatic organisms.  Abstract. Paper presented at Tenth Annual Meeting of Society
        of Environmental Toxicology and Chemistry, Toronto, Ontario, Canada.
 Dillon, T.M., W.H. Benson, R.A. Stackhouse, and A.M. Crider. 1990. Effects of selected PCB
        congeners on survival, growth and reproduction in Daphnia magna. Environmental Toxicology
        and Chemistry 9:1317-1326.
 Diver, D. A. E. Boycott, and S. Garstang. 1925. The inheritance of inverse symmetry in Limnaea
        peregra. Journal of Genetics 15:113-200.
 Hansen, L. G. 1987. Environmental toxicology of polychlorinated biphenyls. In S. Safe and O.
        Hutzinger, eds., Polychlorinated Biphenyls (PCBs): Mammalian and Environmental
        Toxicology. Springer-Verlag, New York, NY, pp. 15-48.
 Heath,'A. G. 1987. Water Pollution and Fish Physiology. CRC Press, Boca Raton, FL.
        Mather, K., and J.  L. Jinks. 1982. Biometrical Genetics. Chapman and Hall, New York, NY.
 Murty, A. S. 1986. Toxicity of Pesticides to Fish. Vol.  II. CRC Press, Boca Raton, FL.
 Niimi, A. J. 1983. Biological and toxicological effects of environmental contaminants in fish and their
        eggs. Canadian Journal of Fishieries and Aquatic Sciences 40:306-312.
 Pirchner, F. 1983. Population Genetics in Animal Breeding, 2nd Ed. Plenum Press, New York,  NY.
 Reznick, D. N. 1981. "Grandfather effects": the genetics of interpopulation differences in offspring size
        in the mosquitoflsh. Evolution  35:941-953.
 SAS Institute. 1985. SAS User's Guide: Statistics, Version 5 Ed., Gary, NC.
 Safe, S. 1984. Polychlorinated biphenyls  (PCBs) and polybrominated biphenyls (PBBs): Biochemistry,
        toxicology, and metabolism of action. CRC Critical Reviews in Toxicology 13:319-395.
 Sturtevant, A. H.  1923. Inheritance of direction of coiling in Limnaea. Science 58:269-270.
 Tanasichuk, R. W., and D. M. Ware. 1987. Influence of interannual variations in winter sea
        temperature on fecundity and egg size in Pacific herring (Clupea harengus pallasi). Canadian
        Journal of Fisheries and Aquatic  Sciences 44:1485-1495.
Weis, J. S., and P. Weis. 1989. Effects of Environmental Pollutants on Early Fish Development.
        Reviews in Aquatic Sciences 1:45-73.
                                           20

-------
                SUBLETHAL RESPONSES OF TWO LARVAL FISHES
                            TO RICE FIELD PESTICIDES

                         Alan G. Heath1'2 and Joseph J. Cech, Jr.1

                                      ABSTRACT

       Newly hatched striped bass, Morone saxatilis, and Japanese medaka, Oryzias latipes,
were exposed to "low" (environmental) and "high" (Vz LC50) sublethal levels of methyl
parathion, carbofuran, molinate, and a combination of the three. Dry weight, RNA:DNA
ratio, swimming performance, spontaneous activity, and acetylcholinesterase activity were
measured immediately after the 4-d exposures and, again, after 10 subsequent days in clean
water.  Whereas pesticides had little effect on growth variables, acetylcholinesterase and
swimming performance were affected, especially by  the high doses in striped bass. Growth
and movement variables of striped bass were greatly affected by food availability.  "High"
sublethal pesticide concentrations and/or low food availability would contribute to a weakened
larval striped bass in the Sacramento-San Joaquin Estuary.

                                   INTRODUCTION

       Populations of striped bass, Morone saxatilis, an introduced predatory fish in the
Sacramento-San Joaquin Estuary of California, USA, have undergone significant reductions
over the past 15 years (Setzler-Hamilton et al. 1988).  A valuable sportsfishing industry for
this species has been severely damaged by this decline.  Two possible reasons for the decline
include: water diversions (and consequent loss of drifting striped bass larvae, as well as larval
food where water is diverted from the River and Estuary for irrigation; and pollutants
(resulting in poor survival of sensitive larvae or their prey. Unfortunately, striped bass
spawning in the Sacramento River occurs at the same time and place when various
insecticides and herbicides are flushed from surrounding  rice fields into agricultural drains
(such as the Colusa Basin Drain which flow into the Sacramento River. In low-runoff years,
drain water can comprise up to 40% of the River flow at this point.  The striped bass larvae
drift in the river for 4-8 days until the water is considerably diluted by flows from the San
Joaquin River and by tidal influences of saltier water in the estuary.  Our research objective
was to assess the effects of two insecticides and an herbicide using both molecular and
whole-animal techniques on the growth and motor function of 4-10 day-old striped bass
larvae and to compare these effects with those shown by Japanese medaka, Oryzias latipes,
This comparison tested possible use  of medaka as a surrogate species for striped bass in
future larval toxicity studies. Medaka are much smaller  as adult fish and can be spawned in
the laboratory at any time, rather than only in May and early June when striped bass  spawn.
    Department of Wildlife and Fisheries Biology, University of California-Davis, Davis, California
95616, U.S.A.

    2Usual address: Department of Biology, Virginia Polytechnic Institute and State University,
Blacksburg, Virginia 24061, U.S.A.

                                         21

-------
                                      METHODS

       We used methyl parathion, an organophosphate insecticide; carbofuran, a carbamate
insecticide; and molinate, a thiocarbonate herbicide; and a combination of all three chemicals
at two sublethal dosages. All were checked by actual gas chomatographic measurements in
the California Department of Fish and Game Water Pollution Control Laboratory.  The high
dose approximated one-half of the larval striped bass LC50 for each chemical,  whereas these
concentrations were divided by 3 where they were mixed together for the combined high
dose. The low dose approximated concentrations in the Colusa Basin Drain water. There
were up to several orders of magnitude differences within dosages (Table 1).  Larvae were
exposed 4 days to these concentrations in 2-1 beakers with daily water changes with new
chemical.  Pesticide exposures were conducted at the California Department of Fish and
Game Aquatic Toxicology Laboratory and simulated their 4 days  in the river.

       Larvae were carefully transported to the University of California Department of
Wildlife and Fisheries Biology and reared in clean water for 10 days to simulate tune in the
estuary. Larvae were held in polypropylene beakers with Nitex mesh panel sides, which were,
randomly assorted in a temperature and solute-constant water bath.  Larvae inflate their
swimbladders and commence feeding at day 6-8 post-hatch, so we fed  10 ml of a
concentrated Artemia nauplii suspension twice daily in a randomized order. Measurements
related to growth and movement were made at two periods: immediately after exposure to the
chemicals, and after 10 days in clean water.

       Growth was measured by comparisons of dry weights and  by whole larvae RNA:DNA
ratios.  Larvae were fixed in 5% buffered formalin in individual wells of tissue plates, rinsed
with distilled water, dried to a constant weight at 50°C, and weighed on a Cahn microbalance.
RNA and DNA concentrations were measured using a modification of the spectrofluorometric
method of Bentie et al. (1981).

       Motor function was assessed by measurements of spontaneous activity, swimming
performance, and acetylcholinesterase concentrations.  Spontaneous activity was also
measured in clean water,  using a pyrex pan over a paper grid. Twenty seconds after being
placed in the center of the pan, the number of grid lines crossed/minute was counted.  In all
of these measurements, a random number table was used to  determine order of use, because
the slowest swimming individuals might be captured first from the holding beakers with our
plastic pipets.  Also, the person conducting a given test was consistent  across treatment
groups, and a code system was used such that the experimenter did not know  the fish's
treatment group.  Swimming performance was measured in clean water using "racetrack,"
consisting of a plastic petri dish with an inverted smaller dish serving as a centerpiece,
forming a circular aquatic "track.".  An individual larva was randomly selected and after 2
minutes adjustment to the "track", was gently chased with a glass  rod as it might be chased
                                         22

-------
                                      TABLE 1
         Mean (+ SD) pesticide concentrations (ppb) and water quality data ranges
                           during 4-d exposures of fish larvae1
Pesticide Concentrations
Low Low
Nom Meas.
Parathion
Carbofuran
Molinate
Combined:
Parathion
Carbofuran
Molinate
0.66
1.3
70.0
0.66
1.3
70.0
1.04+0.59
1.1+0.2
59.0+2.9
0.55+0.08
1.08+0.18
43.0
STRIPED BASS
High High
Nom. Meas.
1800
110
4050
600
37 .
1350
1650+150
79+7.9
3125+330
470+14
29+4.2
,930+57
Water Quality
Temp. pH
17.0-17.7 8.2-8.7
16.6-17.6 8.0-8.9
17.0-17.6 8.1:8.9
16.7-21.7 7.9-8.4
MEDAKA
Parathion
Carbofuran
Molinate
Combined:
Parathion
Carbofuran
Molinate
0.66
1.3
70.0
0.66
1.3
70.0
3.17+0.15
1.5+0.10
54.0+2.1
0:41+0.01
1.1+0.12
41.0+2.1
1800
110
4050
600
37
1350
' 1567+58
88+14
3001+183
515+10
31+5.6
920+48
21.9-23.3' 8.3-8.6
23.1-23.7 8.0-8.6
22.2+22.7 , 8-2-8.5
23.1-24.5 7.8-8.5 ;•
1 As shown in Heath et al. (1993a, b).
                                        23

-------
by a predator, and the number of radial lines it crossed in 1 minute was recorded.  Larvae were frozen
for acetylcholinesterase activity, which was assayed in Dr. Joseph Zinld's laboratory at the University
of California, Davis, using a modified Ellman et al (1961) procedure.

       Analyses of variance, or Mann-Whitney U-tests, were conducted on the pooled triplicate
beakers for each treatment giving an "n" of 8-16 larvae/treatment for the acetylcholinesterase
measurements, 21 for the spontaneous activity measurements, and 15 for the other data. Further
details on methods are in Heath et al. (1993).
                                RESULTS AND DISCUSSION
STRIPED BASS
       There was very little mortality induced by the pesticide exposures.  However, we did observe
that immediately after the 4-day exposure to methyl parathion, the majority of the striped bass larvae
were swimming on their sides. This behavior persisted through the subsequent 10 days in clean water.

       There was virtually no increase in dry weight over the 10 days in clean water in the methyl
parathion and molinate experiments (Table 2). The fish were observed to be feeding and had food in
their guts, but the 300 um Nitex panels were not effectively holding the nauplii in sufficiently high
concentrations in the beakers. We switched to 200 um mesh  panels for subsequent experiments and
larvae grew faster (dry weight tripled). The less-dense food suspension was adequate for maintenance
needs (preventing starvation, but allowing no growth), and may have more accurately simulated the
estuary under food-limited conditions. Larvae from the low parathion dose had reduced weight
immediately after exposure.  This may have been a statistical  fluke, because they were  still operating
on endogenous yolk - they had not yet begun to feed. But the RNArDNA ratio was also depressed,
suggesting  a depressed protein synthesis rate. Weight of the high-dose molinate-exposed striped bass,
after 10 days in clean water was significantly greater than controls.  The same phenomenon was
measured in medaka (Table 3).  This could represent a type of hormesis  where an organism
metabolically over-compensates to an exogenous stressor (Stebbing 1987).

       The RNA:DNA ratio was also elevated in this group (high molinate). The low  ratios in all
three treatments (including controls) in the 10-day combined pesticide group might reflect an error in
preparation of the nucleic acid standard, because the same standard was used for both dosages and the
controls. Overall, from both weight and RNArDNA ratio data, these concentrations of  these three
pesticides show little evidence of growth suppression.

       Striped bass exposed to methyl parathion at both doses showed a slight but significant
depression  in swimming performance immediately after exposure (Table  2). After the  10 days in clean
water, only the high-dose fish showed a significant depression compared with controls.  As stated
above, most of this swimming was on their sides.  The high dose of molinate also reduced swimming
performance, which persisted after the 10  days in clean water. The larvae with reduced food
abundance  (methyl parathion and molinate-exposed, including controls) all showed depressed
swimming, which suggests that this measurement might be useful as a biomarker for low food
consumption.
                                            24

-------
                                       TABLE 2
Mean (+ SE) effects of 4-d exposure of striped bass to 3 pesticides and combined
pesticides (low and high doses: see Table 1).  Numbers of fish were generally 10-15
(see text).  Asterisks indicate significant statistical difference (p<0.05) from
controls, and "na" indicates not available.
Pesticide/
Dose
SwimPerf. Spont. Act. AchE Dry Wt. RNA/DN;
(lines/min) (lines/mm) (ug/mg prot.) (ug)
METHYL PARATHION
Immediately After Exposure
Control
Low Dose
High Dose
20+1
*16+2
*16+2
53+4
47+4
53+4
487+33
*389+20
*356+29
After a Subsequent 10 d
Control
Low Dose
High Dose
MOLINATE
Control
Low Dose
High Dose
9+1
10+1
*4+2
12+2
12+2
*6+4
585+71
*333+23
*402+11
171+12
*125+10
199+17
2.34+0.11
*1.87+0.05
2.44+0.08
in Non-Contaminated Water
176+39
130+9
132+22
na
na
na
Immediately After Exposure
22+1
22+2
*17+1
36+3
42+4
*28+3
409+19
498+38
465+21
After a Subsequent 10 d
Control
Low Dose
High Dose
10+1
*5+l
*4+l
10+1
12+2
11+1
686+4-9
613+23
*487+24
120+8
118+11
*95+7
2.07+0.22
1.74+0.08
2.11+0.13
in Non-Contaminated Water
120+6
125+6
*153+9
1.46+0.10
1.36+0.15
* 1.92+0. 10
                                             25

-------
 Table 2 (continued)
 CARBOFURAN





 Control



 Low Dose



 High Dose








 Control



 Low Dose



 High Dose
             Immediately After Exposure




 23+2     46+6      286+37    113+5        L91+0.06




 20+2     42+6      253+25    113+5       *2.21+0.09




*17+2     18+5      202+16    113+6         1.850.06
        After a Subsequent 10 d in Non-Contaminated Water




 24+2     35+4      386+22   306+27        2.32+0.06




 27+3     31+4     *671+52   386+51        2.20+0.09




 21+2     41+4     *574+46   312+26       * 1.97+0.10
COMBINED PESTICIDES
Control



Low Dose



High Dose








Control



Low Dose



High Dose
            Immediately After Exposure




 22+2     36+5      289+19    138+8       2.95+0.25




 24+2     49+4       245+8   130+11   '    2.71+0.11




 20+2    *53+4     *212+16   154+13       3.30+0.19




       After a Subsequent 10 d in Non-Contaminated Water




 22+2     25+4      612+58   315+25       1.88+0.15




 20+2     26+5      485+52   343+28       2.05+0.13




 23+3     24+4     *371+20   355+44       2.07+0.08
                                       26

-------
        Spontaneous activity measurements were somewhat variable, including a significant elevation
for combined pesticides (escape behavior?). Striped bass showed more significant reductions at the
high dose, indicating impaired movement, but there were fewer reductions than in their swimming
performance data, indicating that this measure of activity is somewhat less sensitive than swimming
performance to pesticides.  However, it appears especially sensitive to food availability (Table 2).
This result plus the significantly-reduced swimming performance may point to a contributing cause of
the striped bass decline in the Sacramento-San Joaquin Estuary.  Significant reductions in larval food
supplies may make these young striped bass much more vulnerable to predation (Webb 1986).

        Acetylcholinesterase activity showed significant reductions, mostly at the high dose for all of
the chemicals, including the combination (Table 2). Our measured reductions are small, but reflect
whole body levels. Methyl parathion and carbofuran are both known acetylcholinesterase inhibitors
(Zinkl et al.  1991).  Acetylcholinesterase inhibition by a thiocarbamate herbicide is more surprising.
The prolonged effect (at 10 days) may indicate  central nervous system developmental inhibition in
larval striped bass or acetylcholinesterase inhibition from a molinate metabolite, because Tjeerdema
and Crosby (1987) measured molinate's half-life to be 1 day!

        Summarizing the larval striped bass results: exposing these newly-hatched fish to individual
pesticides for 4 days resulting in some sublethal effects, even  at concentrations  several  orders of
magnitude below the LC50.  Some of these effects persisted in larvae kept for 10 days in clean water.
Swimming performance and acetylcholinesterase inhibitions were the most sensitive variables
examined, with the swimming being especially  sensitive to food shortage. Methyl parathion appears to
cause the greatest harm of the 3 pesticides tested.  Combining the  3 produced a less than additive
effect on most of the measured variables.

MEDAKA

        Medaka showed virtually no mortality from pesticide exposures and the subsequent 10-day
clean water period.  They were eager eaters and they gained weight rapidly.  Medaka reached sizes 2-4
times that of the striped bass larvae (Table 3).  The significant increases in day weight after the high
molinate dose was the potential hormesis also observed in striped bass larvae.

        A bad batch of DNAase enzymes from the supplier spoiled some of our RNA:DNA ratio
samples, but like the dry weight data these pesticides at these  concentrations seem to have little effect
on growth (Table 3).

        Medaka swimming performance showed some variability (a large SE for methyl parathion low
dose), but the only consistent effect might have been a significant  reduction after exposure to the high
dose of carbofuran.

        Spontaneous activity showed significant increases immediately after exposure  to molinate  and
to combined pesticides (Table 3).  Interestingly, striped bass showed the same response to combined
pesticides (Table 2), but molinate-exposed striped bass shoed significant depressed spontaneous
activity - the opposite response of medaka.  Medaka acetylcholinesterase showed more effects than the
swimming or spontaneous activity showed.
                                            27

-------
                                        TABLES
Mean (± SD) effects of 4-d exposure of medaka to 3 pesticides and combined pesticides (low
and high doses: see Table 1).  Numbers of fish ranged from 7-15 for acetylcholinesterase
activity measurements to 21 for spontaneous activity measurements (see text). Asterisks
indicate significant statistical difference (p<0.05) from controls, and "na" indicates
not available.
Pesticide/     Swim Perf.   Spont. Act.        AchE
  Dose        (lines/min)   (lines/min)      (ug/mg prot.)
                                               Dry Wt.   RNA/DNA
                                               (ug)
METHYL PARATfflON
                 Immediately After Exposure
Control

Low Dose

High Dose
32+2

26+2

24+3
na

na

na
1651+167

 *664+82

 *604+72
      After a Subsequent 10 d in Non-Contaminated Water

Control        76+5         na            2488+227

Low Dose     97+14         na           *1933+129

High Dose      76+7         na            2022+195
86.2+4.4    2.29+0.32

 71.8+4.    1.79+0.75

73.2+4.9    1.70+0.22



 145+20    5.04+0.41

 172+36    5.68+0.71

172+23    4.69+0.82
MOLINATE
                 Immediately After Exposure

Control        18+1        8+1             1464+164

Low Dose      18+1       10+1             1336+239

High Dose      16+1      *13+1              *942+50

      After a Subsequent 10 d in Non-Contaminated Water

Control        44+4       12+3             1960+134

Low Dose      39+5       12+2             * 1533+69

High Dose      44+4       12+3            *1430+124
98+6
91+3
88+5
545+46
616+61
*719+39
na
na
na
na
na
na
                                         28

-------
Table 3 (continued)
Pesticide/ Swim Perf.  Spont. Act.
 Dose     (line/mm)  (lines/min)
    AchE
(ug/mg prot.)
DryWt.  RNA/DNA
  (ug)        :•   •  •
CARBOFURAN

                Immediately After Exposure

Control        16+1        5+1

Low Dose      13+1        6+1

ffighDose    *11+1        3+1

      After a Subsequent 10 d in Non-Contaminated Water

Control        54+4        7+2            3218+247

Low Dose      54+3        9+2            3773+195

High Dose    *42+3        8+3            3044+170
na
na
na
76+4 •*'"•
68+3
66+3
na
na
na
                     108+8   3.19+0.26

                    111+12   3.35+0.33

                      90+8  *2.40+0.20
COMBINED PESTICIDES

                Immediately After Exposure

Control        14+1        6+1

Low Dose      16+2        5+1

fflgh Dose     12+1      *10+1
         na

         na

         na
      After a Subsequent 10 d in Non-Contaminated Water

Control        36+_5        4+2                  na

Low Dose     45+3        5+2                  na

High Dose     46+3        4+2                  na'
   67+3 '  6.02+0.46

   68+2   4.99+0.34

   66+5   5.98+0.31



   90+9   9.17+1.07

 104+11   8.39+0.45

   88+8   8.00+0.90
                                        29

-------
        In conclusion, we have found that there are some sublethal effects on newly hatched striped
bass larve from exposure to pesticide concentrations several orders of magnitude below the LC50.  Of
the variables measured in the larvae, swimming performance and inhibition of acetylcholinesterase
were the most sensitive to the pesticides. Medaka larvae are clearly less sensitive to these pesticides
(as measured here) than are striped bass larve so their use as a surrogate species is not supported.

                                  ACKNOWLEDGEMENTS

        We thank Drs. Rose Russo,  Dave Randall, Vance Thurston, Lin Haoren, Lin Yu-huan, and
Hong-jun Jin  for inviting JJC to speak at Nanjing University, and we thank the USEPA for travel-
related support. We appreciate the logistical support of Dr. Paul Lutes  and Mr. Bill Bentley of the
University of California Aquaculture and Fisheries Program  and facilities of the California Department
of Fish and Game Aquatic Toxicology Laboratory and the University of California Institute of
Ecology. We  acknowledge the USEPA and the CDF&G for supporting our research. JJC was
partially supported by the USEPA (R819658) Center for Ecological Health Research at UC Davis. We
also acknowledge several University of California undergraduate students who assisted with the data
collection: Mike Steele, Kevin Reimer,  Laurie Martin, Jon Hamm and Scott Cech.

                                       REFERENCES

Bentie, L.A., S. Dutta, and J. Metcoff. 1981. The sequential  enzymatic determination of DNA and
        RNA. Anal. Biochem.  116:5-16.
Bulow, F.J. 1987. RNA-DNA ratios as indicators of growth in fish: A review, pp. 45-64 in R.C.
        Summerfeltad G.E. Hall (eds.)  Age and Growth fo Fish. Iowa State University Press.
Ellman, G.L., K.D. Courtney, O. Andres, and R. Featherstone. 1961. A new and rapid colorimetric
        determination of acetylcholinesterase activity. Biochem. Pharmacol. 7:88-95.
Heath, A.G., JJ. Cech, J.G. Zinkl, B. Finlayson, and R. Fujimura. 1993a. Sublethal effects of methyl
        parathion, carbofuran, and molinate on larval striped bass. Amer. Fish. Soc. Symp. 14:17-28.
Heath, A.G., JJ. Cech, Jr., J.G. Zinkl, and M.D. Steele. 1993b. Sublethal effects of three pesticides on
        Japanese medaka. Arch. Env. Contain. Tox. 25:485-491.
Setzler-Hamilton, E.M., J.A. Whipple, and B. MacFarlane. 1988. Striped bass populations in
        Chesapeake and San Francisco Bays: Two environmentallly impacted estuaries. Mar. Poll.
        Bull. 19:466-477.
Stebbing, A.R. D. 1987. Growth hormesis: A by-product of control. Health Physics 52:543-547.
Tjeerdema, R.  and Crosby, D. 1987.  The biotransformation of molinate (ordram) in the striped bass
        (Morone saxatilis). Aquatic Toxicology 9:305-317.
Webb, P.W. 1986. Locomotion and predator-prey relationships, pp. XX in M.E. Feder and G.V.
        Lauder (eds.) Predator-Prey Relationships. Univ. Chicago Press.  Chicago.
Zinkl, J.G., W.L. Lockhart, S.A. Kenny, and F.J. Ward. 1991. Effects of cholinesterase inhibiting
        insecticides on fish. pp. 233-254. In P. Mineau  (ed.) Cholinesterase-inhibiting Insecticides -
        Impacts on Wildlife and the Environment. Elsevier Science Publ. Amsterdam.
                                            30

-------
            THE EFFECTS OF WATER POLLUTANTS ON SOME
                    PHYSIOLOGICAL FUNCTIONS OF FISH

                               Yiming Huang1 and Weihong Lin1

       China, a nation abundant with fish resources, is developing its own aquaculture industry.
However, fisheries production and indeed the survival of many fish species is threatened today by
pollution from industrial wastes, mine runoff, petroleum, pesticides, etc., and the number of natural
bodies of uncontaminated water are becoming scarce.  The, effects of water pollutants,, on the
physiology of fish is receiving wide-spread attention, and this article is a brief, survey of the most
recent studies on this topic.

THE EFFECTS OF WATER POLLUTANTS ON FISH RESPIRATORY FUNCTION

       Respiration is a major function of the fish gill, and recording opercular movement (respiratpry
frequency) is a basic measure of the effect of pollutants on respiration. It has long been known that
pollutants, and in particular industrial wastes such as from synthetic fabric mills, decrease breathing
frequency to a lethal level.  Ammonia inhibits; respiratory frequency, but is also taken up by the gills
and affects metabolism by its alkalinizing effect on the blood.  Severe NH3 poisoning can result in
respiratory failure and fish mortality (28). Phenol is a common industrial pollutant, and concentrations
of 5 - 50 ppm can increase the respiratory frequency of catfish in a manner directly correlated with
concentration and exposure time.  High concentrations of phenols can lead to whole body muscle
convulsions (18). Water containing 0.005 mg/L phenol can strongly inhibit the respiratory center of
Common carp.  Fresh water carp such as Common parpj Silver carp, and Grass carp are very sensitive
to heavy metal ions (Hg2+, Cu2"1", Ag+) in water (6, 7, 8), and respiration is perturbed at very low
concentrations (0.05-0.1 mg/L). Increasing concentrations lead to coughing and a decrease in
opercular movement.  This may be caused by the heavy metal ions flowing into the gills with
respiratory movement and becoming attached to the gill surface. The  ions may combine with the
mucus excreted by the gill epidermis to form a unsoluble metal-protein compound.  The higher the
concentration of heavy metal ions, the more compound formed, which reflexively increases cough
frequency in an attempt to eliminate the accumulated pollutants and restore gas exchange.  Coughing
is apparently a protective mechanism, occurring almost immediately when fish are exposed to heavy
metal ions, and lasting a prolonged period of time.  High concentrations of these ions severely inhibit
respiratory function and lead to death. The effects of heavy metal ions on fish respiratory function
differ depending upon both the ion and the fish species involved.  Exposure to several metal ions
simultaneously increases" cough frequency to a greater degree than exposure to a single ion, implying
an additive or synergistic effect.  Research on the effects of heavy metals on fish respiratory function
is of both theoretical value and of practical significance in the establishment of water quality criteria to
define safe and allowable limits of pollutants and encourage biological monitoring of water
contamination.

THE EFFECTS OF WATER POLLUTANTS ON FISH BLOOD  AND CARDIOVASCULAR SYSTEM

       There has been considerable research lately on the effects of organic materials (petroleum and
its by-products, pesticides, benzene and its salts) and heavy  metals (chromium, zinc, cadmium) on the
cardiovascular system of fish.  Fish exposed for some time to water with a concentration of 200-400
    Department of Biology, Zhongshan University, Guangzhou, China

                                           31

-------
 mg/L of the toxin aniline or its salts (24) exhibited abnormal behaviour and hyperventilation.  Death
 occurred after 2-3 days exposure. When chronically exposed to a low concentration of aniline
 (0.05-50 mg/L), examination of the blood showed altered erythrocyte size and shape, a decreased
 erythrocyte count and hemoglobin content, and an increased number of leucocytes.  These toxins also
 reduced total blood protein and plasma albumin levels, although the content of gamma-albumin
 increased.  The results indicated that aniline and its salts affected not only blood tissue structure and
 function but also protein synthesis. Heavy metal ions have similar effects. Exposure to 15 ppm
 PbNHS for 90 hr decreased the blood cell count 53%, decreased hematocrit 43% and reduced
 hemoglobin content by 37% (16). Blood neutrophils were altered and increased in number when fish
 were exposed to 50 ppm cadmium for 8 hr (3).  Treatment with chromium changed blood ion
 concentrations and decreased plasma osmotic pressure in a dose-dependent fashion (15).  Exposure to
 0.1 ppm mercury for 10 days resulted in anemia and a decreased hemoglobin content (4).

        We determined biophysical characteristics of blood using hemorheology and blood from Silver
 and Variegated carp exposed to industrially polluted water. The characteristics of blood from exposed
 fish were significantly different from those of normal fish  (P<0.05). The most significant changes
 were a decrease in blood viscosity, plasma osmotic pressure, hematocrit, and erythrocyte sedimentation
 rate, and an increase of red cell electrophoresis rate.  Phenol and rogor had a similar effect on
 Common carp and catfish (19).

        Many pollutants also affect cardiac frequency (10), and our results have shown that the heart
 rate of catfish increases dramatically with increased pollutant concentration. In other experiments,
 free-swimming Common carp were exposed to 4 ppm rogor, and blood pressure chronically monitored
 through a dorsal aortic cannula. Systolic and diastolic pressures, and arterial pulse rate gradually
 decreased after 1,7, 12 and 24 hr exposure. It is clear that many pollutants have serious and
 extensive detrimental effects on the cardiovascular system  of fish.

 THE EFFECTS OF WATER POLLUTANTS ON FISH GASTROINTESTINAL FUNCTION

        The effects  of oral perfusion of different concentrations of DDT and petroleum on the
 gastrointestinal function of cod has been studied by chronically recording electromyogram activity
 (30).  The results showed that even short-term perfusion (2-5 min) by 10"8 M DDT destroyed basic
 electrical rhythm, resulting in an irregular frequency and amplitude.  Short-term perfusion of 50 mg/L
 petroleum can strongly inhibit electrical activity.  However, the long-term effects of DDT and
 petroleum (7  days) are different from the short-term effects.  Gastric muscle electrical  activity began to
 decrease in both frequency and amplitude on the third day of exposure. Amplitude then remained
 stable, while the frequency of activity continued to decline and finally vanished by the seventh day.
 Considering the physiological response to short-term exposure  to these toxins,  the data suggests that
 gastrointestinal  activity could recover if the fish were returned  to clean water.  Injection of atropine
 (0.5-1.0 mg/L) to block M-cholinergic receptors could also eliminate the effects of DDT and
petroleum.  The effects of these pollutants on gastrointestinal activity may be indirect through effects
 on the central nervous system.

 THE EFFECTS OF WATER POLLUTANTS ON THE CENTRAL NERVOUS  SYSTEM
 AND BEHAVIOUR OF FISH

       Many pollutants exert effects directly on the central nervous system. An example is petroleum
and its by-products, which have an anaesthetic effect  Salmon  and perch exposed to 10% phenol lose
their sense  of balance  and swim in a circular, spiral or erratic manner before appearing to enter an
essentially anesthetized state. Fish exposed to acute pesticide poisoning display a similar behaviour.

                                             32

-------
In the 1970s, there was considerable electrophysiological research on the effects of pollutants on the
fish nervous system. DDT has been shown to alter salmon lateral line nervous activity (2), paper mill
effluent perturbs fish olfactory function (26), while lethal levels of phenol and pesticides damage
synaptic transmission in the central and peripheral nervous system of perch (29). There have also
been acute and chronic recordings of both spontaneous and evoked action potentials in the brain
during exposure to toxic pollutants.  Electrical activity of the cerebellar dorsal surface changed when
goldfish were placed in 10'6 M DDT, and the changes in frequency and amplitude were correlated with
the behaviour of the fish.  In the first 25  hr of exposure, the fish kept their balance and there were no
marked changes in electrical activity.  Further exposure resulted in the fish losing their balance, with a
corresponding decrease in the mean amplitude of potentials from 49 jiv to  84 fiv, and a decrease in
frequency from  12.8 Hz to 10.5 Hz.  These changes were reversed when the fish were placed in
uncontaminated  water, with balance recovering in 24-120 hr.  Evoked potentials in the brain can be
used not only as a meticulous  method of studying structure-function relationships in the central
nervous system, but also as both normal  and pathological indicators of different parts of the brain.
Evoked potentials in the shark procerebrum and dorsal surface of the midbrain caused by stimulating
the optic nerve were increased in amplitude by exposure to petroleum (27).  The increase in the
amplitude was dose-dependent, and could be raised by up to 50% in the procerebrum and 25% in the
surface of the mesencephalon.  Removal  of the pollutant led to gradual recovery of the potentials to
control levels.  Functional and structural  analysis of the mesencephalic  surface showed that the
negative and positive waves of an evoked potential reflected the nerve cells original synchronous
activity after the afferent impulses arrived along the optic nerve.  The increase in evoked potential
amplitude suggested that excitability was increased in the dorsal, mesencephalic surface of the visual
nervous system.

        Neuromorphological studies indicated that neurons in the medulla  changed in a manner similar
to that during malnutrition when fish were exposed to acute or sub-acute petroleum poisoning. The
neurons became swollen,  the shape of the nucleus was altered, and cell cytoplasm was absent.  Huso
dauricus exposed to 0.02 mg/L pesticide for 60 days showed swelling in the spinal cord and changes
in the shapes of individual neurons and their microstructure.  Behaviourally, the fish were quiescent
and did not feed.  Post-mortem examination disclosed that the pollutants accumulated primarily in the
brain and liver (14).  Fish exposed to 1 X 10"614C-DDT for 5 min had a concentration of 0.3 X 10"6 in
the brain, while exposure for 1 hr increased the concentration in the brain to 1 X 10~6.   Protein
metabolism, and both neurotransmitter and cholinesterase levels changed in a manner compatible with
the functional state of the fish after exposure to pollutants (20, 21).

        We recently recorded the electrical activity of the maxillary beard afferent nerve of catfish to
 study the effects of pollutants on fish taste response (9).  There was no marked change in electrical
 discharge of the taste-sensitive afferent nerve after direct stimulation of the upper jaw beard with a
 solution of DDT (0.02 ml/L) and CuSO4 (0.02 g/L).  However, after 2 hr  of perfusing the surface, the
 pollutants reduced the intensity of the taste response to such stimuli as extract of earthworm,
 L-arginine and ascorbic acid.  Prolonged exposure to 0.01 mg/L phenol first increased, then  decreased
 the taste response of fish to various stimuli.  A weakened taste response can affect normal feeding and
 thus hinder development  and  survival.

        The active mechanism of pollutants on the central nervous system is specific to the type of
 pollutant.  Oil and its by-products have an anaesthetic effect, while organic pesticides act as an
 anti-cholinesterase.  Phenol inhibits oxyphosphorylation, and organic chlorine pesticides can damage
 synaptic transmission.
                                              33

-------
        Fish behaviour is determined by their environment, and is directly affected by pollution.
 Young salmon and Mirror carp exposed to pesticides were lethargic and maintained  a relatively long
 distance from each other even when in the same shoal (17). Experiments investigating the formation
 and maintenance of conditional reflexes in fish have shown that high-level processing of nervous
 activity was damaged if fish were raised in polluted water. Both the growth rate and survival rate of-
 young sturgeon exposed to 50-70 mg/L petroleum were significantly inhibited (23).  Exposure to lower
 concentrations could increase the survival rate to 100%, but behaviour was still abnormal in that the
 fish had a decreased appetite and a lethargic response to environmental stimuli. If crude petroleum
 was added to the water (30 mg/L), fish feeding reflexes arid defence reflexes were seriously damaged.  -
 Exposure for 1 day to water containing petroleum decreased the feeding reflex in  sturgeon by 80%,
 and in Mirror carp by 100%. If fish were returned to clean water the second day, the defence
 response remained inhibited for that day but returned to control levels on the third day.

        Pollutants also affect the ability of a fish to distinguish different illumination, temperature and
 salinity levels.  The ability of salmon to distinguish light levels decreased linearly, and defence
 reflexes were inhibited after feeding on food containing 30-80% of the LD50 concentration of DDT
 (12).  Salmon exposed to low concentrations of DDT (1 x 10"11) chose a water temperature 2°C lower
 than the water temperature  selected by  control salmon, while exposure to high  concentrations of DDT
 (5 x 10'") led to selection of a water temperature 4°C higher than the temperature selected by control
 fish (13). The ability to distinguish water salinity was also affected in mosquito fish by DDT (5).

        There are sufficient facts to prove that pollutants in the aquatic environment damage normal
 fish physiological functions, and threaten fish survival.  In order to protect fish production, yet
 rationally utilize fish and other aquatic  resources, the active mechanism of pollutants on fish must be
 quickly discerned.

 REFERENCES

 Aubin, A.C. et al., 1969.  The effect of an acute DDT exposure on the spontaneous electrical activity
        of goldfish cerebellum.  Can. J. Zool. 47: 163-166.
 Bahr, T.G. et al., 1971.  Action of DDT on evoked and spontaneous activity from the rainbow trout
        (Salmo gairdneri) lateral line nerve.  Comp. Biochem.  Physiol. A 38: 279-284.
 Gardner, G.R. et al.,  1970.  Histological and hemotological responses of an estuarine teleost to
        cadmium.  J. Fish. Res. Board Can. 27: 2185-2196.
 Gutierrez, M. et al., 1978.  Accumulation and effects  of inorganic mercury in the blood of robolo
        (Dieertrarechus labrax). Invest. Pesq. 42: 317-324.
Hansen, D.J., 1972. DDT and malathion: effect on salinity selection by mosquito  fish. Trans. Amer.
       Fish. Soc. 101: 346-350.
Huang, Y. et al., 1986. The effect of electroplating mill effluents on the cleaning  movement on Mud
       carp (Cirrhitms molitorella). Acta Sci. Nat. Univ. Sunyatseni 1:  103-105.
Huang, Y. et al., 1987. Effects of the heavy metal ions (Cu2+,  Hg2+, Ag+) on respiratory function of
       common carp (Cyprinus carpio, L.) and  crucian carp (Carassius auratus).  Acta Sci. Univ.
       Sunyatseni 4: 80.
Huang, Y. et al., 1988. Effects of the heavy metal ions (Cu2+,  Kg2*, Ag+) on the cough response of
       mud carp (Cirrhinus molitorella). Envir. Sci. 8: 216-222.
Huang, Y. et al., 1989. Effects of several pollutants on gustatory response of catfish (Clarias leather).
       Theses abstract compilation 18th meeting of Chinese Physiology.  150 pp.
Hughes, G.M.  et al., 1977.  The effects of zinc on the cardiac and ventilatory rhythms of rainbow trout
       (Salmo gairdneri Richardson) and their responses to environmental hypoxia.  Water Res. 11:
       1069-1077.
                                             34

-------
Jiesheng, L. and Y. Huang, 1988. Effect of rogor on the blood and function of cardiovascular system
       in common carp (Cyprinus carpio, L.). Chinese J. Appl, Physiol. 4: 67.
Mcnicholl, P.O. et al., 1975. Effect of DDT on discriminating ability of rainbow trout (Salmo
       gairdneri).  J. Fish. Res. Board Can.  32: 785-788.
Ogilvie, D.M. et al., 1965.  Effect of DDT on temperature selection by young Atlantic salmon (Salmo
       salar, L.).  J. Fish. Board Can. 22: 503-513.
Premdas, F.H. et al., 1963.  The uptake and detoxification of 14C-labelled DDT in Atlantic salmon
       (Salmo salar).  J. Fish Res. Board Can. 20: 827-837.
Putte, V.D., 1983.  Respiration and osmoregulation in rainbow trout (Salmo gairdneri) exposed to
       hexavalent chromium at different pH values.  Aquat Toxicol. 2: 99-112.
Srivastava, A.K. et al., 1979.  Blood dyscrasia in a teleost, Colisa fasciatus after acute exposure to
       sublethal concentrations of lead. J. Fish Biol. 14: 199-203.
Symons, Ph.E.K., 1973. Behaviour  of young Atlantic salmon (Salmo solar) exposed to or force-fed
       fenitrothion, an organophosphate insecticide.  J. Fish. Res. Board Can. 30: 651-655.
Wang, D. et al., 1980.  Effect of phenol on respiratory center of common carp.  Thesis abstract
       compilation of 2nd environmental science of universities. 162 pp.
Zheng, M. et al., 1984.  Studies on some biophysical properties of blood in polluted silver carp
       (Hypophlhalmichthys morifrix) and fathead (Aristichthys nobilis). Acta Sci. Nat. Univ.
       Sunyatseni, Suppl. 4, Envir. Sci.  148 pp.
                                             35

-------
        THE PHYSIOLOGICAL RESPONSES OF RAINBOW TROUT
       TO COPPER AND AMMONIA:  MECHANISMS OF TOXICITY
                           AND AMMONIA EXCRETION

                                             by
                                E.W. Taylor1 and R.W. Wilson2


                                     INTRODUCTION

       Estuaries are environments in which large variations in the physical and chemical nature of the
water, such as temperature, salinity and dissolved oxygen, occur both seasonally and tidally.  In
addition to these natural variations, the estuaries of industrialised countries frequently suffer from the
man-made problems of pollution via effluent discharges from refineries, chemical plants and power
stations as well as sewage outfalls. Many anadromous or catadromous fish species such as the
salmonids make their periodic migrations via these polluted estuaries. The present study considered
the responses of a salmonid fish, the rainbow trout, Oncorhynchus mykiss to ammonia and copper, two
pollutants discharged into the Tees estuary in N.E. England. The assessment of the toxicity of copper
and ammonia to fish has been the  subject of numerous studies and reviews (Alabaster & Lloyd, 1980;
Campbell & Stokes, 1985;  WHO,  1986).  Such studies of toxicity and the modulating effects of
various water qualities, such as temperature, pH and relative hardness, are useful in predicting the
toxicity of pollutants in the natural environment but we require a better understanding of the
mechanisms of toxic action, including the physiological changes produced by lethal and sublethal
levels of toxicants.

AMMONIA PRODUCTION AND TOXICITY IN FISH

       Ammonia is one of the commoner pollutants of rivers and estuaries, and can enter these waters
from a number of sources including sewage, agricultural and industrial discharges (Alabaster & Lloyd,
1980). It is also the chief metabolic waste product of fish and can therefore become a problem in the
crowded conditions typical of intensive fish culture.

       Ammonia is highly toxic to all  vertebrates if allowed to accumulate internally (WHO,  1986)
and so must either be rapidly excreted or converted to a less toxic form of waste nitrogen. Aquatic
animals have the advantage that water acts as an infinite sink for ammonia and so continuous
excretion over permeable outer surfaces such as gills is usually sufficient to balance the rate of
production and maintain low, sub-toxic, internal levels. This is the case for most
teleosts studied, which excrete the  majority of their nitrogenous waste (60-95%) simply as ammonia
(e.g. Smith, 1929; Sayer & Davenport, 1987), and are therefore referred to as ammoniotelic.  In fish
an acute increase in the ambient ammonia concentration can cause a reduction in excretion and a net
gain of ammonia from the environment; the result is an accumulation of ammonia in the body which
may be deleterious to the fish (Randall  & Wright, 1987).
    School of Biological Sciences, University of Birmingham, Birmingham B15 2TT, UK.

    Department of Physiological Sciences, University of Manchester, Manchester M13 9PT, UK.

                                          36

-------
COPPER TOXICITY

       Copper is one of the more abundant transition metals and a small amount (less than 0.1/umol
I"1) is normally present in both freshwater and marine environments.  Elevated concentrations of copper
in natural waters occur mainly through mining and smelting processes, but more recently the
acidification of freshwater lakes by acid precipitation has been associated with high levels of copper
due to leaching from bedrock  (Spry et al., 1981).  Copper is also frequently found in certain industrial
wastes, which has resulted in toxic concentrations occuring in heavily populated and industrialised
estuaries and coastal areas (Abdullah et al., 1972; Tort et al., 1986).

       The toxicity of pollutants such as metal ions, protons and surfactants can often be explained by
their surface activity, particularly on the structurally and physiologically delicate gills (Stagg &
Shuttleworth, 1987;  McDonald et al., 1989). The fish gill is a complex and multifunctional organ,
being the principal site of gas exchange,  ion regulation, acid-base balance and excretion of nitrogenous
waste.  Any pollutant interfering with one or more of these functions will have a potentially toxic
effect on the  fish.  Laur0n & McDonald (1985;  1986) showed that during short term (12"24 hours)
exposure of freshwater rainbow trout to copper, toxicity resulted from the disruption of gill
ionoregulatory function.  Plasma Na1" and Cl" levels declined due to an inhibition of active ion uptake
at copper concentrations as low as 0.2 pmol L"1.  At higher copper concentations stimulation of passive
ionic effluxes also occured.  The effects of copper on marine fish are less well documented.  However,
high levels of ambient copper (2.67 umol L"1) caused elevations in plasma Na+ and Cl" in the
seawater-adapted flounder, Platichthys flesus (Stagg & Shutfleworth, 1982a).

                                RESULTS AND DISCUSSION

       A recent study in our  laboratory considered the separate and combined effects of high levels of
ammonia and copper on rainbow trout. The levels of pollutants were based on those measured in the
Tees estuary  (NE England) and the fish was chosen to represent the salmonids (trout and salmon);
migrating species which have to negotiate estuaries, are sensitive to pollution and of great commercial
value.

EXPOSURE TO HIGH EXTERNAL AMMONIA
       Exposure to 1 uM total ammonia O^m) at pH 7.9 was non-toxic to trout adapted to either
fresh water ([NH3] = 21.6 uM) or sea water ([NH3] = 14.3 |oM). The fish survived and were able to
maintain plasma ammonia lower than ambient levels despite the high permeability Of the gill
epithelium for NH3 and its continuous production as a waste product of the fish's metabolism (Fig
l,a). Maintenance of a negative gradient for ammonia was accompanied by a net accumulation of
non-respiratory protons (but no net pH changes) in freshwater, but a reduction in non-respiratory
protons (and a subsequent alkalosis) in seawater (Fig l,b).  We interpreted these data as indicating that
the fish were able to actively excrete ammonia against a diffusion gradient, most likely as the
ammonium ion NH4+, in exchange for sodium in seawater (leading to net efflux of protons) and in
exchange for an  influx of protons in freshwater (Wilson and Taylor, 1992). However, more recent
data suggest that, in freshwater at least, trout may not possess an active NH/ exchange mechanism
(Wilson et al., 1993). Instead, an acidified boundary layer at the external gill surface may sufficiently
reduce the local  [NH3] to allow passive diffusion of NH3 from blood to the water, even when the
[NH3] in the bulk water is much higher than that in the blood.  Currently, our understanding of the
mechanisms for excretion of ammonia over the gills of fish is incomplete.
                                            37

-------
      lOOOi-
_    800
                                                11000 pmol I
                                                             -i
 o
 E
  E
  E
  CJ

600
400
        200
            0
_      +4
i
         + 2

            0

         -2

         -4
                                                              O  FW
                                                              •  SW
                                                                      I
                  C   0
                                 8       12      16
                                  Time  (h)
20      24
   Figure la: Plasma total ammonia [T^J in freshwater and seawater acclimated trout prior to and during 24 hours
   of exposure to high external ammonia concentration (Tann 1,000 umol I'1 pH 7.9, 15°C). External T^ during the
   control period (C) was less than 30 umol I'1.  * denotes values significantly different from control values and +
   denotes values significantly different when compared with the corresponding value from freshwater trout (P<0 05-
   Student's unpaired t-test).  Mean values are  shown +1  S.E.M., N=6 and 5 for freshwater and seawater trout'
   respectively.
   Figure Ib: The non-respiratory acid load for freshwater and seawater trout during the control period (C) and during
   24 hours of exposure to high external ammonia concentration (mean +1 S.E.M.; for N values see Fig 1 a) * denotes
   a value significantly different from the control mean within the  group (P<0.05).

                                             38

-------
RESPONSES TO COPPER IN FRESHWATER TROUT

       Exposure to copper was acutely lethal to freshwater trout.  The fish died within 24 hours of
exposure to 4.9 uM copper (the level measured in the Tees estuary). The physiological changes
occurring during this period gave clues to the physiological effects of the toxic metal ion and these are
illustrated in Figure 2 a-i.  The fish's ability to ionoregulate broke down so that sodium and chloride
levels in plasma progressively fell (Fig. 2a).  Plasma calcium levels were regulated at first but rose
precipitously at the point of death.  In contrast potassium levels increased progressively (Fig. 2b), and
we interpretated this as indicating potassium efflux from tissues 1 secondary to plasma ionic dilution.

       The disruptive effects of copper on branchial ionoregulation were likely to arise from:  i) the
inhibition of active ion uptake at the gills,  and ii) an increase in the permeability of the gills, and
consequently accelerated diffusive efflux of ions. For example, it has been demonstrated that copper
inhibits branchial ATPases both in vitro (Shephard & Simkiss, 1978; Bechman & Zaugg, 1988), in
isolated gill or opercular membrane preparations (Stagg & Shutfleworth,  1982b;  Crespo & Karnaky,
1983), and in vivo in freshwater rainbow trout (Lauren & McDonald, 1987b).  The inhibition of active
NaCl uptake is  most likely to be due to the high affinity of copper for sulphydryl groups (-SH) on
transport enzymes such as Na+, K+-ATPase located in the gill epithelium (Shaw & Grushkin, 1964;
Stagg &  Shutfleworth, 1982b).

       The increase in gill permeability mentioned above is probably related to the ability of copper
to displace calcium from biological ligands (Nieboer & Richardson, 1980).  Since calcium is known to
be important in controlling the integrity and permeability of the branchial epithelium in fishes (Potts &
Fleming, 1971;  Cuthbert & Maetz, 1972;  McDonald & Rogano, 1986), Lauren & McDonald (1985)
proposed that the stimulation of ionic effluxes by copper may be a result of increased ionic
permeability of the gill, secondary to the displacement of Ca++ ions from paracellular tight junctions.
The toxicity of  copper is therefore dependant on the ambient calcium levels during exposure (e.g.
Alabaster & Lloyd, 1980;  Reid & McDonald, 1988).  At low enough concentrations of copper,
ionoregulatory disturbances can occur without apparent physical damage to the gills (Lauren and
McDonald, 1985). However, at higher levels of copper the inhibition of active Ion pumps and
displacement of intercellular and membrane-bound  calcium may lead to  a breakdown in branchial cell
volume control  and epithelial organisation.  Indeed, in our freshwater trout exposed to 4.9 uM copper
we observed severe gill histopathologies which included cell adhesion failure, swelling, and pillar cell
detachment leading,to numerous haematomas (red cells pooling in regions of lamellae).

       The structural disruption of the gill epithelium  in freshwater trout exposed to copper had
drastic effects on respiratory gas exchange. The fish became-progressively hypoxic and  as a
consequence haemoglobin oxygenation was reduced so that oxygen supply to tissues  was reduced (Fig.
2,c).  Carbon dioxide  accumulated (Fig. 2,d) and the fish switched to anaerobic metabolism so that
lactic acid levels in the blood rose (Fig. 2,e).  The blood became progressively acidotic (Fig. 2,f) due
to a combined respiratory (CO2) and metabolic (lactic acid) acidosis; this in turn caused a rightward
Bohr shift on the fish's haemoglobin and oxygen uptake and transport were further curtailed. Physical
damage to the gills (and reduced diffusion of NH3) may also be responsible for the progressive
accumulation of plasma ammonia (Fig. 2,g). However, inhibition of active ammonium ion transport
NH//Na't' or H+ exchange) may also contribute to this effect. As part of its attempt to compensate  for
these changes by  increasing oxygen transport the fish exhibited an increase in haematocrit to values as
high as 50% (Fig. 2,h).  This was partially due to a reduced plasma volume, partly due to increased
red cell numbers (probably released from the spleen) and partly due to red cell swelling.  These
changes are likely to have resulted in an increase in apparent blood viscosity.  Simultaneously the fish
                                             39

-------
     160



£   M

i_>


i   120

*a
z
—   100

•

r     «


O     ^
   3
                  COPPER (4-9pM)-
   a
   o.
   O
   a
   Q-
   o
   a.
   a
   a.
 a —
   O.

   a
   JO

   a.
      10




       5



     0-5



     0-4



     0-2



      12


       e


       4


       0

     8-0
   e   7-6
       7-2

      2000
   O
   E
   J5-


   "e
   E
   a
    1000
     40



     20


     •0
"5 e
Si S
g fe.
t -g
a o
o  •
                  J	I
               f-
                  J	I
                                     a)
                                      c)
I	I
                                      f)
                                •  J  h)
                            I    i    i
            C  0
         100



         80



         60



         40
                4   6   12  16  24


                   Time (h)   40
                                                                     is
                                                                     S§
                                                                     l
                                                                     i
                                                                        g
                                                                       "
                                                                    1 1
                                                                    =3 8
                                                                    s'
                                     c S
                                    _O oj

                                    1 S
                                                                     .•8
                                                                    o  «j
                                                                    S  «
                                                                    c  rt

                                    en

                                    E

-------
showed increased heart rate and blood pressure (Fig. 24), and together with the changes in haematocrit
these indicate that circulating catecholamines were elevated.  The combined effect of these responses
was probably cardiac failure leading to death. Similar trauma have been described from fish exposed
to very acid waters (the product of acid-rain) (Milligan & Wood, 1982).  The fish dies because its
integrated physiological responses to the toxic environment culminate in its destruction. The
breakdown appears to start with disruption of ionoregulation which in freshwater is vital to survival
because the ionic gradients between the fish and its environment are very large.

RESPONSES OF SEAWATER TROUT TO COPPER

        We have seen that acute exposure of freshwater trout to 4.9 umol copper  I'1 at pH 7.9 caused
100%  mortality within 24 hours, as a result of severe ionoregulatory and respiratory disturbances.  In
contrast, trout acclimated to full-strength sea water (external [NaCl] three- to fourfold higher than
blood  [NaCl]) and exposed to 6.3 (jrnol copper r1 at the same pH for 24h, exhibited no significant
change in plasma ion concentrations, no respiratory problems and survived with  apparently
undamaged gill lamellae. It is clear that copper toxicity, and the physiological responses during acute
exposure, are greatly reduced at high salinities.  These results were unexpected; (i.e.  our prediction
was that an ionoregulatory inhibitor such as copper would cause a net increase in plasma ions in fish
maintained in a hypersaline environment).
        Thus  seawater-adapted teleosts are less sensitive to copper than freshwater species as
previously described by Eisler & Gardner (1973); Voyer (1975);   Taylor et al. (1985). This may be a
function of physicochemical factors such as higher calcium levels, resulting in greater competition with
copper for membrane binding sites (Laure"n & McDonald, 1985; Reid & McDonald,  1988), or high
carbonate alkalinity and the presence of chelating agents which can reduce the toxic forms of copper
by complexation (Chakoumakos et al., 1979; Miller & Mackay, 1980).  Alternatively, the lower
toxicity could be related to  physiological differences in the mechanisms of branchial salt transport and
the location of the key transepithelial ion transporting enzymes in seawater acclimated fish.

        Environmental Ca2+ and Mg2+ are thought to play an important role in controlling the branchial
 permeability to Na+ in both freshwater and seawater teleosts (Potts and Fleming 1971;  Bornancin et
 al. 1972;  Isaia and Masoni 1976). Copper is thought to displace surface Ca2+ in freshwater trout
 causing increased gill permeability Lauren and McDonald (1985).  However, the ambient [Ca +] in
 seawater is more than 26 fold higher than in fresh water which would obviously provide more
 competition for external binding sites with copper.  This may explain the differential responses found
 in seawater trout in the present study. If it is assumed that the basolateral Na+K+ -ATPase enzyme is
 unaffected in 24h (due to the slow intracellular uptake of copper), then any ionoregulatory disturbances
 can only be attributed to changes in the passive diffusion of Na+ and Cl" which is Ca2' (and Mg +)
 dependent. In seawater the high external [Ca2+] may simple out-compete the external copper, resulting
 in no ionoregulatory disturbances despite the large Na+  and Cl" gradients across the gills.  In contrast,
 in freshwater trout osmotic and ionic gradients are large and calcium levels relatively low.  Addition
 of copper may displace much of the calcium leading to spatial disruption of the  gill epithelium  and
 affecting specific ATPases  (see Fig. 3).  However, to fully test this hypothesis would require the use
 of isotopic unidirectional ion flux measurements and synthetic seawater media with controlled Ca+ and
 Mg2+ levels to determine whether differences in [Ca2+] and [Mg2+] are in fact the determining factors in
 the differential responses to copper exposure seen in trout acclimated to freshwater and seawater.
 Longer term exposure of fish to copper results in uptake over the gills or via the gut and subsequent
  damage to internal organs such as the liver (Baker 1969).  Consequently, basolateral ion regulation
 may  be affected by long term exposure.
                                              41

-------
    NH
                                                                                      •^ ^ i

                                                           =L_2*     NH4
   CO
Figure 3: Plasma total ammonia concentration in the arterial blood of the seawater acclimated rainbow
trout exposed to 6.3 nmol'1 copper I'1 at  12°C.  Points represent mean  values +1 S.E.M.) during the
pre-exposure period and following 1, 5 and 24 hours of exposure at pH 7.9  *  denotes mean values
significantly different from the pre-exposure mean (C) (P<0.05; Student's paired t-test),

                                         42

-------
THE INHIBITORY ACTION OF COPPER ON AMMONIA EXCRETION

       Exposure of freshwater trout to copper caused a large and progressive increase in plasma  '
ammonia levels (Fig. 2,g). Seawater trout similarly experienced elevated ammonia levels when
exposed to copper but they reached a plateau after 4 hours at levels half that reached by 4 hours and
one tenth of the value after 19 hours in  freshwater trout (Fig. 4).  Copper exposure caused no increase
in O2 consumption in the seawater trout. If the normal relationship between MO2 and MNH3 (Brett
and Zala 1975;  Heisler 1984) remained constant throughout exposure to copper, then the elevation of
plasma [TAmm] cannot be explained by any change in the rate of endogenous ammonia production.
Under normal pre-exposure conditions branchial ammonia excretion was probably the resultjOf..a.
combination of NH3 diffusion and apical Na+/NH4+ exchange (Wright and Wood 1985).  It seems
unlikely that any changes in the gill NH3 diffusion capacity occurred since no major changes in the
other, blood gas tensions PaO2 or -PaCO2  were observed. This argues for an inhibition of apical
Na+/NH4+ exchange by copper as suggested by LaurGn and McDonald (1985) in freshwater rainbow
trout. 'The stabilisation of plasma [TAmm] after 5 hours in seawater trout could then  be explained if it
is assumed that the rate of branchial NH3 clearance by diffusion will again match the endogenous rate
of NH3 production once a new, larger PNH3 gradient over the relatively undamaged-gill lamellae is
established.  In freshwater the gills are structurally damaged and  ionoregulation breaks down after
copper exposure (Fig. 2), so that apical  Na+/NH4+ exchange and NH3 diffusion are likely to be
impaired. Consequently, ammonia levels continue to rise rather than reaching a plateau.
         300
 -x'    200
                                          Copper  6-3 jjmol  I
  o
  e
  E
  E
100
             0
                    C   0
8      12      16
  Time  (h)
                                                          20     24
 Figure 4: Schematic diagram illustrating the possible routes for NH3 and CO2 excretion,
 ionoregulation and acid-base regulation of a generalised gill epithelial cell from a freshwater trout. As
 well as Na+/K+(NH4+) and Ca2YMg2+ exchanges on the basolateral membrane and Na+/H+(NH4+) and,
 C1"/HCO3:.  exchangers on the apical membrane the diagram includes the proton pump (*), additional
 acidification of the boundary layer by hydration of CO2 catalysed by carbonic anhydrase (CA) **  and
 ammonia trapping in the acid boundary layer,-*** (for additional description see text).'  ,      '   -
                                             43

-------
                                         CONCLUSION

         Exposure of trout to elevated levels of ammonia or copper in freshwater or seawater has
 provided useful indicators of the mechanisms of ammonia exchange over the fish gill. Trout are able
 to maintain ammonia excretion against high external levels either by active excretion of NH4+ (in
 exchange for H+ or Na+) or by excretion of NH3 into an acid boundary layer (sustained by an
 electrogenic H+ pump (Lin & Randall, 1991) and/or by local hydration of CO2, catalysed by carbonic
 anhydrase (Randall & Wright, 1987)).  These processes are summarised on Fig. 4.  Copper, because it
 inhibits branchial ATPases and competes with calcium for ion channels and ion binding sites,  disrupts
 ionoregulation and in freshwater can cause breakdown of cell homeostasis and cell adhesion in gill
 epithelia (cf Fig. 4).  In seawater calcium levels are relatively high so that elevated levels of copper
 are less likely to cause physical damage to the gill lamellae and ionoregulation is apparently
 unaffected. However, ammonia accumulates in the plasma, indicating that some  apical transport
 processes involved in ammonia excretion (e.g.  Na+/NH4+  exchange) are affected  by copper;  whereas,
 the principle basolateral mechanisms involved in sodium and chloride regulation  are apparently
 unaffected. In support of this, we observed a significant increase in plasma [HC(V)  ] in the
 seawater trout exposed to copper, indicating inhibition of  apical CITHCCV exchange in addition to
 apical Na*/NH4+  exchange.

        The breakdown of the gill epithelia in freshwater trout, in the presence of copper,
 compromised ionoregulation and respiratory gas exchange.  The  fish became markedly hypoxic and
 hypercapnic and accumulated lactic acid; resulting in a combined respiratory and metabolic  acidosis.
 As part of a physiological response to these accumulating problems, probably mediated by elevation of
 circulating catecholamines, the fish showed greatly increased haematocrit, heart rate and blood
 pressure, resulting in death from heart failure.

                                   ACKNOWLEDGMENTS

        Ted Taylor's attendance at the Symposium was funded by the US Environmental Protection
 Agency and he is grateful both to the EPA, the staff and students and in particular Professor Hongjun
 Jin of Nanjing University, PRC for organising this successful international meeting.
        Rod Wilson was supported by a studentship from the Science and Engineering Research
 Council and the research was conducted in cooperation with Zeneca Environmental Research
 laboratories, Brixham, Devon, UK. The  authors are grateful to Sheila Craggs for her patient
 processing of the manuscript.

                                        REFERENCES

 Abdullah, M.I., Royle, L.G. and Morris, A.W. 1972. Heavy metals concentration in coastal waters
        Nature (London) 235: ISS-'eO.
 Baker, J.T.P. (1969) Histological and electron microscopical observations on copper poisoning in the
        winter flounder Pseudopleuronectes  americanus J.  Fish Res. Bd.  Can. 26:  2785-2793.
 Beckman, B.R. and Zaugg, W.S. 1988. Copper intoxication in Chinook salmon Oncorhynchus
        tshawytscha induced  by natural springwater:  Effects on gill Na+, K+-ATPase, haematrocrit,
        and plasma glucose.  J. Fish. Aquat. Sci. 45: 143Q-1435.
Bornancin, M., Cuthbert, A.W. and Maetz, J. 1972.  The effects of calcium on branchial sodium fluxes
       in sea-water adapted eel Anguilla anguilla L. J. Physiol. (London) 222: 487-496.
                                           44

-------
Brett, J.R. and Zala, C.A. 1975. Daily pattern of nitrogen excretion and oxygen consumption of
       sockeye salmon (Oncorhynchus nerka) under controlled conditions. J. Fish. Res. Bd. Can.
       32:2479-2486.
Campbell, P.G.C. and Stokes, P.M. 1985.  Acidification and toxicity of metals to aquatic biota. Can.
       J. Fish. Aquat. Sci. 42: 2034-2049.
Chakoumakos, C., Russo, R.C. and Thurston, R.V. 1979.  Toxicity of copper to cutthroat trout (Salmo
       clarki) under different conditions of alkalinity, pH and hardness.  Environ. Sci. Technol. 13:
 :<     213-219.
Crespo, S. and Karnaky, K.J. Jr. 1983. Copper and zinc inhibit chloride transport across the opercular
       epithelium of seawater-adapted killifish, Fundulus heteroclitus. J. Exp.  Biol. 102: 337-341.
Cuthbert, A.W. and Maetz, J. 1972. The effects of calcium and magnesium on  sodium fluxes through
       the gills of Carassius auratus L..  J. Physiol.  221:633-643.
Eisler, R. and Gardner, G.R. 1973. Acute toxicity to an estuarine teleost of mixtures of cadmium,
       .copper and zinc salts.  J. Fish Biol.  5: ISrM-Z.
Heisler, N. 1984. Acid-base regulation in fishes. In Fish Physiology, Vol. 10A.  Edited by W.S.
       Hoar and D.J. Randall.  Academic Press Inc., New York  pp. 315-401.
Isaia, J., Masoni, A.  1976. The effects of calcium and magnesium on water and ionic permeabilities in
       the seawater adapted eel Anguilla  anguilla L. J. Comp. Physiol.  109: 221-233.
LaurQn, D.J.  and McDonald, D.G. 1985.  Effects of copper on branchial ionoregulation in the
       rainbow trout, Salmo gairdneri Richardson.  Modulation by water hardness and pH. J. Comp.
       Physiol.  B 155: 635-644.
LaurOn, D.J.  and McDonald, D.G. 1986.  Influence of water hardness, pH, and alkalinity on the
        mechanisms  of copper toxicity in juvenile rainbow trout,  Salmo gairdneri.  Can. J. Fish.
        Aquat. Sci.   43: 1488-M96.
Lauren, D.J.  and McDonald, D.G. 1987.  Acclimation to copper by rainbow trout, Salmo gairdneri:
        Biochemistry. Can. J. Fish. Aquat. Sci. 44: 105'1!!.
Lin, H. and Randall, D.J. 1991
McDonald, D.G. and Rogano, M.S. 1986.  Ion regulation by the rainbow trout,  Salmo gairdneri in
        ion-poor water.  Physiol. Zool. 59: 318-331.
McDonald, D.G., Reader, J.P. and Dalziel, T.R.K. 1989.  The combined effects of pH and trace metals
        on fish ionoregulation.  In Acid Toxicity and aquatic animals.  Edited by R. Morris, E:W.
        Taylor, D. Brown and J.A. Brown.  Soc. Exp. Biol. Semin. Ser. 31: pp 221-242.
Mifler, T.G. and Mackay, W.C. 1980. The effects of hardness, alkalinity and pH of test water on the
        toxicity of copper to rainbow trout. Wat. Res. 14: 129"133.
Milligan, C.L. and Wood, C.M. 1982. Disturbances in haematology, fluid volume, distribution and
        circulatory function associated with low environmental pH in the rainbow trout, Salmo
        gairdneri. J. Exp. Biol. 99: 397-415.
Nieboer, E. and Richardson, D.H.S. 1980. The replacement of the nondescript term 'heavy metals' by
        a biologically and chemically significant classification of metal ions. Environ. Pollut. 1: 3-26.
Potts, W.T.W. and Fleming W.R. 1971.  The effects of environmental calcium  and ovine prolactin on
        sodium balance in Fundulus kansae.  J. Exp. Biol. 55: 63-76.
Randall, D.J. and Wright, P.A. 1987.  Ammonia distribution and excretion in fish.  Fish Physiol.
        Biochem. 3: lOT^O.
Reid, S.D. and McDonald, D.G. 1988.  Effects of cadmium, copper, and low pH on ion fluxes in the
        rainbow trout, Salmo gairdneri.  Can. J. Fish. Aquat. Sci. 45: 244-253.
 Sayer, M.D.J. and, Davenport, J. 1987.  The relative importance of the gills to ammonia and urea
        excretion in five seawater and one freshwater teleost species.  J. Fish. Biol. 31: 561-570.
 Shaw, W.H. and Grushkin, B. 1957.  The toxicity of metal ions to aquatic organisms.  Arch. Biochem.
        Biophys. 67: 447-453.
                                             45

-------
 Shephard, K and Simkiss, K. 1978. The effects of heavy metal ions on Ca^ ATPase extracted from
        fish gills.  Comp. Biochem. Physiol. 6 IB: 69-72.
 Smith, H.W. 1929. The excretion of ammonia and urea by the gills of fish.  J. Biol. Chem. 81-
        727-742.
 Spry, D.J., Wood, C.M. and Hodson, P.V. 1981. The effects of environmental acid on freshwater fish
        with particular reference to softwater lakes in Ontario and the modifying effects of heavy
        metals. A literature review. Can. Tech. Rep. Fish. Aquat. Sci. 999: 149p.
 Stagg, R.M. and Shuttteworth, T.J.  1982a.  The accumulation of copper in Platichthys flesus L. and its
        effects on plasma electrolyte concentrations. J. Fish. Biol. 20: 491-500.
 Stagg, R.M. and Shuttieworth, T.J.  1982b.  The effects  of copper on ionic regulation by the gills of the
        seawater-adapted flounder Platichthys flesus L..  J. Comp. Physiol.  149: 83-90.
 Stagg, R.M. and Shutfleworth, T.J.  1987. Sites  of interactions of surfactants with beta-adrenergic
        responses in trout Salmo gairdneri gills.  J. Comp. Physiol.  157: 429-434.
 Taylor, D., Maddock, E.G. and Mance, G. 1985. The acute toxicity of nine 'greylist' metals (arsenic,
        boron, chromium, copper, lead, nickel, tin, vanadium and zinc) to two marine fish species:
        Dab Limanda limanda and grey mullet Chelon labrosus.  Aquat. Toxicol. 7: 135'144.
 Tort, L., Torres, P. and Flos, R. 1986.  Effects on dogfish haematology and liver composition after
        acute copper exposure.  Comp. Biochem. Physiol. 87C: 349-353.
 Voyer, R.A. 1975. Effect of dissolved oxygen concentration on acute toxicity of cadmium to the
        mummichog, Fundulus heteroclitus (L.)  at various salinities. Trans. Am. Fish. Soc. 104:
WHO 1986.  Environmental Health Criteria No. 54; Ammonia.  World Health Organisation, Geneva,
       IPCS, 1986.
Wilson, ,R.W. and Taylor, E.W. 1992.  Transbranchial ammonia gradients and acid-base responses to
       high external ammonia concentration in rainbow trout Oncorhynchus mykiss acclimated to
       different salinities. J. Exp. Biol. 166: 95'112.
Wilson, R.W., Taylor, E.W. 1993a.  The physiological responses of freshwater rainbow trout
       Oncorhynchus mykiss during acutely lethal copper exposure.  J. Comp. Physiol. B  163: 38-47.
Wilson, R.W. and Taylor E.W. 1993b.  Differential responses to copper in rainbow trout
       Oncorhynchus mykiss  acclimated to seawater and brackish water.  J. Comp. Physiol.  163:
       239-246.
Wilson, R.W., Wright, P.M., Munger, S. and Wood, C.M. 1993.  Ammonia excretion in the freshwater
       rainbow trout, Oncorhynchus mykiss and the importance of gill boundary layer acidification:
       lack of evidence for Na+/NH4+  exchange.  J. Exp. Biol. (submitted).
Wright, P.A.  and Wood, C.M. 1985. An analysis of branchial ammonia excretion in the freshwater
       rainbow trout: effects of environmental pH change and sodium uptake blockage. J. Exp. Biol.
       114: 329-353.
                                           46

-------
    ECOTOXICOLOGICAL STUDY OF FENPROPATHRIN ON SOME
 COMMON CHINESE AQUATIC ORGANISMS AND EVALUATION OF
               ITS EFFECT ON THE AQUATIC ECOSYSTEM

              Yiwei Yin1, Zhaohui Wang1, Yongyuan Zhany2, Ying Xu2, Lihong Xu2,
                            Wengzhong Wu2 and Guoshong Chen2
                                       ABSTRACT

       Fenpropathrin is acutely toxic to some common Chinese aquatic organisms. The toxic
mechanisms of this pesticide in fish involve inhibition of Na+,K+-ATPase, modification of the Na+
channel in the nervous system, perturbation of ion metabolism, and ultrastructural injury of the fish
gill.  However, this pesticide does not accumulate persistently in the body of fish.  Most of it is
rapidly absorbed by sediment after entering a body of water, and the residue in the aquatic phase
degrades within a few days. The pesticide contained in the sediment also completely degrades within
several days. While the toxic effect of fenpropathrin is not persistent, use of this pesticide within any
area containing fish must be carefully monitored to avoid acute poisoning.

                                    INTRODUCTION

       Fenpropathrin (a-cyano-3-phenoxy-henzyl-2,2,3,3-tetramethyl-cyclopropane-propane-carboxylic
ether) is a new synthetic analogue of pyrethroids that was synthesized in China in 1991 and
manufactured under the commercial name of Meothrin. Fenpropathrin is not very toxic to mammals
(including humans), but are highly toxic to a wide variety of micro-organisms at a relatively low dose.
The toxicity of fenpropathrin to aquatic organisms, especially fish, is essentially unknown. Coats and
O'Donnell-Jeffery (1979) have  reported that the octanol/water partition coefficient of fenpropathrin is
1070 and the 24 hour LC50 is 76.7 ppb in Rainbow trout, which is considered to be of high toxicity.
The half-life of fenpropathrin in soil under aerobic and anaerobic conditions at 25 °C is 10 and 25
weeks, respectively (Su and Fan, 1989). It is not known whether the degradation  pattern would be
similar in the benthic environment of water bodies.  Since fenpropathrin is a highly toxic agent,
utilization of this pesticide could be a significant problem to fisheries in China.  Therefore, this study
examined the ecotoxicological effects of fenpropathrin on fish and other aquatic organisms, and its
degradation in both the aquatic and benthic phase.

                         METHODS, RESULTS AND DISCUSSION

ACUTE TOXICITY OF FENPROPATHRIN TO AQUATIC ORGANISMS

       While most pyrethroids are considered to be highly toxic to aquatic organisms, some are not,
making it necessary to precisely determine the toxic level of fenpropathrin to some common Chinese
organisms. Fenpropathrin was dissolved in acetone to form a stock solution, and appropriate volumes
of stock solution were added to dechlorinated tap water to make up the required concentrations. Since
    Institute of Aquatic Ecoscience, Jinan University, Guangzhou, 510632

    2State Key Laboratory for Freshwater Ecology and Biotechnology of China, Institute of Hydrobiology,
 Chinese Academy of Sciences, Wuhan, 430072
                                           47

-------
  preliminary studies showed that the lethal concentration of fenpropathrin was lower than its soluble
  limit (340 ppb), no emulsifiers or solvents were used in these tests. Acute toxicity tests were carried
  out on the following representative Chinese aquatic organisms; Silver carp (Hypophthalmichthys
  mohtnx), Grass fish (Ctenopharyndon idell), Mosquito fish (Gambusia affinis), Daphnia magna H.B.,
  Daphnta carinata and Chronomus sp.. The results of these tests are given in Table 1.  It is clear that
  fenpropathrin is highly toxic to aquatic organisms, and extreme caution in its use around fisheries
  areas is highly recommended.

         Table 1. Acute toxicity of fenpropathrin on some common Chinese  aquatic organisms.
         Species
LC50 or EC50 Time  Temperature
       (ppb)  (hr)    (°Q
Mosquito fish
(Gambusia affinis)
Silver carp
(Hypophthalmichthys molitrix)
Grass fish
1.8
1.8
2.5
96
96
96
23-25
25-27
16.5-17.5
        (Ctenopharyndon idell)

        Daphnia magna HB

        Daphnia carinata

        Chronomus sp.
       1.8

       0.7

       5.5
48

48

48
20±5

20±5

22±5
' TOXICOLOGICAL STUDY OF FENPROPATHRIN IN FISH

        While no aquatic lexicological studies of fenpropathrin have been conducted, pyrethroids in
 general appear to effect fish Na+ channel activity (Khan, 1983; Cole et al., 1984; Laurence and Caside,
 1983), ATPase activity (Clark and Matsumura, 1982), ion metabolism (Bradbury, 1989) and gill
 structure. The lexicological studies outlined below investigate these four perturbations.

 A. Effect of Fenpropathrin on ATPase Activity in the Gill and Kidney of Silver Carp

        Silver carp were exposed to fenpropalhrin in vivo and the gills and kidneys removed for assay
 Gills and kidneys were also removed from unexposed fish and then immersed in fenpropathrin in vitro.
 Isolation of gill and kidney tissues, preparation of tissue homogenates, and measurement of ATPase
 activity were done according to the method of Matsumura et al. (1982) and Xu et al (1987)
 Results showed that Na+,K+-ATPase and total ATPase activities were higher in gill homogenates than
m kidneys (results not shown). The inhibition of ATPase in gill and kidney tissue of Silver carp after
exposure to 10-3 M fenpropathrin is shown in Table 2.  At a fenpropathrin concentration of 10-3 M,
inhibition of Na+.K^-ATPase in gill and kidney homogenates was 43.9% and 29.2%, respectively, but
inhibition of total ATPase in the two homogenates was only 23.2% and 16.8%, respectively.  The
inhibition of ATPase in the two homogenates was dose-dependent, and inhibition was significant
above a concentration of 10-6 M.
                                           48

-------
                 Table 2.  Inhibition of ATPase activity in gill and kidney tissue of
                        Silver carp after exposure to 10-3 M fenpropathrin
              Total ATPase

              Na+,K+-ATPase
                             Gill (% inhibition)

                             23.0 (P<0.01)

                             43.9 (P<0.10)
                                    Kidney (% inhibition)

                                    16.8 (P<0.05)

                                    29.2 (P<0.01)
       The inhibition of Na+,K+-ATPase activity in gill homogenates of Silver carp exposed to a
fenpropathrin concentration near the LC50 for 24 hr is shown in Table 3. Table 4 shows inhibition of
total ATPase and Na+.K^-ATPase in gill homogenates of Silver carp after exposure to fenpropathrin at
a concentration well above the LC50 (5 ppb) for 5 hr.  The results indicate that similar inhibition of
Na+,K+-ATPase activity (approximately 40%) occurred when fish were exposed to either 1 ppb for 24
hr or 5 ppb for 5 hr.  Since fish mortality occurred while Na+,K+-ATPase activity was significantly
inhibited, this suggests that inhibition Na+,K+-ATPase activity could be one of the causes of death.

       Table 3. Gill Na+,K+-ATPase activity of Silver carp in vivo when exposed to different
                    concentrations of fenpropathrin for different periods of time
Exposure
              Control
              0.5 ppb
              0.75 ppb
              1.0 ppb
       Activity
24 hr  (uM Pi/ug/hr)  4.03±0.26     4.01±0.33      4.04±0.25
       % inhibition
       % mortality                          0             0
        Activity
48 hr   (uM Pi/ug/hr)
        % inhibition
        % mortality
72 hr
 96 hr
Activity
(uM Pi/ug/hr)
% inhibition
% mortality

Activity
(uM Pi/ug/hr)
% inhibition
% mortality
              4.07±0.26      4.02+0.30
4.15+0.12
                                    0
                             3.62±0.24
                             11.07
                             12.3
4.02±0.21
3.66±0.17
8.35
8.7
3.22±0.26
19.90**
29.5
                                                          2.52±0.11
                                                          37.47**
                                                          37.5
                             2.45±0.19
                             39.80**
                             41.8
3.86±0.14
6.89
3.61±0.32
13.01*
15.6
2.50+0.24
39.76**
46.3
2.32±0.20
42.29**
47.5
 n=4; t-test; results expressed as average ± standard deviation
                                             49

-------
                        Table 4.  ATPase activity in gill tissue of Silver carp
                           after exposure to 5 ppb fenpropathrin for 5 hr.
                Total ATPase


                Na+,K+-ATPase


                Mortality (%)
activity*
% inhibition
Control

9.18±0.39
activity*      4.10±0.20
% inhibition
Exposed

7.15+0.46
22.11**

2.29±0.29
44.15**

55.5
             Pi/vg/br, mean±SD; ** t-test, significantly different
 B. Effect of Fenpropathrin on the Na* Concentration in the Blood of Grass Carp

        Urinary Na+ and K* concentration and excretion rates are significantly elevated in trout after
 exposure to fenvalerate (Bradbury et al., 1987), which may be associated with the effect of fenvalerate
 on ion metabolism and osmoregulation (Symonik, 1989). We therefore examined the effect of
 fenpropathrin on the Na* concentration in the blood of Grass carp.
 Grass carp fingerlings (14 cm in length) were exposed to fenpropathrin at concentrations between 1-3
 ppb.  The test solution was changed every 2 hours.  Blood was sampled from the gill aorta, and
 weighed and nitrated.  The blood Na+ concentration was measured with an ELC-I liquid chromatograph
 (Waters Company).

        Blood Na* concentration decreased approximately 25% after several hours of exposure (Fig. 1)
 and decreased more rapidly in the higher concentrations  of fenpropathrin (3 ppb). The fall in Na+
 concentration in the blood may be associated with disturbances in ion metabolism and osmoregulation,
 and could be involved in the lethal action of fenpropathrin on fish.

 C.  Effect of Fenpropathrin on Na+ Channel Activity of Cardiac Muscle of Silver Carp

       Pyrethroids have been reported to increase transmembrane sodium influx by delaying
 inactivation of the sodium channel (Narahashi et al., 1967).  Decamethrin caused a positive inotropic
 effect on isolated guinea pig atrial muscle which was abolished with tetrodotoxin, implying that the
 effect of decamethrin is due to modification of the sodium channel (Berlin et al., 1984).  The current
 study was undertaken to investigate  the inotropic action of fenpropathrin on the isolated heart of Silver
 carp.

       Cardiac muscle function was studied using the method of Berlin et al. (1984) and Li et al.,
 1989.  The heart was removed from the fish and a cannula inserted into the liver vein to perfuse the
heart.  The perfusate was pumped out through the aorta.  The heart was equilibrated in physiological
buffer solution for 20-30 minutes and then 2  ml fenpropathrin (10-4 - 10-6 M) dissolved in 10%  ;
ethanol was injected into the cannula.  The results are shown in Fig.  2.  Fenpropathrin caused a
dose-dependent (10-4 M to 10-6 M) increase in heart rate which peaked in 8-10 sec and then
recovered to resting levels.  When 1 ml 10-7 M tetrodotoxin was injected during the tachycardia
induced by 2 ml 10-5 M fenpropathrin, the heart stopped beating for several seconds but then regained
its rhythm. Therefore, the positive inotropic effect of fenpropatiirin on cardiac muscle was blocked by
tetrodotoxin.  Since tetrodotoxin blocks the increase in Na+ influx in nerve endings  (Berlin et al.,
                                             50

-------
                                                    O
                                                   JC
                                                        I
                                                        U
                                                        a
                                                        Q
                                                        §
                                                        a
                         CM
           (spuosnoLyJ)
(6/5n)  uoi -f-DN  ^o

-------
                                          •8
                                           a
                                           o
                                          &

                                          "8
                                          p
UQ3H
     52

-------
 1984), modification of the sodium channel and imbalances in ion metabolism may also be involved in
 the lethal effect of fenpropathrin on fish.

 D. Effect of Fenpropathrin on Gill Morphology of Silver Carp

        Kumaragura et al. (1982) and Bradbury et al. (1987) have reported that fenvalerate may
 damage gill structure.  Damage to the gill surface was characterized by local inflammation, increased
 mucous secretion and separation of adjacent lamellae from the connective tissue. Numerous necrotic
 cells were also found in these separated interlamella zones. These histological changes may alter
 respiratory function in addition to ion metabolism.  The following study  evaluated the effect of
 fenpropathrin on gill ultrastructure by histopathological examination of Silver carp gill tissue..
 Silver carp fingerlings (10-15 cm in length) were exposed to 0.5, 0.75 and 1.0 ppb  fenpropathrin for
 24-96 hours. Another group was exposed to 5.0 ppb fenpropathrin for 5 hours. Gill tissue was
 removed after 5, 24, and 72 hr,, fixed in Bouins' fixative, sectioned in paraffin, and stained with H.E..
 Results are shown in Figs. 3-6.  Histopathological changes in the gill showed that the degree of
 damage was directly related to the duration of exposure and the concentration of fenpropathrin the fish
 were exposed to.  The gill sections showed that lamellae were swollen, hyperplasic, and fused
 together.  Mucous secretion was increased  and  some blood cells  were  free from the lamellae. Because
 of the importance of the gills in respiration, absorption,  secretion and excretion in fish, injury to the
 gill may contribute to fish mortality.  These changes also provide some histological basis for the
 functional alterations discussed above.

 ACCUMULATION AND ELIMINATION OF  FENPROPATHRIN IN  FISH

       There is little  data on the toxicokinetics of fenpropathrin to date, although it has been reported
 that the octanol/water partition coefficient of fenpropathrin is 1070 (Coats and O'Donnell-Jeffery,
 1979). It was therefore necessary to determine some toxicokinetic parameters of the pesticide in order
 to evaluate its effect on the aquatic ecosystem.

       A routine shaking method (Coats and O'Donnell-Jeffery, 1979) was used to determine the
 partition coefficient of fenpropathrin, and the actual biological  concentration factor (BCF) and
 elimination of the pesticide in the body of Silver carp were also determined. The average weight of
 the fingerlings was 10 g.  The quantity of fenpropathrin was measured with a Hewlett-Packard 5890 A
 gas chromatograph (Wengzhong and Ying,  unpublished).

       The octanol/water partition coefficient of fenpropathrin was 1096.48, which corresponds
 closely to previously published results (Coats and O'Donnell-Jeffery, 1979).  Accumulation and
 elimination of fenpropathrin in Silver carp carcasses are shown in Figs. 7 and 8.  The concentration of
 fenpropathrin which accumulated in Silver  carp reached a maximum after approximately 8 hours
 exposure (Fig. 7), at which time the BCF was 411.3.  The fenpropathrin concentration then began to
 decrease.  The half-life of fenpropathrin (Fig. 8) was 30.14 hours, and over 70% of the pesticide was
 eliminated after 60 hours.  Accumulation therefore occurred much faster than elimination.  The
 octanol/water partition coefficient of most pyrethroids ranges from 104 to 107 M, and the BCF lies
 generally in the range of several thousand in fish (Coats et al., 1989).  Fenpropathrin has a much
lower partition coefficient and BCF than most pyrethroids and  can be eliminated rapidly, suggesting
that the problem of bioconcentration may not be serious.
                                            53

-------
          Figure 3.  Control (x 100).
Figure 4.  5 ppb (x 100) after 5 hours exposed.
                   54

-------
Figure 5.  0.75 ppb (x 100) after 24 hours exposed.
Figure 6.  0.75 ppb (x 100) after 72 hours exposed.
                    55

-------
I   I—I—I—I—1—i—1—1—I
                                                         I
                                                         I?
                                              CO
   /[poq gey. ui
                 56

-------
                                                                            o
                                                                            '-O

                                                                            to
                                                                           'CM
                                                                                       B
                                                                                       •s

                                                                                  _
                                                                                 3
                                                                                 o
                                                                                       §
                                                                                       ftO
                                                                                       E
o
o
         o.'
Q
CO
O
r--
O .
ID
        ~r
         o
                                                                            o
                                                      C5
                                                      O
                                       57

-------
 ELIMINATION AND TOXICITY OF FENPROPATHRIN IN A WATER-SEDIMENT SYSTEM

        A detailed report concerning the degradation of fenpropathrin in soil was given by Su and Fan
 (1989), but there is no data on the fate or toxicity of the pesticide in water or sediment.  The present
 study investigates the degradation and toxicity of fenpropathrin in a water-sediment system.
 Sediment was collected from the fish pond at the Institute of Hydrobiology, Chinese Academy of
 Science, Wuhan.  The sediment was sifted through a 45  mesh/cm sifter, and allowed to settle
 overnight.  Amounts of 1000 g of the settled sediment were added to 15 L dechlorinated tap water in
 20 L glass aquaria. Fenpropathrin was added to the aquaria to make final concentrations of 1, 5, 10,
 15, and 100 ppb.  Seven hours after addition of the pesticide, samples of both the water and the
 sediment were taken. Both samples were analyzed for fenpropathrin with a gas chromatograph.
 Sediment samples were dried with a vacuum filter. Fenpropathrin was extracted with an
 acetone/hexane (10:3) solvent by agitation with an ultrasonic generator for 15 min. The extract was
 then filtered through a flouride clay and silica gel column,  and further concentrated for quantification
 by gas chromatography.

        An acute toxicity test was carried out when fenpropathrin was added to the tap water.  Results
 are shown in Table 5.

                    Table 5.  Acute toxicity of fenpropathrin on Daphnia magna
                               (48 hr; water temperature 16-17 °C).
               Concentration (ppb)    1      1.8    3.2    5..4    7.5

               Mortality (%)          10     50     100    100    100
        The elimination of fenpropathrin in water and sediment is shown in Figs. 9A and 9B.  The
 fenpropathrin concentration in water decreased rapidly (Fig. 9B), but was rapidly absorbed at the same
 time by the sediment (Fig. 9A).  After 7 hours, 3.5 ppb  was measured in the water and 1800 ppb in
 the sediment. The half-life of fenpropathrin in water was between 3-7 days and no pesticide could be
 detected after 7 days. The decrease in fenpropathrin concentration corresponds with the mortality of
 Daphnia magna shown in Table 5.  The half-life of fenpropathrin in sediment was between 7-14 days
 and no pesticide could be detected after 45 days.  Furthermore, fenpropathrin residue in the sediment
 after 7 days was not found to be released into the water. It appears  that the elimination of
 fenpropathrin in the aquatic ecosystem is more rapid than in soil (Su and Fan, 1989).

                                       CONCLUSIONS

        Fenpropathrin is acutely toxic to some common  Chinese aquatic organisms. The toxic
 mechanism and lethal effect on fish may be related to inhibition of Na+,K+-ATPase, modification of
 the Na+ channel in the nervous system, disruption of ion metabolism, and ultrastructural damage to the
 gills.  However, this  pesticide did not persistently accumulate  in the  body since it was rapidly
 eliminated into the water.  Furthermore, the pesticide in  the water was  rapidly absorbed by the
sediment The residual fenpropathrin in the aquatic phase, and the pesticide absorbed by the sediment
were both degraded within a few days.  Despite the lack of persistence of fenpropathrin in the aquatic
ecosystem, it is recommended that this pesticide be used cautiously in fisheries  areas to avoid acute
toxic effects.
                                            58

-------
  fi
  £
  "1
  u
  .G
1.B
1.7 -
1.6 -
1,5 -
1.4 -
i.a -
1.2 -
1.1 -
  1 -
0.9 -
Q.B -
D.7 -
a.B -
a. a
                                                                                            ^,
                                                                                               14
                                              Tim* (
-------
                                         REFERENCES


 Berlin, J.R. et al., 1984. The inotropic effects of a synthetic pyrethroid decamehtrin on isolated guinea
        pig atrial muscle. Europ. T. Phannacol. 98: 313-322.
 Bradbury, S.P.,  1989.  Toxicokinetics and toxicodynamics of pyrethroid insecticides in fish.  Environ
        Toxicol. & Chem. 8: 373-380.
 Bradbury, S.P. et al., 1987.  Physiological response of rainbow trout (Salmo gairdneri) to acute
        fenvalerate intoxication. Pest. Biochem. Physiol. 27: 275-288.
 Casida, J.E. et al., 1983. Mechanisms of selective action of pyrethroid insecticides.  Ann. Rev.
        Phannacol. Toxicol. 23: 413-438.
 Coats, J.R. and O'Donnell-Jeffery, 1979. Toxicity of synthetic pyrethroid insecticides to rainbow
        trout.  Bull. Environ. Contain. Toxicol. 23: 250-255.
 Coats, J.R. et al., 1989.  Toxicology of synthetic pyrethroids in aquatic organisms: an overview.
        Environ. Toxicol. & Chem. 8: 671-679.
 Cole, L.M. et al., 1984.  Similar properties  of s-t-butyl biclo-phosphorothioate receptor-ionophore
        complex in brains of human, cow, rat, chicken and fish.  Life Sci. 35: 1755-1762.
 Khan, N.Y., 1983.  An assessment of the hazard of synthetic pyrethroids to fish habitat. In: "Mode of
        action, metabolism and toxicology" 3: 437-450.
 Kumuragura, 1982. Dkect and circulatory path of permethrin (NRX-143) causing histopathological
        changes in the gill of rainbow trout. J. Fish. Biol. 20: 87-91.
 Laurence, LJ. et al., 1983.  Stereospecific actions of pyrethroids on the GAMMA-aminobutyric and
        receptor-inophose comples.  Science 221: 1399-1401.
 Matsumura, F. et al., 1982. Influence of chlorination and pyrethroids on cellular calcium regulatory
        mechnisms.  In: Pesticide chemistry, human welfare and the environment.  Miyamote, I. and
        Kearney, P.C., eds. New York: Pergamon. 3: 3-13.
 Narahashi, T. et al., 1967. Mechanism of excitation block by the insecticide allethrin applied
        externally and internally to squid giant axon.  Toxicol. Appl. Pharmacol. 10: 529.
 Su, D. and Fan, D., 1989. Degradation and movement of fenpropathrin in soil. Acta Sci. Circ 9-
       446-453.
 Symonik, D.W. et al., 1989.  Effect of fenvalerate on metabolic ion dynamics in the fathead minnow
       and bluegill fish. Bull.  Environ. Contam. Toxicol. 42: 821-828.
Xu, L. et al., 1987.  Using the effects of environmental toxicants on the ATPase activity of grass carp
       tissues as an ecotoxicological index - a preliminary study.  Acta Hydrobiol. Sinica.  11:
        193-202.
                                            60

-------
             BIOACCUMULATION OF ORGANIC CHEMICALS
                               IN TISSUES OF FISHES

                           Robert V. Thurston1 and John F. Neuman1
                             Colin J. Brauner2 and David J. Randall2

                                     INTRODUCTION

       During the past 3 years we have been studying the rates of uptake by fishes of selected
synthetic organic chemicals, foreign to natural systems, known as xenobiotics.  We have conducted
several tests using an automated respirometer in which we have exposed fish to 1,2,4,5-
tetrachlorobenzene (TCB).  Our principal test animal has been the rainbow trout (Oncorhynchus
mykiss), although we have also tested other fish species to compare results. One of our long-range
objectives is to determine to what extent uptake and depuration rates of xenobiotics by fishes may be
related to their respiration rates, in the hope of developing a simple predictive model describing the
rates of uptake and depuration of xenobiotics as a function of the oxygen uptake rate and chemical
properties of the chemical in question.  In this presentation we will describe our approach to this
research and discuss some of our results.

                                        METHODS

       Test Conditions: Many of our preliminary tests have been conducted at Fisheries Bioassay
Laboratory, Montana State University (MSU), and most of our respirometry experiments have been
conducted in a modified Brett-type respirometer at the Zoology Department, University of British
Columbia (UBC). This respirometer was designed and built specifically for our studies, and has been
described in some detail by Gehrke et al. (1990).  The respirometer can test fishes up to 2 kg in size
and at swimming speeds as great as 2.5 m/second. Experiments can be conducted over several days
with water velocity,  temperature, pH, dissolved oxygen, and carbon dioxide all computer-controlled at
predetermined levels, and data monitored continuously by the computer as well.  For the tests at UBC
described here, fish were sacrificed immediately after  removal from the respirometer, and whole fish
or fish tissues, and test water samples, were frozen and stored at -20C until shipment to MSU for
chemical analysis; shipment was by overnight air carrier with samples packed in dry ice.  Fish tested
at MSU were prepared for analysis the same as those  tested at  UBC. All samples were stored at -30C
at MSU prior to analysis.

       TCB Analysis: Small whole fish (<10 g), subsamples of homogenates of larger fish, and
individual organ tissues, were blended to a fine powder with the aid of dry ice and anhydrous Na2SO4
while still frozen. At least two separate samples were prepared from homogenates of larger fish.
Only single samples were prepared from smaller fish and fish tissues because of limitations in sample
quantity. TCB was  soxhlet extracted (~>'8 hours) from each sample using hexane. Surrogate
(1,2,3,4-TCB) in amount comparable to the amount anticipated for TCB results (usually 100 ug) was
added at the initiation  of the extraction step.  Lipids were removed from portions of this extract using
florisil column chromatography; TCB and surrogate were eluted with 5% methyl-t-butyl ether.
Pentachlorobenzene  (PCB) internal standard (0.25 ng) was added to the purified extract and component
    fisheries Bioassay Laboratory, Montana State University, Bozeman, Montana 59717, U.S.A.

    Department of Zoology, University of British Columbia, Vancouver, B.C. V6T 1Z4, Canada

                                           61

-------
 concentrations were determined by electron capture gas chromatography (ECD-GC).  Packed GC
 columns (1.8 m x 2 mm i.d.) containing 3% SE-30 or 3% Carbowax on 100/120 Supelcoport were
 used. A multipoint calibration curve (0.5 - 10 pg/l) was employed during quantitation.  Both standards
 and sample extracts were injected in duplicate,  When plasma samples were required, the caudal
 peduncle was severed immediately after the fish was sacrificed and blood was collected from the
 caudal artery into heparinized capillary tubes.  Plasma was separated by centrifugation and 5.0 ul of
 plasma was removed by means of a 10 |jl glass syringe and transferred to a 10 ml volumetric flask
 containing 0.05 ug internal standard (PCB). This solution was made to volume with hexane, and the
 contents mixed for approximately 1 minute using a vortex mixer.  TCB was extracted from test water
 samples using hexane, and internal standard was added to the diluted extracts.  Both plasma and water
 samples were analyzed for TCB as described above.

                    TESTS CONDUCTED, RESULTS, AND DISCUSSION

        In several of our initial experiments rainbow trout were exposed to TCB for different time
 periods up to 6 hours under either static test conditions at MSU or forced swimming conditions in the
 respirometer at UBC. After test, fish were sacrificed and different tissues analyzed to determine any
 apparent sequence among them in increase of toxicant concentration.  Tissues analyzed were blood
 plasma, gills, liver, kidney, brain, spleen, heart, upper gut, lower gut, white muscle, pink muscle, and
 adipose. We also reviewed the data to determine if the rate of toxicant concentration increase in any
 one tissue might be representative of the fish as a whole.  Because TCB is lipophilic, this presupposes
 one tissue might have a  fat content representative of the fish as a whole, or at least with a fairly
 constant ratio to that of the fish as a whole. Standard deviations from the mean concentrations of
 TCB  among  the tissues analyzed were so great that we  were unable to detect any sequential build-up
 among the tissues.  Indeed, concentrations were not appreciably different between 1 and 6 hours
 except in the case of adipose tissue and possibly  muscle (Table 1). We also looked at the tissue/blood
 plasma ratio of the test fish, and once again did not see any consistent ratio (Table 2).

       It was apparent we would need to measure TCB concentrations in whole fish, and our next
 tests were designed to compare blood plasma TCB concentration against that of whole body over time.
 In our first experiment the concentrations of TCB in blood plasma  and whole body were very similar
 after 2 hours of exposure, but blood plasma concentration at 6 hours had not increased, whereas whole
 body  concentration had more than doubled (Figure 1).  The obviqus conclusion was that TCB
 equilibrium between exposure water and blood plasma was achieved within 2 hours or less, and after
 reaching saturation the blood was acting as  a conduit to pass TCB from gills to body tissues.  Thus,
 blood plasma could not be used to indicate total body content and therefore uptake rate of the toxicant.

       In our next experiment 5-g rainbow trout  were exposed to nominal TCB concentrations of 10,
50, and 100 ug/1 in holding tanks for several days, during which they were transferred to a fresh
solution every 24 hours.  Four fish were removed from each tank at 1, 2, 4, and 8 days, sacrificed, and
blood plasma and whole  body TCB concentrations measured.  A tight pattern of TCB increase over
time emerged for both blood plasma and whole body, and these patterns were also correlated with
exposure concentrations (Figure 2), but once again equilibrium between TCB in exposure water and
whole body lagged behind TCB equilibrium reached in blood plasma.
                                           62

-------












p
s
g
S3
o
1
«J
C
_o
a
^
8
4J
3
00
CO
TJ
•a
&;

T-H

H
E-i








"




1

•1
PQ


3
i
0
5*
i
cu
1
^

5

3
0

^
g
»^
<£•


|



is
5



CO
i
CTt
C§
CM

a

0
1 f















j«
^
o
PQ
c

ll
cQ

8


oo
T-H
T-H
£"


















o\
01

8
0

,_,
^F
o
T-H
oo
t-H
T-H

P


CO

o
U-l



„
T-H
V


0
oo


O

oo



oo
rj-
CO
>n


cs


>n
o
t~-
r— 1

O
a\
JQ

Vt
^f
1— 1
o
CO
T-H
CO

cs
oo
VO

t
c3
i— i
CO
•^


s
ON


\
CO
T^


>o
CO
VO
CO



VO
oo
s
o


01


0

s^

8
5?
*
^
ON
T-H
8
CO
oo

ON

oo
T— *

oo



50
c5
rH


>n
oo
>— 1


CO

ON
CO



oo
oo
g
o


CO


0
CO
CO
00

fO
u-i

0
ci)
T— t
CO
CO
CO
VO

r^

ON
1— 1
VO

oo
01



s
f-«.



R
1— <


r-
oo
T-H



cs
«n
ON
oo


CO


o
VO














J
in
T-H

r\

^
^
O
Vt

^c

T— (
f2


















81

>n
T-H
J5

CO
CO

0
oo
<^H

CO

oo
t^

CO
ON
T—t


OO
r-
cs



oo
co
vi


to
oo
f-



CO

£
CO


CN


o
T-H
00

o\
01
T-H
T-H

CO
T-H
VO
8
oo
T-H

1

^
t^
*
CO
CO


o\
01



VO


•n
0)
T-H



^

00
VO


ts


O
01
ON
>n
ON
OO

ON
T-H

VI
oo
CO

t~

Is!
•s
8
SI
rj-


T-H
01



ON
OJ
ON


OO
c~-*
c3
T-H



VO
ON
01
c--
co


01


0
VO


























!
g
o
II

*

S3
^
1

*

-S?
*P

-------
 g
 3
 o

 |H
 I
 3
 •a
rO

8
1
 o
 u
 o
 g
 CO
••d

I
•8
cs
•S
I
          CH

                         CO
                         5:
                        o<

                        §
                        i
                        o
                        VO,
                        rH
                        00

                        o
                       CO
                       C~7
                       in
                       cs
                       s
                             s
                             ON

                             O
                                    OO

                                    o
                             CO
                             «n
                             OO
                             
                                                        o\
                                                        CO
                                                        vS
                                                        VO
                                                        00
                                                        vd
                                                        fN
                                                       CN
                                                              OO

                                                              00
                                                              CM'
                                                              04
                                                              04
                                                              04

                                                             I
                                                              g
                                                              CO
                      VO
                      O
                                                              CO

                                                              3
                                                             CO
                                                             04
                                                             VO
                                                64

-------
                Q.
                Q.
                O
                <
                o:
                Ul
                o
                o
                o
                m
                o
20 -
                                        2                 4

                                           TIME - HOURS
                    Figure 1.  TCB plasma level vs whole body burden static
                               conditions - water level 260 ng/1.
       Earlier studies at MSU had shown the rate of uptake of some of the chemicals we might be
working with can be extremely rapid, with equilibrium between whole body and water sometimes
reached in just a few hours. Tischmak (1984) demonstrated that equilibrium for 2,4,6-trichloroaniline
between fathead minnows (Pimephales promelas) and their water environment was achieved in less
than a day (Figure 3), and although equilibrium for TCB required 6-8 days, it was also apparent that
the rate of uptake of TCB decreased considerably between initial exposure and 24 hours into test
(Figure 4).
                                              65

-------
         250 r
                                                                DOSE LEVEL
                                                                100 ua/i
                                                               DOSE LEVEL
                                                               SO ug/l
                                                               DOSE LEVEL
                                                               1Q uo/I
E
o>

I
<
    m
    o
         100
          50
                                                           DOSE LEVEL
                                                           100 ug/l
                                                           DOSE LEVEL
                                                           SO ug/l
                                                           DOSE LEVEL
                                                           10 ug/l
Figure 2.  TCB plasma vs whole body burden static conditions with periodic dosing.
                                      66

-------
              100,000 =
                IO,OOO
              o  1,000
                  100 Or
                           a
                           Q
                    10
                         G TCA Test I.
                         0 TCA Test2.
                         A Mixture
I   I   I   I   I   I  I   I   I   I   I   I   I   I   I   I
2  4  6  8  10 12  14  16  18 20 22 24 26 28 3O 32
         EXPOSURE  PERIOD (DAYS)
               (From Tischmak,  1984)
                           Figure 3.  TCA bioconcentrations profiles.

       Our next step was to conduct a series of short term tests in which we exposed rainbow trout to
TCB in the UBC respirometer to measure TCB  uptake vs. oxygen consumed.  We chose a 2-hour test
period for several reasons, but principally because this was long enough to measure oxygen differences
in the exposure water for as little as 100 g of test fish, and short enough so that rate of toxicant uptake
by the test fish would not measurably vary between start and finish of the test period. The uptake of
TCB by rainbow trout per gram of fish for three different weight classes is shown in Figure 5. We
also looked at uptake of TCB in relation to oxygen consumption rate for each of the three different
weight classes of rainbow trout, each tested at two different swimming speeds (Figure 6).

       Separate from our chemical uptake studies, we have compiled a file of data from the published
literature on respiratory oxygen requirements of fishes (Thurston and Gehrke,  1993).  This file, called
OXYREF, contains data from over 6800 individual laboratory tests in which oxygen consumption was
measured.  The data in our file include fish species, fish weight, certain test water conditions,
measured respiratory oxygen requirements, and  mode at which each fish was tested: "standard"
(resting), "routine" (moving about), or "active"  (measured swimming rate).
                                              67

-------
                 100,000
                  IO,OOO
                   IjOOO
                    IOO
                      10
                             A
                             G
A
o
A
0
                                                   ores Test I.
                                                   A Mixture
                          1   I   I   I   I  I   I

              L_JL
                             4  6 8  10 12 14 16 18 202224 26 283032
                                  EXPOSURE PERIOD (DAYS)
                 (From Tischmak,  1984)
                            Figure 4.  TCB bioconcentration profiles.
       From OXYREF, we plotted oxygen consumption vs. weight for fishes in an active swimming
mode, which is that mode under which we tested our fish at UBC. If we now superimpose on this
curve the data from our rainbow trout TCB uptake experiments, the species we tested in more than
one weight range, one sees an excellent fit; the bulk of those data points from the higher swimming
speed tests are slightly above the OXYREF curve, and the data points from the slower swimming
speed tests are slightly below the curve (Figure 7).

       There is a clear correlation between the rate  of oxygen uptake and the rate of uptake of TCB
among the freshwater fishes we have tested.  Thus, if the oxygen uptake rate and the water
concentration of a chemical to which a fish is exposed are known, it may be possible to  predict uptake
rates of that chemical during exposure. We are continuing to expand OXYREF as more data become
available in the literature, but in the present laboratory study if we can establish correlations between
oxygen consumption and toxicant uptake and depuration, then OXYREF can provide a powerful tool
for the prediction of bioaccumulation by fishes of xenobiotics.
                                            68

-------
    en
    3
   CD
   QL
   O
   in
   CO
   CD
   O
        10000
         1000
          100
           10
                         10          100

                           WEIGHT (g)
                                                A   HIGH BL/s


                                                O   LOW BL/s
                                           1000
               Figure 5.  TCB absorbed vs weight rainbow trout.
-C


3
    LU
    CD
    o
    co
    en
    m
    o
         10000
          1000
           100
            10
             1
             0.1
                                                367-469g
                                 31.9-46.1g


                           5.11-6.67g
                    1         10       100

                  OXYGEN CONSUMED  (mg/h)
     A   HIGH BU/l

     o   LOW eu/t
                                               y - 3X76 x""
                                               r - 0.995
                                               H - 0.990
                                               n - 48
1000
Figure 6.  TCB absorbed vs oxygen consumed rainbow trout at three weight ranges.
                                  69

-------
               o>
               £
              o
              Q.
              in
              o
              o
                                              — OXYREF ACTIVE CURVE WITH 99% Cl
                                              o  U8C RESPIROMETER TESTS
                                                 WITH RAINBOW TROUT
                                                                           1000
                                        FISH WEIGHT (g)
          Figure 7.  OXYREF data file active swimming regression curve vs rainbow trout
                                oxygen uptake during UBC tests.


                                  ACKNOWLEDGEMENT

       This research was funded by the U.S. Environmental Protection Agency, Environmental
Research Laboratory, Athens, Georgia, through.Cooperative Agreements CR 813424, CR 816369, and
CR 816778.

                                   LITERATURE CITED

Gehrke, P.C., L.E. Fidler, D.C. Mense, and DJ. Randall.  1990. A respirometer  with controlled water
       quality and computerized data acquisition for experiments with swimming fish. Fish
       Physiology and Biochemistry, 8:61-67.
Tischmak, DJ.  1984.  Separate and simultaneous bioconcentration in fathead minnows of five organic
       chemicals.  M.S. Thesis, Department of Chemistry, Montana State University, Bozeman
       Montana, USA.
Thurston, R.V., and P.C. Gehrke.  1993.  Respiratory oxygen requirements of fishes: description of
       OXYREF, a data file based on test results reported in the published literature, pp  95-108 -In'-
       Fish Physiology, Toxicology, and Water Quality Management.  Proceedings of an International
       Symposium, Sacramento, California, USA, September 1990.  U.S. Environmental Protection
       Agency, Environmental Research Laboratory, Athens, Georgia, USA. EPA/600/R-93/157.
                                             70

-------
      THE RELATIONSHIP BETWEEN THE UPTAKE OF ORGANIC
           CHEMICALS AND OXYGEN CONSUMPTION IN FISH

                              Colin J. Brauner1, John F. Neuman2,
                           Robert V. Thurston2, and David J. Randall1

                                        ABSTRACT

       Physiological models used to predict toxicant uptake are dependent on parameters such as
epithelial thickness, perfused area of the gills, gill blood and water flow rates and water boundary layer
thickness. These parameters are difficult to measure and are species specific. In addition, they are adjusted
in accordance with metabolic rate to ensure adequate oxygen delivery to  the tissues. Oxygen is a lipid
soluble molecule and crosses the gills transcellularly as do most lipid soluble xenobiotics. Thus, conditions
for oxygen transfer may be indicative of those for toxicant transfer.

       In five species of teleost fish similar in size (3 to 11 g) and for rainbow trout weighing 40 and
400 g, the uptake rate of a model xenobiotic,  1,2,4,5-tetrachlorobenzene (TCB, log  K^, of 5.0), is
dependent upon external TCB concentration and oxygen consumption rate of the fish. In 3-5 g fish, the
elevation of the regression line describing TCB uptake relative to oxygen consumption rate is significantly
greater than in fish larger  than llg. The slope of the regression lines; however, are not significantly
different, supporting the concept of using a general coefficient to predict toxicant uptake as a function of
consumption rate (MO2) in  fish. The greater elevation of the regression line in the small fish data is likely
due to TCB binding to the  body surface and reflects the relationship between surface area and volume.

       The  coefficient relating the initial uptake of TCB with MO2 in fish has been used to predict the
uptake of tetrachloroguaiacol in 400 g rainbow trout and 700  g large scale suckers. Tetrachloroguaiacol
has a similar log K^  (4.41) to TCB and the predicted levels approximate  the measured values.
       We have created a large oxygen data base OXYREF which can be used to estimate the oxygen
consumption rate of a fish based on fish size and activity level. Potentially, the coefficient relating toxicant
uptake with MO2 can be multiplied by the MO2 value retrieved from OXYREF to predict chemical uptake
rates. The coefficient can be corrected to account for the uptake of different chemicals provided log K^,,
weak acid dissociation constant, water pH and freely dissolved concentration of the chemical are known.

                                      INTRODUCTION

       With the advancement of technology there is a massive production of chemicals synthesized to
meet the demands of industry. Most of  these chemicals are foreign to the body (xenobiotics) and
eventually enter the environment where they may be available for uptake by organisms. Once within the
body many xenobiotics are extremely toxic; particularly the lipid  soluble, chlorinated organics which are
not susceptible to oxidation by the multi-function oxidases. The mode of toxicity is often specific to each
xenobiotic and may be localized to a specific target site or more likely it may act at multiple sites. Thus,
a chemicals mode of toxicity is difficult to quantify, let alone predict. Most xenobiotics, however, have
one thing in common;  they must enter the animal before they can exert their toxic effect(s). Uptake can
    Department of Zoology, University of British Columbia, Canada

    fisheries Bioassay Laboratory, Montana State University, U.S.A.

                                            71

-------
 occur through biomagnification in the food chain or by direct uptake from the environment across the
 respiratory surface or skin. This discussion is concerned with direct uptake of xenobiotics from water by
 fish.

         Generally, the gills comprise the major surface area of the body in fish. In addition, this boundary
 between the animal and the environment is only a few  cells thick (1-10 u) and is highly  perfused with
 blood which flows counter currently to water. These conditions ensure adequate gas exchange across the
 gills but also make the gills an ideal organ for xenobiotic uptake. The  gills are often thought to be the
 major site for xenobiotic uptake (Neely, 1979, Gobas, Opperhuizen and Hutzinger, 1986, Randall and
 Brauner, 1993); however, uptake across the skin can not be neglected, particularly in small fish when the
 body surface  area can make up a large proportion of the animals total surface area.

         The uptake of xenobiotics by fish directly from water is determined by numerous factors, the most
 dominant of which are the physico-chemical properties of the compound and the physiological processes
 which occur at the gills.

         PHYSICO-CHEMICAL CHARACTERISTICS WHICH INFLUENCE UPTAKE

         The physico-chemical characteristics which most strongly influence the flux of xenobiotics across
 biological membranes are: water and lipid solubility, molecular weight and volume, "active concentration",
 tendency to ionize, and susceptibility to metabolic transformation.  The amount of toxicant which can be
.carried in  the inhalant water is determined by the chemicals water solubility; The rate  at which the
 chemical can  diffuse across the lipid membrane of the respiratory epithelium from inhalant water will in
 turn  depend on lipid solubility. Subsequent distribution of the chemical across the lipid and aqueous
 barriers of neighbouring cells will be determined by a compromise between lipid and water solubility. The
 importance of a ratio between lipid and water solubility to toxicant uptake has long been recognized and
 is the basis for the well established octanol:water partition coefficient (K.J. Octanol was chosen to
 represent "biological lipids"  and K^ became the standard for quantifying the  partitioning characteristics
 of compounds (Connell, 1990).

        Xenobiotics diffuse across membranes transcellularly and thus diffusion  depends largely on lipid
 solubility but  molecular weight and volume of a compound cannot be ignored. In general, the diffusion
 coefficient of a chemical decreases as  molecular weight  increases.  Saito et al. (1990) used electron
 microscopy to qualitatively demonstrate that compounds greater than 2000 in molecular weight are not
 absorbed across the gills in carp.
Zitko and Hutzinger (1976) have proposed that the upper molecular weight limit for xenobiotic uptake is
approximately 600.

       The driving force for diffusion is supplied by the "active concentration" gradient between the
environment and the fish. Most xenobiotics are extremely hydrophobic and  consequently will bind to
organic matter (Black and McCarthy, 1988) or become oriented into  small, hydrophilic droplets called
micelles. In both situations, the chemical is not aqueously dissolved and therefore not available for uptake
by diffusion. The uptake rate of a xenobiotic to which a fish has not previously been exposed is directly
proportional to the aqueously dissolved concentration of the compound in the environment. Thus, it is
crucial to know the "active concentration" or the concentration of the chemical available for uptake.

       Lipids are nonpolar and thus noncharged molecules diffuse across lipid membranes much more
easily than charged molecules. This is of significance in the uptake of compounds which dissociate into
weak acids because the membrane diffusivity of the dissociated form of the compound is very low relative
to the non-dissociated form. The dissociated proportion of a compound is dependent upon the pH of the

                                             72

-------
medium in which it exists and its tendency to dissociate which is quantified by the weak acid dissociation
constant (pKJ. The non-dissociated, aqueously dissolved concentration of the compound in water can be
considered the "active concentration" which is available for uptake by the fish. This concentration can be
calculated if water pH and the pK^ of the compound are known. The pH required for this calculation is
that of the water in the interlamellar channels of the gills (see Randall this volume). Saarikoski et al. 1987
attempted to predict uptake of a variety of weak acids in guppies by taking into account the water pH and
compound pKa but were only partially successful.

       The final physico-chemical characteristic mentioned above is the resistance of the chemical to the
action of multi-function oxidases (MFO). These enzymes convert compounds to a less hydrophobic form
which reduces the toxicity of the  xenobiotic and  facilitates excretion of the compound. Multi-function
oxidase activity in fish has been demonstrated in many tissues but exists predominantly in the liver and
gills (Barren et al. 1989). If xenobiotics are oxidized at the gills before they enter the fish or are converted
to a less toxic form  at the same  rate that they are taken up,  the compounds will not bioaccumulate.
Unfortunately, the most harmful  xenobiotics are often resistant  to metabolic transformation  which
contributes to the toxicity of the compound.

       All the characteristics described above  can effect uptake of xenobiotics. However, in general, the
most important physico-chemical characteristics in predicting toxicant movement across membranes are
the KOW and the "active concentration" of the compound which encompass both the concentration of the
aqueously dissolved compound and the concentration of the non-dissociated form if the chemical is a weak
acid.

                       XENOBIOTIC UPTAKE ACROSS FISH GILLS

       There has been considerable effort to model the uptake of xenobiotics using a variety of different
types of models which range from  simple, one  compartment models to complex pharmacokinetic models
(Barron et al. 1990). Hayton and Barron (1990) have recently  proposed a physiological model to describe
xenobiotic uptake across the gills in fish:

Uptake rate = (Cw - C^^D^) A Km)+(h(Da-1)A)+Vb(Kb-1+Vw-1]-1                             eq.(l)

       This model consists of three components: 1.) the concentration gradient between the environment
and the fish, where Cw and Cf are the concentration of the xenobiotic in the  external water and the plasma
water respectively, 2.)  physiological  and anatomical  characteristics of the gills, where d is epithelial
thickness, h is aqueous stagnant layer thickness, A is gill surface area, Vb is effective gill blood flow, Vw
effective water flow, and 3.) physical constants specific to the compound, where Dm is diffusion coefficient
in epithelium, Da is diffusion coefficient in water,  K^ is epithelium/water distribution coefficient, and K^,
is the blood/water distribution coefficient.

       During initial exposure to  a xenobiotic, the first  component, the concentration gradient between
the environment and the fish, will be equal to the concentration of the xenobiotic in the environment. This
will persist for some time because immediately  after entering the blood the lipophilic compounds will bind
extensively with blood lipid and proteins (Schmeider and Henry,  1988) keeping the aqueously dissolved
concentration of the xenobiotic in  the plasma water very low.

       The second component describing the physiological and anatomical characteristics of the gills are
not easy  to quantify.  For example, it is very difficult to estimate the surface area of the gill and even if
an anatomical value is derived it probably does not reflect the functional surface area of the gills because
in resting fish, only 60% of the gill is perfused  with blood (Booth, 1978). The thickness of the respiratory

                                            73

-------
 epithelium ranges from 0.5 to 11 jam depending upon the species (Hughes, 1984) and the thickness is
 probably not constant from the proximal to distal portion of the secondary lamellae. To further complicate
 matters many model parameters are not constant and are influenced by the metabolic rate of the fish. For
 instance during exercise, there is an increase in cardiac output and ventral aortic blood pressure (Kiceniuk
 and Jones, 1977) which increases the perfused gill surface  area and gill blood flow, and reduces the
 thickness of the respiratory epithelium (Jones and Randall, 1978). The thickness of the aqueous stagnant
 layer or boundary layer will also be greatly influenced by the velocity of water flow over the gills. Thus,
 all of these parameters are difficult to measure, they are species specific but most importantly they are all
 adjusted according to the metabolic demand of the animal to  ensure adequate gas exchange at the gills.

        Oxygen is a  lipid soluble molecule as  are most xenobiotics. Therefore, conditions for oxygen
 transfer may be indicative of those for toxicant transfer. The development of a coefficient (X) describing
 xenobiotic uptake as a function of oxygen consumption rate can be used to replace the physiological and
 anatomical parameters of the gills in equation 1.  This  coefficient can then be multiplied by oxygen
 consumption rate (MO2) and be incorporated into a general model to predict xenobiotic uptake. Oxygen
 consumption rate is easy to measure or it can be retrieved from an oxygen data bank (OXYREF) compiled
 by Thurston and  Gehrke (1990) (see Thurston et al, this volume). This bank contains over 15 000
 measurements of oxygen consumption rate in over 300 species of fish exposed to different temperatures
 at rest and during exercise. Analyses  of these data indicate that the main determinant of oxygen
 consumption rate in fish is body weight.

        The third component in equation 1 has to do with characteristics specific to the xenobiotic to
 which the fish is exposed. As mentioned above the most important physico-chemical characteristic which
 determines  xenobiotic uptake rate is K^ Both K^ and K,, can be calculated from the K^  of the
 xenobiotic. Saarikoski et al.  (1987) have demonstrated in guppies that the logarithm of chemical uptake
 rate increases linearly with log K^ between 1 and 4.3 for a variety of compounds. The uptake  rate is
 approximately constant for compounds with log K,,w between 4.3 and 6 and decreases with log K^, greater
 than 6 (McKim, Scheider and Veith, 1985). We chose 1,2,4,5 tetrachlorobenzene (TCB) as our model
 toxicant to calculate X because it does not dissociate, is not easily metabolized and has a log K<,w of 4.97
 which is in the range  where K^, does not effect maximum uptake rate (Saarikoski et al. 1987). The data
 of Saarikoski et al. (1987) may be used to derive a coefficient (K,,w coeff.) which can be multiplied by X
 to predict the uptake of xenobiotics which differ in log K^ from TCB.  Thus, during initial exposure to
 a compound, equation (1) can be simplified to:
Uptake rate = (Cw) X (K,w coeff.)
eq. (2)
        where X is a coefficient describing xenobiotic uptake as a function of oxygen consumption rate,
and K,,w coeff. is the coefficient used to account for uptake in xenobiotics which differ in log K<,w from
TCB. For xenobiotics with log K^, between 4.3 and 6, this coefficient is 1, for those between 1 and 4.3
this coefficient is io°-429k>sKow-1-842 (Randall and Brauner, 1993 calculated from Saarikoski et al. 1987).

                                     DERIVATION OF X

        One of the requirements of using X to predict the uptake of xenobiotics in fish based upon MO2
values retrieved from OXYREF, is that there is no influence of TCB exposure on MO2. In rainbow trout
exposed to a large range of TCB concentrations there is no significant effect of toxicant exposure  on
oxygen  consumption rate in resting and active  fish (Fig. la and Ib). Thus, there are no direct effects of
TCB exposure to complicate the relationship between toxicant uptake and oxygen consumption rate in fish.
                                            74

-------
  CO I
        100 r
         90
         80
< 'oo 70
O  J.
55 J. 50

       50
         40
                o
            0
                       o
                     500
1000
1500
                                                                  o

                                            2000
2500
                                                 -1,
                                WATER TCB (ug-1   )
 X
 §
        300 r
        250 cx
     QD 200
         100
0
                     0
                       o
                            oo
                       200
                                   o
                                                     o
                                 400
            600
            800
 1000
                                                  -1,
                                WATER TCB (ug-1  )
Figure 1.   A) The effect of water 1,2,4,5 tetrachlorobenzene (TCB) concentration on the oxygen
consumption rate of adult rainbow trout at rest (^=0.01); or B) swimming at 1.25 Body lengths per second
(Bis'1) (r^O.02).
                                    75

-------
           30 r
           25
      7    20
 I   V5
 CQ     bo
 a     S  10
             o
              100
200             300

       M02  (mg-kg   • hr   )
400
500
 Figure 2. The effect of oxygen consumption rate on TCB uptake rate (mg-kg^hr1) in five species of fish
 exposed to one of two  external TCB concentrations  (260, open symbols, or 780 ugt1 TCB, closed
 symbols) at a water velocity of 2.25 or 3.75 Bis'1 for 2h.(circle, goldfish; diamond, largemouth bass;
 upright triangle, channel catfish; inverted triangle, fathead minnow; and square, rainbow trout) (n=8,
 vertical lines represent 1 standard error, ^=0.72 at the high TCB concentration, and ^=0.29 at the low
 concentration).

        The relationship  between MO2 and TCB uptake rate was examined in five species of fish similar
 in size, and for several different weight classes of rainbow trout. Forced exercise was chosen as the means
 to influence oxygen consumption rate because during exercise, fish can increase MO2 relative to resting
 rates by up to 10 fold (Brett, 1964). In all tests, fish were introduced into a 130 1 Brett-type respirometer
 at the University of British Columbia (described by Gehrke et al. 1989) for two hours at a water velocity
 of 18 em's'1. Tetrachlorobenzene  was added to the respirometer, a water sample was taken and fish were
 forced to swim for 2 h at either a low or a high water velocity (less than  80% of maximal swimming
 velocity) while MO2 was measured. At the end of the test, a water sample  was taken and the fish were
 removed from the respirometer,  killed and immediately frozen. The fish and water samples were  kept
 frozen at -80°C and then  transported to Montana State University in dry ice for TCB analyses. The entire
 fish was homogenized, TCB from the tissue was concentrated in hexane by Soxhlet extraction  and total
 body and water TCB concentration was measured by Gas Chromatography.

       Five species offish: goldfish (Carrasius auratus, 11.43  ± 0.99 g), large mouth bass (Micropterus
 salmoides, 4.8 ± 0.18 g), channel  catfish (Ictalurus punctatus, 3.7 ± 0.06 g), fathead minnow (Pimephales
promelas, 3.92 ± 0.34 g)  and rainbow trout (Oncorhynchus mykiss, 5.02 ± 0.07g) were exposed to water
                                         76

-------
TCB concentrations of 260 or 780 jiT1 at low or high swimming velocity with exception of the large
mouth bass which were only capable of swimming at the low velocity. There is a significant correlation
between TCB uptake rate and MO2 at both the low TCB concentration (?= 0.29) and the high TCB
concentration (r2= 0.79)  (Fig. 2). The rate  of toxicant uptake at  a given oxygen consumption rate is
dependent upon water TCB concentration (Fig. 2). During initial exposure to a toxicant, the gradient for
toxicant uptake is equal  to the toxicant concentration in the water and uptake  is proportional  to the
external concentration (Spacie and Hamelink, 1982).

       In an attempt to standardize for external TCB concentration and generate one regression equation
describing toxicant uptake relative to MO2, the measured TCB uptake rate was divided by the mean water
TCB  concentration  over the exposure  period (Fig. 3).  It is apparent; however,  that two distinct
relationships exist, one for each TCB exposure concentration. One explanation for this is that the measured
TCB concentration in the water is an overestimation of the aqueous concentration. The high water TCB
concentration (760 ngl~l) is near to the maximal water solubility for TCB and it is possible that TCB is
binding to organic matter or forming micelles. Thus, the measured content of TCB in the water is greater
than the aqueously dissolved  concentration.  An overestimation of the aqueous TCB concentration in the
high TCB treatment would result in a lower placement of the regression line.
a,
6Q
 E-
 O,
 D
 OQ
     E-
O
          100 r-
           75
            50
            25
                                                                V
                                                  D
              100
                         200            300
                                M02  (mg-kg
400
500
                                                    -i
Figure 3. The effect of oxygen consumption rate on the TCB uptake (mg kg'1 hr'1) per unit gradient TCB
(mg-r1) in five  species of fish. See Fig. 2 for  further  explanation. (n=8, r^O.79 at the high TCB
concentration, and ^=0.317 at the low concentration).
                                         77

-------
     CU
            8
           0
                       goldfish
                                 bass
                                                 catfish
                                                        fathead
                                                        minno-w
              rainbow
              trout
DQ
O
E-
    Figure 4. The percent body lipid determined for five species of fish. * indicates statistically different from
    those without symbol. (n=8)
         100 r
          80
5
<
a:
o
          60
          40
          20
           0
             0
                                10
15
20
25
                                                      "1
                                       (mg-kg   • hr  ) x LIPID
30
    Figure 5. The relationship between TCB uptake per unit gradient and the product of oxygen consumption
    rate and proportion of body lipid, in five species of fish. See Fig. 2 for further explanation. (n=8, ^=0.69
    at the high TCB concentration and ^=0.85 at the low concentration).
                                               78

-------
       Lipid content in fish is important during long term exposure to xenobiotics particularly as fish
approach equilibrium with the environment (Connell, 1990, Geyer et al. 1985) but the importance of lipid
content to toxicant uptake during initial exposure is not known. The lipid content of the five species was
measured and significant differences were found between goldfish and bass relative to channel catfish,
fathead minnow and rainbow trout (Fig. 4). The inclusion of lipid in the relationship describing toxicant
uptake as a function of MO2 significantly improved the coefficient of determination at the low TCB
concentration (^=0.85) and only marginally reduced the coefficient in the high TCB concentration (from
0.79 in fig. 3 to 0.68) (Fig. 5).  These results indicate that the  body lipid content in fish may be an
important determinant of toxicant uptake during initial exposure  to TCB; however this requires further
investigation. The data presented so far indicates that for five species of fish weighing between 3  and 1 Ig,
there is a significant relationship between TCB uptake rate and MO2 during initial exposure to TCB. Does
this relationship exist for fish over a range of body masses?
cu
E-
tx
oa
o
E-
o
E-
          100  i-
            80
            60
           40
           20
             0
              100
                         200             300

                                M02 (mg-kg
                                                                 400
500
                                                     -1
Figure 6. The effect of oxygen consumption rate on the TCB uptake (mg-kg'1 'hr'1) per unit gradient TCB
(mg-r1) in medium (39 g, small solid square, n=4) and large (412 g, large solid square, n=8) rainbow trout
exposed to 260 ugT1 TCB (^=0.70).

       Medium (38.6 ± 0.96) and large (412.5 ± 6.9) rainbow, trout were forced to swim at one of two
water velocities during exposure to  an aqueous TCB concentration of 260 ul"1. Toxicant uptake rate was
divided by the mean TCB concentration during the exposure duration and when regressed against MO2
the coefficient of determination is significant (r2=0.59)(Fig. 6). However, the slope of this regression line
differs from that for the five species of small fish exposed to the low TCB concentration (Fig. 3). In Fig.
3, the low coefficient of determination (^=0.32) for the five species of small fish exposed to the low TCB
concentration is predominantly due to the low TCB uptake rate measured in goldfish and large mouth bass.
The goldfish are more than twice the size of the other fish in this group and when the goldfish data are
combined with the large rainbow trout data, there are still two distinct regression lines but only the
                                          79

-------
  elevations differ significantly,  the slopes do not (Fig. 6). The constant slope of the regression lines
  indicates that TCB uptake rate relative to M02 is the same for all fish sizes and justifies the use of MO,
  as a general tool in the prediction of xenobiotic uptake. The point at which the regression lines intersect
  the abscissa (when M02 is zero) likely reflects TCB binding to the body surface. The proportion of whole
  animal TCB uptake rate which  is determined by TCB binding to the skin will not be great in large fish
  but may be quite large in small fish where the surface area to body mass ratio is large. The significant
  difference between regression line elevations probably reflects this difference in surface area to body mass
  ratios between the two groups. These data indicate that in fish greater than 1 Ig there is very little effect
  of fish size on the uptake rate of TCB due to TCB binding  to the skin, but there is a large effect in fish
  less than llg. Therefore, in fish greater than llg, TCB uptake during initial exposure in  fish can be
  predicted for a variety offish species using equation (2) where k= 15.63 + 0.0820MO2
          TOO  r
            80
§   Q
£   §
CQ   2:
£   D
            60
           40
           20
             o
             100
200
                                               300
                                  400
500
                                     M02
Figure 7. The effect of oxygen consumption rate on the TCB uptake (mg-kg-'-hr1) per unit gradient TCB
(mgl ) where open symbols are data for small fish from Fig. 3 and closed symbols are data for medium
and large rainbow trout from Fig 6. The goldfish data (open circles) are grouped with the medium to large
rainbow trout data (solid squares)  (iM).59, equation of regression line is Y=15 63+0 0820X) The
                     SrouPed  into  ^cond  regression  (r*=0.22, equation  of regression  line  is
       As discussed earlier, there is very little effect of log K^ on log uptake rate for xenobiotics with
log K™ between 4.3 and 6 (Saarikoski et al. 1987). Tetrachloroguaiacol (TCG) has a log K,,  of 4 41 (Xie
and Dryssen, 1984) and was chosen to determine whether the relationship between TCB uptake rate and
MO2 can be applied to other xenobiotics. Rainbow trout (409.5 ± 0.02 g) were exposed to 260 nT1 TCG
at low or high swimming speed for 2 h and large scale suckers (Catastomus macroheilus) (706 ± 0.05 g)
which could not be forced to swim at a high velocity, were exposed to 260 nT1 TCG at two different
                                            80

-------
temperatures (9 or 18°C). Tetrachloroguaiacol dissociates and the concentration of the non-dissociated
form was calculated from water pH and the pR, for TCG (6.19, Xie and Dryssen, 1984). The uptake rate
of TCG was divided by the concentration of non-dissociated TCG and is significantly correlated with MO2
(r2=0.97)(Fig. 8). The regression line (and 95% confidence interval) describing TCB uptake rate relative
to MO2 in fish greater than 1 Ig (from Fig. 7) is also plotted in figure 8 and approximates the TCG data.
.Thus, the equation of the regression line describing TCB uptake rate relative to MO2 can be used to
predict the uptake of TCG and likely other xenobiotics similar in log
K
6Q
CU
 E-
 CU
U

O
5
 X
     z.
     Q

     o
           100 r
            80
            60
            40
            20
              0
                0
                           100
200
300
400
500
                                                     	 i     	 i
                                      M02  (mg-kg  • hr   )
Figure 8. The effect of oxygen consumption rate on tetrachloroguaiacol (TCG) uptake (mg-kg^hr1) per
unit gradient TCB (mgT1) in rainbow trout 410 g  (solid squares, n=8) and large  scale suckers 706 g
(n=8)(r2=0.97). The regression line and 95% confidence interval is for TCB uptake  per unit gradient for
goldfish, medium and large rainbow trout (the lower regression line) in Fig. 7.

       The  model proposed above at present  is  restricted to conditions of normoxia, during initial
exposure to xenobiotics. It is simple, and requires limited information about the xenobiotic to which a fish
is exposed (log K^, and pK^ if it is a weak acid) and the metabolic rate of the fish. The metabolic rate of
the fish can be measured directly or estimated from the oxygen database OXYREF. The use of this model
to  predict the uptake of xenobiotics over the log K^, range of 1 to 4.3 remains to be  tested. We are
presently testing and expanding this model to predict xenobiotic uptake over a longer  duration of exposure
as  the animal reaches equilibrium with the  environment.
                                          81

-------
                                   ACKNOWLEDGEMENT

        This  work  was  supported in  part by Cooperative Agreement CR816369  from The U.S.
Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia.

                                        REFERENCES

Barren, M.G., Schultz, I.R., and W.L. Hayton. 1989. Presystemic branchial metabolism limits di-2-
   ethylhexyl pthalate accumulation in fish. Toxicol. Appl. Pharmacol. 98:49-57.
Barron, M.G., Stehly, G.R. and Hayton, W.L.  1990. Pharmacokinetic modeling in aquatic animals. I.
   Models and concepts. Aquatic Toxicology 18:61-86.
Black, M.C., and J.F. McCarthy. 1988.  Dissolved  organic macromolecules reduce the uptake of
   hydrophobia organic contaminants by the gills of rainbow trout (Salmo gairdneri). Environ. Toxicol.
   Chem. 7:593-600.
Booth, J.H. 1978. The distribution of blood flow in the gills of fish: application of a new technique to
   rainbow trout (Salmo gairdneri). J. exp. Biol. 73:119-129.
Brett, J.R.  1964. The respiratory metabolism and swimming performance of young sockeye salmon. J.
   Fish. Res. Bd. CANADA, 21(5):1183-1226.
Connell, D.W. 1990. Bioaccumulation of xenobiotic compounds. CRC Press, INC., Boca Raton, Florida.
   213pp.Black and McCarthy, 1988
Dobbs, A.J. and N. Williams. 1983. Fat solubility- a property of environmental relevance? Chemosphere.
   12(1):97-104.
Gehrke, P.C., Fidler, L.E., Mense, D.C. and D.J. Randall 1990. A respirometer with  controlled water
   quality and computerized data acquisition for experiments with swimming  fish. Fish Physiology and
   Biochemistry Vol. 8(l):61-67.
Geyer, H., Scheunert, I. and F. Korte.  1985. Relationship between the  lipid content of fish and their
   bioconcentration potential of 1,2,4-Trichlorobenzene. Chemosphere 14(5):545-555.
Gobas, F.A.P.C., Opperhuizen, A. and O. Hutzinger. 1986. Bioconcentration of hydrophobic chemicals
   in fish: relationship with membrane permeation. Environ. Toxicol. Chem. 5:637-646.
Hayton, W.L. and M.G. Barron. 1990. Rate-limiting barriers to xenobiotic uptake by the gill.  Environ.
   Toxicol. Chem. 9:151-157.
Hughes, G.M. 1984. General anatomy of the gills. In: Fish Physiology (W.S. Hoar and D.J. Randall eds.)
   vol 10 A ppl-63. Academic Press, New York.
Jones, D.R. and Randall, D.J. 1978. The respiratory and circulatory systems during exercise. In: Fish
   Physiology. (Hoar, W.S. and D.J. Randall, eds.). Vol 7. pp. 425-501.
Kiceniuk, J.W. and Jones, D.R. 1977. The oxygen transport system in trout  (Salmo gairdneri) during
   sustained exercise. J. Exp. Biol. 69:247-260.
McKim, J., Schmieder, P., and G. Veith. 1985. Absorption dynamics of organic chemical transport across
   trout gills  as related to octanol-water partition coefficient Toxicol. Appl. Pharmacol. 77:1-10.
Neely, W.B. 1979. Estimating rate constants for the uptake and clearance of chemicals by fish.  Environ.
   Sd. Technol. 13(12):1506-1510.
Randall, D.J.  and C.J. Brauner. 1993.  Toxicant uptake across fish gills. Proceedings of an International
   symposium, Sacremento, California, September  18-20,  1990. United States  Environmental Protection
   Agency. Fish Physiology, Fish Toxicology, and Fisheries Management. (In Press).
Saarikoski, J., Lindstrom, R., Tyynela, M. and M.  Viluksela. 1986. Factors affecting the absorption of
   phenolics  and carboxylic acids in the  guppy  (Poecilia reticulatd).  Ecotoxicol. Environ. Safety.
   11:158-173.
Saito. S., Tateno, C. Tanoue, A. and T.  Matsuda. 1990. Electron microscope autoradiographic examination
   of uptake behaviour of lipophilic chemicals into fish gill.  Ecotoxicol. Environ. Safety.  19:184-191.
                                            82

-------
Schmieder, P.K.  and T. A.  Henry.  1988. Plasma  binding  of 1-butanol, phenol,  nitrobenzene and
   pentachlorophenol in the rainbow  trout and rat: a comparative study. Comp. Biochem. Physiol.
   91C(2):413-418.
Spacie, A. and J.L. Hamelink. 1982. Alternative models for describing the bioconcentration of organics
   in fish. Env.Toxicol. Chem. 1:309-320.
Thurston, R.V. and  Gehrke P.C.  1993. Respiratory oxygen requirements of fishes: Description  of
   OXYREF, a datafile based on test results reported in the published literature. Proceedings of  an
   International symposium, Sacremento, California, September 18-20,1990. United States Environmental
   Protection Agency. Fish Physiology, Fish Toxicology,  and Fisheries Management. (In Press).
Xie, T.M. and Dryssen, D. 1984. Simultaneous determination of partition coefficients and acidity constants
  'of chlorinated phenols and guaiacols by gas chrQmotaography. Analytical Chimica Acta 160:21-30.
Zitko,  V.  and  Hutzinger,  O.  1976. Uptake of  chloro-  and  bromobiphenyls,  hexachloro- and
   hexabromobenzene by fish. Bull. Environ. Contain. Toxicol.  16:665-673.
                                            83

-------
                   TOXIC EFFECTS  OF MEOTHRIN ON THE
                  GILL ULTRASTRUCTURE IN GRASS CARP,
                          CTENOPARYNGODONIDELLUS

                         Bingsheng Zhou1, Yongyan Zhang1 and Ying Xu1

                                         ABSTRACT

        Fingerling Grass carp, Ctenopharyngodon idellus, were exposed to two different concentrations
 of meothrin close to the 24 hr LC50,  and gill morphology subsequently examined by light and electron
 microscopy.  Fish exposed to meothrin experienced convulsions, coughing, ataxia, intermittent
 paralysis and death. Post-mortem examination revealed significant morphological alterations in
 exposed fish compared to control fish. In both exposed groups, short exposure to meothrin resulted in
 hypertrophy of the secondary lamellae, which degraded to local telangiectasia (fusion between
 secondary lamellae) upon longer exposure.  Ultrastructural changes in the gills included collapse of the
 pillar cell system which resulted in large non-tissue spaces which were invaded by leucocytes. Myelin
 figures and electron-dense deposits were accumulated in the cytoplasm of the epithelial cells. Chloride
 cells appeared severely damaged, showing degenerated mitochondria and ruptured nuclear membranes.
 The higher concentration of meothrin  caused a loss of adhesion between the epithelial cells,
 accompanied by collapse of the structural integrity of the primary lamellae and degeneration of
 epithelial and chloride cells.

                                      INTRODUCTION

        Meothrin or fenpropathrin ((s)-a-cyano-3-phenoxybenzyl(s)-a-(4-chlorophenyl)-3-
 methylbutylate), a synthetic pyrethroidal insecticide, is used to control insects and mites on grain and
 cotton crops, fruit trees, vegetables, flowers and other crops. Meothrin is therefore wildly distributed
 in the surrounding environment.  There have been many reports of cypermethrin uptake,  depuration,
 metabolism and toxicology in fish and aquatic animals (Coats et al., 1979; Mclease, 1980). Although
 some reports of meothrin hydrolysis in water and its metabolism in plants have been published
 (Mikami et al.,  1985), there are only a few systematic descriptions of the toxic syndromes associated
 with exposure to the pesticide.  Such descriptions are vital to an understanding of the mechanisms of
 action of the pesticide.  Meothrin is highly toxic to fish, with a 24 hr  LC50 of 2.2 ppb. No detailed
 histological studies of the effects of meothrin on aquatic animals are available, and the exact
 mechanism of meothrin poisoning in fish has not been ascertained.  The gill, which serves as a major
 organ for respiration and osmotic regulation, is easily damaged by even low concentrations of
 numerous pollutants (Skidmore & Tovell, 1972; Papathanossiou & King, 1983; Lars et al., 1983;
 Karlsson et al.,  1985; Waster et al., 1988) and pesticides (Drewett & Abel, 1983). Zinc,  copper,
 cadmium, sodium bromide (Waster et  al., 1988), as well as lindane and DDT, have been  reported to
 cause histopathological changes in fish gill and other organ structures  (Mathur, 1962).  Gill damage
 can, of course, be severely debilitating to fish in many ways, including impairing oxygen uptake (Lars
 et al., 1983; Waster et al., 1988).  The purpose  of this paper is to identify the histopathological effects
 of meothrin on the gill structure of the Grass carp using light and electron microscopy, and to evaluate
 potential mechanisms involved in the reaction of gill tissue to this pesticide.
    'State Key Laboratory of Freshwater Ecology and Biotechnology, Institute of Hydrobiology, Chinese
Academy of Sciences, Wuhan, 430072

                                           84

-------
                               MATERIALS AND METHODS

EXPOSURE OF FISH TO MEOTHRIN

       Seventy Grass carp, with a wet weight of 9.5-10.5 g, and a total length of 10-11.5 cm, were
obtained from the fish pond of the Institute. The fish were acclimated for 1 week at 23±1 °C in
aquaria filled with tap water filtered by active carbon. A stock solution of 99.9% pure meothrin
(Sumitomo, Japan) was used in all experiments.  Test solutions of 1 and 3 ppb were prepared by
adding the appropriate quantity of meothrin acetone solution to water to make a final volume of 10 L.
Control solutions contained the same concentration of acetone as the test solutions. Test solutions
were completely replaced with a fresh solution every 12 hr in the 1 ppb test group and every 2 hr in
the 3 ppb test group by siphoning out the old solution and replacing it with fresh solution.

PREPARATION OF THE SAMPLES

       Fish gills were removed from experimental animals after 6, 12, 24, 48, and 72 hr of exposure
in the 1 ppb test group and 1, 2, 4,  6, 8, and 10 hr in the 3 ppb test group. In order to avoid changes
in the gill filaments before chemical fixation, in situ fixation with 0.1  M PBS buffered (pH 7.2) 2.5%
glutaraldehyde was used to minimize shrinkage.  Gill filaments were then dissected out, fixed in the
same glutaraldehyde solution as above for at least 4 hr, and then post-fixed with osmium tetroxide (1%
in the same buffer) for 2 hr. After  acetone dehydration, specimens were embedded in Epon 812 resin
and polymerized in an oven at 60°C.  Semi-thin sections were obtained with a LKB ultratome and
stained with 1% toluidine blue in 1% Borax.  A number of these sections were used for light
photomicrography.  Ultra-thin sections were cut with glass knives, stained with a saturated solution of
uranyl acetate in 50% alcohol, post-stained with lead citrate, and examined under a JEM-100CX
transmission electron microscope.
                                          RESULTS
TQXICITY
       Behavioural symptoms of a toxic reaction were observed among the Grass carp after
immersion in meothrin.  The affected fish in 1 ppb meothrin frequently burst into bouts of rapid
swimming and often attempted to burrow into the bottom of the tank after 6 hr exposure. Some fish
experienced coughing and convulsions.  Affected fish in 3 ppb meothrin suffered coughing and
convulsions after 2 hr of immersion.  These fish subsequently showed symptoms of paralysis and
ataxia and eventually died.  No mortality was recorded for up to 10 hr in the remaining 3 ppb groups.
No abnormal behaviour or mortality was recorded in the control group.

LIGHT MICROSCOPY EXAMINATION

       In the 1 ppb meothrin test group, light microscopy examination of the gills revealed disarray
of the secondary lamellae and a slight curl-over as compared with the control fish gills (Fig. 1) so that
the maximum length from the lamellar margin to the gill filaments was reduced.  The lesions of the
secondary lamellae included hypertrophy of the secondary lamellar epithelium, and collapse of the
pillar cell system (Fig. 2). Some of the secondary lamellae  were fused, and occasionally a group of
secondary lamellae fused so completely that there were no conspicuous interlamellar spaces.
In the 3 ppb meothrin test group, the secondary lamellae were severely degenerated compared with the
lamellae of the 1 ppb test group. After 2 hr exposure to meothrin, approximately 50% of the
secondary lamellae were grossly curled over.  After 4 hr exposure, the secondary lamellar epithelium

                                            85

-------
 was completely separated from the pillar cell system to form a large non-tissue space.  Some
 secondary lamellae were clogged with a large accumulation of blood cells at the distal ends as a result
 of telangiectasia (Fig. 3). Following 8 hr immersion in meothrin, the secondary lamellae showed a
 significant increase in the number of chloride cells and hypertrophy of the epithelial cells. The apical
 epithelial tissue of the secondary lamellae had increased in thickness to more than 1 cell layer.

 TRANSMISSION ELECTRON MICROSCOPY EXAMINATION

        The normal epithelium of secondary lamellae in the Grass carp  gill is simple, consisting of a
 thin, single or double cell sheet separated by  the pillar cell system and blood lacunae (Fig. 4).  The
 epithelia of the primary lamellae have chloride cells, mucous cells and accessory cells interposed
 between the epithelial cells.  After immersion in 1 ppb meothrin, the apical portion of the secondary
 lamellae became enlarged. Lesions of the secondary lamellae were mainly characterized by an
 increased diffusion distance, due to the swelling of the epithelia.  Necrotic cells, pillar cells, myelin
 figures and leucocytes were frequently found in these spaces (Fig. 5). The epithelial cell membrane
 was irregular in shape and protruded in the ambient water (Fig. 6). Degenerated and necrotic
 epithelial cells were often torn from the epithelium and many epithelial  cells had lost their typical flat
 shape  and had become rounded.  The cytoplasm of the epithelial  cells was vacuolated and showed
 accumulation of myelin figures (Fig. 7), organelles and electron-dense deposits (Fig. 8). The pillar
 cell, which is dominated by the nucleus and thin cell flanges, were irregular in shape and microridges
 protruded into the blood vessels associated with the epithelium.  Some sections of the  secondary
 lamellar pillar cell system had severe ultrastructural damage and formed extensive clubbings  with
 blood  cells (Fig. 3).

        The most pronounced change compared to control gill tissue was the  hypertrophy of the
 secondary lamellae, and the increased number of epithelial and chloride  cells  in this part of the gill.
 The central lamellar blood vessels were markedly reduced in size and in some cases were not easily
 identified (Fig. 6). The tips of the secondary lamellae were more heterogeneous than normal,
 containing accumulations of chloride and mucous cells (Fig. 5).  Chloride cells were often
 degenerated, showing loss of cytoplasm and loss of contact with neighbouring epithelial cells, which
 resulted in large intercellular gaps. Chloride cell mitochondria were severely damaged with decreased
 cristae (Fig. 9) and swollen or ruptured nuclear membranes (Fig.  10).

        Gills from the 3 ppb test group were similar to that of the 1 ppb test group. Again, the most
 pronounced change in the secondary lamellae was an increase in the number of chloride and  mucous
 cells, and the vacuolation and cytoplasmic protrusion of the epithelium (Fig. 11).  Surface extensions
 of these rounded epithelial cells were longer and more numerous than extensions of epithelial cells
 from control gills.  Loss of adhesion between the epithelial cells and chloride cells, and rupture of the
 basement membrane, resulted in dispersion and degeneration of cell contents with a resultant decrease
 in staining quality  (Fig.  12).  Some cells were almost completely  disconnected from the lamellae.

                                         DISCUSSION

        The secondary lamellae, where  the major portion of gas exchange takes place,  are a very
 sensitive part of the respiratory system (Karlsson, 1985).   The most overt syndromes of meothrin
poisoning in fish gills, i.e. hypertrophy and telangiectasia, are similar to  the structural changes
observed in the gills of other fish species exposed to other pollutants (Smart, 1976; Temmink et al.,
 1983; Stoker et al., 1985). Telangiectasia was observed in all test fish in this study. This type of
structural damage shows a close similarity to lesions caused by Cd (Karlsson et al., 1985), Zn
(Skidmore & Tovel, 1972), and ammonia (Smart, 1976).  Hypertrophy of the second lamellae also

                                            86

-------
                                                                                 M   ..**"•  S V-^V-
                                                     ~tfft.
                                                                               *6SL * *&*-*, Wi*^. «. fc ^^^
                                                                                                                  2
                                F  "
                                                                                                                   6
Figure 1. Section through the gill filament of control fish, showing the longer secondary lamellae (x360). Figure 2. Section
through gill filament of fish exposed to 1 ppb meothrin for 6 hr.  Hypertrophy of the secondary lamellae can be seen (x250).
Figure 3. Local extensive clubbing with blood (short arrow) and large non-tissue spaces (long arrow) of secondary lamellae
in fish exposed to 3 ppb meothrin for  10 hr (x575).  Figure 4. The tip of a normal secondary lamella from a Grass carp gill.
The double later of epithelial cells and a pillar cell are seen (x4,5SO). Figure 5. The tip of a secondary lamella of a Grass
carp gill after exposure to 1 ppb meothrin for 12 hr.  Note the epithelial layer is more heterogenous than normal, containing
an increased fraction of mucous cells,  chloride cells, and large non-tissue spaces which are invaded by leucocytes. A
necrotic epithelial cell can also be seen (arrow) (x2,4GO).  Figure 6. The centre part of a secondary lamella after exposure to
1 ppb meothrin for 12 hr. The epithelial cells have become rounded and hypertrophied, and irregular erythrocytes can also
be seen (x3,500). Abbreviations on following page.
                                                       87

-------
Abbreviations: SL, secondary lamella; PC, pillar cell; EC, epithelial cell; NT, non-tissue space; MC, mucous cell;
N, nucleus; PL, primary lamella; L, leucocyte; CC, chloride cell; E, erythrocyte; M, mitochondria; MF, myelin figures

Figure 7.  An epithelial cell after exposure to 1 ppb meothrin for 12 hr containing many myelin figures and cell inclusions
         (arrow) which were composed of degenerating mitochondria (x8,450).
Figure 8.  A degenerating epithelial cell after exposure to 1 ppb meothrin for 12 hr containing electron-dense inclusions
         (x7,200).

-------
                  §
      1
                  •rf
                  00
                  at
                    V


                    1

                  I
                  z
2 &
& 1
                  I
                  s
   &

   1
   i-n
   3
                       f

                       1
                        8
                        1
                       i
                      g »
                      i
     iP
     IJ 52
                    I
                  i" If
                  > l-l
                     f» M »
                     K co 2
                     r*sl
                  £l31«]fi
                  8 « 1 s-al.0
                  •a 1511 & *
                   tillI!
                    2  da
89

-------
  occurred in all test fish and is closely comparable to the damage caused by tributyltin compounds
  (Holm et al., 1991) and lindane (Drewett & Abel, 1983). Fusion of the secondary lamellae may be a
  secondary result of gill hypertrophy, and has been reported in the gills of fish exposed to chromate.
  The results suggest that hypertrophy and telangiectasia of secondary lamellae are the primary
  symptoms of exposure to toxic pollutants.  The secondary lamellar epithelium can vary in thickness
  between one or two cell layers, and is usually thinnest adjacent to the blood vessels. The nuclei of the
  epithelial  cells  frequently align with the pillar cell nucleus to minimize the gas exchange distance.
  The epithelial cells are sometimes, separated by an intercellular space, which has been suggested to
  contain lymph  (Hughes & Morgan, 1973).  Such a lymphatic space may have  an important function in
  the regulation of respiratory or osmotic exchange. Since swelling of these spaces and infiltration by
  leucocytes is a common tissue reaction to toxic exposure, the increased non-tissue space could reduce
  the diffusion capacity and result in reduced gas exchange.  The accumulation of leucocytes is probably
  a defense  reaction by the immunological system to the toxin (Temmink  et al.,  1983).

        The plasma membrane of the epithelial cells normally forms microvilli-like structures.  Such
 folding may increase the functional lamellar area and produce microturbulence, enhancing the
 effectiveness of the exchange processes over the surface epithelial (Karlsson et al., 1985).  Reduction
 of these structures and other changes in the epithelial cell surface could  diminish the capacity for gas
 exchange. The loss of normal epithelial cell shape, and the appearance of microridges and cytoplasmic
 protrusions after exposure to meothrin may be due to osmotic perturbation and resultant swelling of
 the secondary lamellae.  The increased number of degenerating and necrotic epithelial cells are
 indicative  of the reduced life span of these cells after exposure to the toxin.  Such gill lesion combined
 with the increased number of myelin bodies in the secondary lamellar epithelial cells could affect
 blood circulation, and further reduce respiratory function.

        Most chloride cells are found in  the stratified epithelium separating the bases of successive
 lamellae, although some chloride cells are also found in the secondary lamellae.  Accumulation of
 chloride cells in the secondary lamellae was detected only following acute, experimental, heavy metal
 contamination and other severe environmental changes. Chloride cells are easily identified partly due
 to their dense population of mitochondria, and the abundance of granular endoplasmic reticulum.  The
 structure and function of chloride cells have/been studied by many authors (Karlsson, 1983; Cioni et
 al.,  1991).  The basal lateral  cell surface  has been shown to be the site of the K-dependent,
 ouabam-sensitive phosphatase component of the Na-K-ATPase enzyme complex (Karnaky et  al.,
 1976), and many reports suggest that chloride cells are  involved in the osmoregulatory mechanism of
 fish gills (Karlsson, 1983; Karnaky, 1986).   The most pronounced effect of meothrin on chloride cells
 was the degeneration of the mitochondria, which has seldom been reported in fish gills exposed to
 other toxins.  Such damage of the mitochondria may be caused by the meothrin itself or by changes in
 osmolarity, and  is likely to severely affect the functioning of the chloride cells. The swelling and
 degeneration of  the chloride cell nuclear membrane may have been the result of changes in osmolarity.
 Lesions of mitochondria can cause severe inhibition of  Na-K-ATPase, enzymes which are vital for
 energetic processes such as oxidative phosphorylation (Cionic et al., 1991).  Chloride cell mitochondria
 are the first organelles to have shown alteration and signs of degeneration implying dysfunction due to
 exposure to toxins. Several models of the mechanistic  action of pollutants such as tributyltin  (Holm et
 al., 1991) and cadmium (Papathanassiou &  King, 1983) have been proposed, and include the inhibition
of mitochondria! oxidative phosphorylation.  The present study confirms the usefulness of morphology
in the identification of gill damage. Possible effects on energy metabolism and on the brain and
nervous system should be investigated through biochemical, histological and physiological studies.
                                            90

-------
                                      REFERENCES

Cionic, C, Merich, D.D., Cataldi, E., and Cataudelia, S., 1991. Fine structure of chloride cells in
        fresh-water and seawater adapted Oreochromis mossambicus (Peters) and Oreochromis
        niloticus (Linnaeus). J. Fish Biol., 39: 197-209.
Coats, J.R., Odonnell-Jeffery, N.L., Mcleese, D.W., 1979.  Toxicity of four synthetic pyrethroid
        insecticides to Rainbow trout. Bull. Environ. Contam. Toxicol., 23: 250-255.
Drewett, N., and Abell, P.D.,  1983. Pathology of lindane poisoning and of hypoxia in the brown
        trout, Salmo trutta L.  J. Fish Biol., 23: 373-384.
"Holm, G., Norrgren, L., and Linden, O., 1991.  Reproductive and histopathological effects of
        long-term experimental exposure to bis(tributylin) oxide (TBTO) on the three-spined
        stickleback, Gasterosteus aculeatus Linnaeus.  J. Fish Biol., 38: 373-386.
Hughes, G.M., 1973. The structure of fish gills in relation to their respiratory function. Biol.
        Rev., 48: 419-75.
Karlsson, L., 1983.  Gill morphology in the zebrafish, Brachydanio rerio (Hamilton Buchanan).  J.
        Fish Biol., 23: 511-524.
Karlsson, L., Runn, R., Haux, C., and Forlin, L., 1985.  Cadmium-induced changes in gill
        morphology of zebrafish, Brachydanio rerio (Hamilton-Buchanan), and rainbow trout,
        Salmo gairdneri Richardson.  J. Fish Biol., 27: 81-95.
Karnaky, K.J., Jr.,  1986.  Structure and function of the chloride cells of Fundulus heteroclitus and
        other teleost. Amer. Zool, 26: 209-224.
Karnaky, K.J., Jr.,  Ernst, S.A., Philpott, C.W., 1976. Teleost chloride cell.  Response of Pupfish  „
        Cyprinodon varieatus  gill Na,K-ATPase and chloride cell fine structure to various high
        salinity environments. J. Cell Biol., 70: 144-156.
Lars, P.D., Rasmassen, E.H., and Ole,  K., 1983.  Light and electron microscopic studies of the
        acute and chronic toxic effects of N-nitroso compounds on the marine mussel, Mytilus
     •   edulis (L.).  Aquatic Toxicol., 3: 285-299.
Mathur, D.S.,  1962. Studies on the histopathological changes induced by DDT in the liver,
        kidney and intestine of certain fishes.  Experientia, 18: 506-509.
 Mclease, D.W., 1980.  Lethality of permethrin, cypermethrin and fenvalerate to salmon, lobster
        and shrimp. Bull. Environ. Contem. Toxicol., 25: 950-955.
 Mikami, N., Baba, Y., Kataji, T., and Miyamoto, Jl, 1985.  Metabolism of the synthetic
        pyrethoroid fenpropathrin in plants. J. Agric. Food Chem., 33:  980-987.
 Papathanassiou, E., and King, P.E., 1983. Ultrastructural studies on the gills ofPalaemon
        serratus (Pennant) in relation to cadmium accumulation. Aquatic Toxic., 3: 273-284.
 Sala, R., Crespo, S., Martin, V., and Castell, O., 1987.  Presence of chloride cells in gill filaments
        and lamellae of the skate Torpedo marmorata.  J. Fish Biol., 30: 357-361.
 Skidmore,  J.F., and Tovell, P.W.A., 1972. Toxic effects of zinc sulphate on the gills of rainbow
        trout.  Water Res., 6:217-230.
 Smart, G.,1976. The  effect of ammonia exposure on gill structure of the rainbow trout (Salmo
        gairdneri).  J. Fish Biol., 8: 471-475.
 Stoker, P.W., Larsen, J.R., Booth, G.M., and Lee, M.L., 1985. Pathology of gill and liver tissues
        from two genera of fishes exposed to two coal-derived materials.  J. Fish Biol., 27: 31-46.
 Temmink,  J.H.M., Bouumeister, P.J., and Defong, P., 1983. An ultrastructural study of
        chromate-induced hyperplasia in the gill of Rainbow trout (Salmo gairdneri). Aquatic
        Toxic. 4: 165-179.
 Waster, P.W., Canton, J.H., and Dormans, J.A.M.A., 1988. Pathological effects in freshwater fish
        Poecttia reticulat (guppy) and Oryzias latipes (medaka)  following methyl bromide and
        sodium bromide exposure. Aquatic Toxic., 12: 323-344.
                                            91

-------
                   CHEMICAL AND BIOLOGICAL CONTROLS
       ON THE BIOACCUMULATION OF HYDROPHOBIC ORGANIC
                            COMPOUNDS IN FOODWEBS

                                    Deborah L. Swackhamer1
                                      INTRODUCTION

        The accumulation of hydrophobic organic compounds (HOCs) in fish is of grave concern due
 to the possible adverse health effects on aquatic organisms as well as potential human health effects in
 fish-eating populations. In the Laurentian Great Lakes of the United States and Canada, the lack of
 reproduction of lake trout has been attributed to contamination by chlorinated organic contaminants
 such as polychlorinated biphenyls (PCBs) and polychlorinated dibenzodioxins and furans (PCDDs and
 PCDFs) (Mac et al., 1991). Deformaties in several colonies of different species of fish-eating birds
 correlate with concentrations of organochlorine contaminants (Giesy et al., 1994).  Also, developmental
 deficiencies have been observed in children born to women who consume large amounts of Great
 Lakes fish (e.g. Jacobson and Jacobson, 1984) raising concerns about trans-generated effects in
 humans due to chronic exposures to chlorinated organic compounds.

    Fish accumulate contaminants from water exposure and also from the food web. In the Great
 Lakes, top predators such as lake trout are thought to accumulate more than 90% of their body burden
 of PCBs from food consumption, making transfer across the gill a minimal exposure route. This
 species has an extensive foodweb consisting of four to five trophic levels. Thus to fully understand the
 process of bioaccumulation of contaminants in fish, one must understand bioaccumulation in each step
 of the food web. This paper focuses on the bioaccumulation process in the primary trophic level  of
 foodwebs.

    The primary trophic level plays an inportant role in determining the fate and transport of HOCs in
 aquatic systems. HOCs preferentially associate with phytoplankton because of their high lipid content
 of 10-40%. The particulate-associated HOCs can then be transferred to the foodweb by grazing, or be
 transported to bottom sediments by settling.

    It follows that the accumulation of HOCs by phytoplankton should be dependent on both the
 physical-chemical properties of the HOC, as well as biological characteristics of the phytoplankton.
 The chemical properties include solubility, octanol-water partition coefficient (Kow; a measure of
 lipophilicity), molecular weight and size, and molecular configuration. The phytoplankton
 characteristics that should be considered are lipid content differences among species, lipid composition,
 surface area, and surface type, and the growth rate of the population. While previous research has
 focused on the chemical controls affecting the association of HOCs with phytoplankton, our premise is
 that biological controls are as, if not more, important in regulating the bioaccumulation process.

    The chemicals of concern in the Great Lakes ecosystem are typical of aquatic systems impacted by
 both long-range atmospheric transport and point source inputs. Organic chemicals that bioaccumulate
in fish include PCBs, DDT and metabolites, dieldrin, chlordane, nonachlor, oxychlordane, PCDDs and
    Environmental and  Occupational Health,  School of Public  Health, University  of  Minnesota
Minneapolis, MN 55455 U.S.A.
                                           92

-------
PCDFs, hexachlorobenzene, toxaphene, and mirex. These chemicals share certain properties, generally
imparted by the presence of chlorine substituents: they are low in aqueous solubility, they are low in
reactivity and thus persistent, they are hydrophobic (log Kow > 4), and they exhibit toxicity.

    Mathematical models have been constructed that predict HOC concentrations in fish from known
water concentrations, by making assumptions about the uptake of chemical at each step of the
foodweb. These so-called "foodchain" models typically begin by assuming that the primary trophic
level is the first step in the pelagic food web, and that the HOC reaches an equilibrium partitioning
between the water and the phytoplankton lipids (e.g. Thomann and Connolly, 1984). Implicit in this
approach is that HOCs have a similar solubility in lipids as they do in octanol. Thus it is assumed that
the lipid-normalized bioaccumulation factor (BAF) is equal to the Kow of the HOC:
and
BAF = [HOC]pWnkta/[HOC]wata

   BAF/fraction lipid = Kow
 Several researchers have demonstrated the validity of this assumption in fish. The BAFs for fathead
 minnows determined in lab experiments for a wide range of chemicals  were correlated to the
 chemicals' Kows resulting in a relationship having a slope near to one  (Veith et al., 1979; Mackay,
 1982). The slopes near unity support the above assumptions to a first order approximation. As a result,
 current phytoplankton bioaccumulation models assume that the process is a thermodynamic
 partitioning to lipids that occurs quickly relative to other processes (i.e. equilibrium is assumed).

    This paper will report on laboratory experimental  results that evaluate the effect of biological
 controls on bioaccumulation and call into question the conventional modeling approach described
 above.

                                         METHODS

    The methods used on our studies have been described in detail elsewhere (Swackhamer and
 Skoglund, 1991; Swackhamer and Skoglund, 1993; Skoglund and Swackhamer, 1993; Stange and
 Swackhamer, 1994), and will be summarized here. Uni-algal cultures of phytoplankton (obtained from
 the Starr Collection, University of Texas at Austin) were incubated in batch experiments with a
 mixture of 40 polychlorinated biphenyl (PCB) congeners. These 40 congeners were chosen as
 representative HOCs having a wide range of known physical-chemical properties. Under specified
 growth conditions, experiments were maintained for 20-30 days and samples taken at specified time
 points. The PCB concentrations were measured in the biomass and media to calculate a BAF for each
 compound at each time point, and the results correlated to controlled changes in the experimental
 variables. Experiments were conducted at two different growth rates, "low"  (0.3-0.9 day'1) and "active"
 (0.13 day'1), and for four genus of algae, Scenedesmus, Selenastrum, Synedra, and Anabena. Cultures
 . were maintained with 16 hours light/8 hours dark, and growth was regulated by temperature (10-12 C
 for low, 20-21 C for active). Bold-Allen media (Nichols and Bold, 1965; Allen,  1968) made in filtered
 Lake Superior water was used in all experiments. Total PCB concentration was approximately 7
 ng/mL.

     Phytoplankton biomass was Soxhlet extracted in  1:1 hexane-acetone, and the media phase batch-
 extracted with hexane. Extracts were cleaned with alumina and silica gel column chromatography, and
 analyzed by gas chromatography with, electron capture detection using a 60 m DB-5 column by the
 internal standard method.
                                             93

-------
      Lipid content and composition were determined using a micro-extraction technique after Gardener
  et al. (1985) as described in Stange and Swackhamer 1994).

                                  RESULTS AND DISCUSSION

         We have shown from these experiments and have discussed elsewhere the dramatic effect of
  growth rate on the uptake and magnitude of PCB bioaccumulation by algae (Swackhamer and
  Skoglund 1993, Skoglund and Swackhamer 1993). We found that under limited or "low" growth
  conditions, the uptake of PCB congeners with log Kow < 6 was related to the congener's Kow
  indicating that the process is a thermodynamic partitioning into lipophilic materials. However  we ob-
  served the process to be much slower than assumed, with equilibrium not fully achieved after 20 days
  of exposure. The  uptake for congeners with log Kow > 6 was not correlated to Kow, indicating that
  the uptake of these very hydrophobic compounds was strongly limited by the kinetics of uptake or
  physically inhibited by the cell membrane in some way. This will be discussed further below.

     The uptake of PCB congeners under "active" growth conditions was not related to  Kow We
  interpreted this as a dilution phenomena caused by the biomass increase during growth  The uptake
  rate is slow relative to growth rate, such that equilibrium is never reached during active growth Thus
  concentrations of PCBs were lower in actively growing cultures compared to dormant cultures and
  could not be predicted by Kow. These experiments are summarized in Figure 1  The implication of
  these experiments is that the conventional approach to modeling phytoplankton bioaccumulation of
  HOCs is invalid for actively growing blooms, leading to large uncertainties in fish bioaccumulation
  models. While the physical-chemical properties of the compound control bioaccumulation when growth
  is minimal, it is the biological control of growth rate that influences bioaccumulation in active
 populations.

    These results also indicated that bioaccumulation was a two-step process, conceptually seen as a
 relatively quick sorption to the surface of the cell followed by a much slower transport across the lipid
 bilayer membrane into the cell. Thus given enough time, equilibrium is reached with cell internal
 lipids; in short time, the compounds are  surface-associated and do not reach equilibrium We have
 modeled the effect of growth by assuming this  two step process and using rate constants for uptake
 and loss from the literature (Connolly and Pedersen,  1988) to demonstrate the effect of growth rate
 and Kow on the magnitude of bioaccumulation (Figure 2). When growth is zero, BAF has a direct and
 linear relationship  to Kow. At higher growth rates, BAF is sharply decreased for the more hydrophobic
 compounds, and is independent of Kow.  For more highly hydrophobic  compounds, the BAF in
 phytoplankton can vary by 3 orders of magnitude for growth rates varying from 0 to 1 day1  Thus the
 effect of growth on BAF and the subsequent effect on contaminant concentrations at higher trophic
 levels cannot be ignored.

    We repeated the "low" growth experiments  across different species to further evaluate the
 biological controls on uptake. We chose a representative green, blue-green, and diatom species to
 compare  differences in  lipid content, lipid composition, and surface area. The results from these
 experiments were similar to those described above. The uptake was slow, and was proportional to
 Kow. However, the rates of uptake  and the magnitude of uptake differed among the different species
 Differences were not related to the surface area  of the different species. Normalization of the BAFs to
 total hpid content slightly decreased the differences among species. To  further elucidate the effect of
 lipids, the BAFs were normalized to one  of three different lipid classes:  glycolipids (internal lipids)
 neutral lipids (storage lipids), and phospholipids (membrane structural lipids). The results are shown'in
Figure 3  Interestingly, differences among species for congeners with log Kow < 6.5 were minimized
by normalization to glycolipids, whereas differences among species for congeners with log Kow > 6.5

                                            94

-------
                         •s  f
                         ^  £
                         s£
                         O  O
C3
Q

c
o
-*— >
03
•»*
3
s
3
 03
 O
• »H
CQ
 o
H-J
 03
 U
 O
^5
 03
                                                           o  • .2
                                                             -a co
                                                           «
                                                               
-------
            di/
               •'
CO
0
                                           0
      is
                                                                 fi
                                                                 I
                                                                 a
                                                                 •a
                                                                S
                                                                •o
                                                                ;a
                                                                •a
                                                                CQ
                                                                f* T^*
                                                                ~ O\

                                                                'o S


                                                                2* fe


                                                                o j~
     K  &
    '13OT
     •S'S
     a  w
     o -o
     
-------
                      Lipid Class Normalized BAFs at 12 Days
                      7.0-
                2

                £•    6.0-

                "o
                o
                O
                     5.0-
                      4.0
                                ~T~

                                 5
          LL.

          <
D)
O
                      7.0-
      2

      5-   6.0 H
                 a>
                 •z.
           5.0-
                      4.C
                                       an

                                      1 D
•
 *
                                                                       n  Anabena

                                                                       *  Selenastrum
                     7.0-
                CO
               2

               £-    e.o H

                o
               .c
                Q.
                CO

                ?    5.0-
                     4.0
                            an

                           a. o
                        a
                       , a
                             "0
                                        log Kow
Figure 3.  Observed BAFs for 40 PCB congeners in laboratory experiments where growth rate was

minimized for two species of algae vs. log Kow of the congener. The BAFs are normalized to three

different lipid classes.    .
                                        97

-------
  were minimized by normalization to phospholipids. This provides support for our earlier interpretation
  that uptake for the medium hydrophobia compounds is a thermodynamic partitioning to internal
  cellular lipids (albiet a slow process). Thus species differences in BAF are minimized by normalizing
  to that pool of lipids. However, the very hydrophobic compounds appear to associate most strongly
  with the phospholipid membrane, supporting the concept that uptake of these compounds is physically
  restricted  rather than limited by the kinetics of transport.

                                         CONCLUSIONS

         The results presented here underscore the importance of physical-chemical properties of
  contaminants as well as the biological properties of phytoplankton in controlling the bioaccumulation
  of contaminants in the primary trophic level. The hydrophobicity of the chemical, expressed as Kow,
  plays an inportant role in determining the magnitude of accumulation,  and is linearly related to the
  BAF at equilibrium for compounds with 4 < log Kow < 6. These compounds appear to associate with
  the glycolipid fraction of total lipid, representative of the internal lipids of the phytoplankton.  In
  contrast, the more highly hydrophobic compounds with log Kow > 6 seem to accumulate at the lipid
  bilayer and do not move into the cell, as indicated by their association with phospholipids. However,
  when phytoplankton populations are actively growing the increase in biomass occurs more quickly
  than the uptake of contaminants, thus preventing equilibrium from being achieved. Thus modeling
  frameworks describing the accumulaion of HOCs by phytoplankton must incorporate both chemical
  and biological processes, and include Kow, growth rate, and lipid content and class as variables.

                                         REFERENCES

 Allen, M.M. 1968. Simple conditions for growth of unicellular blue-green algae on plates. J. Phycol
         4:1-4.
 Connolly, J.P. and Pedersen, C.J. 1988. A thermodynamics-based evaluation of organic chemical
         accumulation in aquatic organisms. Environ. Set. Technol. 22:99-103.
 Gardener, W.S., W.A. Frez, E.A.  Chichock, and C.C. Parrish.  1985. Micromethod for lipids in aquatic
         invertebrates. Umnol  Oceanogr. 30:1099-1105.
 Giesy, J.P., J.P. Ludwig and D.E. Tillett. 1994. Deformaties in birds of the Great Lakes region
        Environ.  Set. Technol. 28:128A-136A.
 Mac, MJ. and C.C. Edsall.  1991. Environmental contaminants and reproductive success of lake trout
        in the Great Lakes: an epidemiological approach. J. ToxicoL Environ. Health. 33:375-394.
 Mackay, D. 1982. Correlation of bioconcentration factors. Environ. Sci. Technol. 16:274-278.
 Nichols, H.W. and H.C. Bold. 1965.  Trichorosarcina polymorpha Gen. et Sp. Nov. J. Phycol 1:34-38.
 Skoglund, R.S.and D.L. Swackhamer. 1994. Processes affecting the uptake and fate of hydrophobic
        organic contaminants by phytoplankton. In: L.A. Baker, ed, Environmental Chemistry of Lakes
        and Reservoirs. ACS Advances in Chemistry Series, Lewis Publishers, Chelsea, MI.
 Stange, K. and D.L. Swackhamer, 1994.  Factors affecting phytoplankton species-specific differences in
        accumulation of 40 PCBs. Environ.  ToxicoL Chem. 13(11): in press.
 Swackhamer, D.L. and R.S.  Skoglund. 1991. The role of phytoplankton in the partitioning of
        hydrophobic organic contaminants in water. In: R. Baker, ed., Organic Substances and
        Sediments in Water. Lewis Publishers, Chdsfta, MT pp 01405
 Swackhamer, D.L. and R.S.  Skoglund. 1993. Bioaccumulation of PCBs by algae: kinetics vs.
        equilibrium. Environ. Toxicol. Chem. 12:831-838.
Thomann, R.V. and J.P. Connolly. 1984.  Model of PCB in Lake Michigan lake  trout food chain
       Environ. Sci. Technol. 18:65-71.
Veith, G.D., D.L.  DeFoe, and B.V. Bergstedt. 1979. Measuring and estimating the bioconcentration
       factor of chemicals in fish. J.  Fish. Res. Bd. Can. 36:1040-1048.

                                          98

-------
    ECOTOXICOLOGICAL MONITORING OF MAJOR INDUSTRIAL
                        EFFLUENTS IN NANJING, CHINA

                  Hongjun Jin1, Xiao Lou2, Zhehai Zhang2 and Guoxiang Wang2

                                        ABSTRACT

       Physico-chemical analysis is ait present the only method employed to supervise and manage the
levels of industrial discharges in China.  It is important to demonstrate that biological assays should be
incorporated in the studies that are used to determine industrial effluent discharge limits.  By using
toxicity testing on aquatic organisms, this study evaluated the effects of effluents discharged from 24
factories, to determine which effluents were toxic and required control and regulation, and examines
the efficiency of effluent treatments. Twenty-four effluents form twelve types of industries were
tested. The biological toxicity data from these effluents did not entirely agree with the
physico-chemical data gathered during the monitoring of these effluents.

       Acute toxicity tests were conducted with two species of fish and one species of daphnids.
Fourteen treated effluents, based on their chemical compostion, contained toxic chemicals in
concentration which exceeded National discharge standards.  Some of these effluents were toxic to the
aquatic life. Ten treated effluents met National discharge standards, but were still highly toxic.  The
concentration-based acute toxicity measurements of the effluents were translated into acute toxicity
units (TUa), and the annual discharge (AD) data of the effluents were tabulated. The Annual
Discharge-Acute Toxicity Unit (AD-TUa) is recommended to assess the environmental impact of
effluents, and to establish effluent priority with regard  to toxicity.

       Three high priority effluents were then studied in a 21 day chronic toxicity test using
daphnids.  Reproductions rate was the most sensitive indicator to these effluents, and the no observable
effect concentation (NOEC) and lowest observable effect concentration (LOEC) of each effluent to
Daphnia magna were derived.  The range of acute and chronic ratios for these effluents was found to
be from 10 to  100. Acute toxicity tests  were performed on fish and daphnids using three selected
effluents both before and after treatment. It was determined that the effectiveness of effluent treatment
could not be monitored adequately by physico-chemical testing alone. The toxicity of one chemical
effluent was tested downstream from the effluent outlet at three stations, and the water downstream
was found to be lethal to D.  magna with a short time.

       The supervision of toxic effluent discharge must be strengthened by means of biological
toxicity assays on the effluents.  Effluent toxicity limits should be established and incorporated into
discharge permits to prohibit the discharge of pollutants in toxic amounts.

                                      INTRODUCTION

       The impact of industrial effluents on the aquatic ecosystem is increasing with the development
of modern industry in China. This impact has drawn the attention of both the public and
environmental agencies. Pollution control is of urgent concern world-wide, and in the West, biological
    Department of Environmental Sciences, Nanjing University, Nanjing 210008, P.R. China

    2Nanjinhg Municipal Environmental Monitoring Center, Nanjing 210008, P.R. China

                                            99

-------
 assays have become an essential test in the monitoring of toxic effluents (Thomas, 1988; Zwart, 1988;
 USEPA, 1991).  Physico-chemical analysis is at present the only method used in the supervision and
 management of industrial discharge in China, and it appears to be less reliable in establishing effective
 guidelines especially when the effluent is a complex mixture of toxic pollutants or a mixture with
 unknown constituents.  As control over conventional pollutants is strengthened in China, emphasis
 should be placed on the reduction of toxic chemical discharge. In order to control toxic discharge via
 water quality criteria, it is important to understand that the results of biological assays must be
 incorporated into the criteria to be used to set industrial effluent discharge limits.  Nanjing is one of
 the major petrochemical industrial bases in China,  and there are many factories along the Yangtze
 River. By using toxicity tests on aquatic organisms, this study aims to evaluate the ecological effects
 of effluents discharged from 24 factories in order to determine which are the most toxic, and to
 examine the efficiency of effluent treatment.  Further studies and the application of ecotoxicological
 monitoring are recommended to control toxic discharge levels in  China.
                                MATERIALS AND METHODS
TEST EFFLUENTS
        Effluents were sampled directly from 24 factories in Nanjing representing 12 different kinds of
industries (chemical, metallurgy, gas-making, pharmaceutical, chemical fiber, textile, hide-processing,
electronic, instrument and meter-making, engine manufacturing, paper, and food plants).  Both
composite and grab samples were used. The chemical parameters of the effluents were measured at
the Nanjing Municipal Environmental Monitoring Center according to the  methods recommended by
the NEPA of China (China NEPA,  1989). The ambient toxicity of one of the chemical effluents was
tested at 3 different sites downstream from the outlet of the discharge station. Effluent samples were
diluted with tap water aerated for 3 days prior to use. The dilution water had a pH of 7.2-8.5,
dissolved oxygen of 6.9-8.1 mg/L, COD of 1.08 mg/L, conductivity of 294 us/cm, and a hardness of
1.082 me/L.

TEST ORGANISMS

        Juvenile  Crucian carp (Carassius auratus) and Silver carp (Hypophthalmichtys molitrix) were
obtained from the Freshwater Fisheries Institute of Jiangsu province, Nanjing. The fish were kept in
holding tanks at  a water temperature of 22±2 °C and a 12 h light: 12 h dark photoperiod.  The fish
were last fed 24  h before experimentation. Daphnids (Daphnia magna) were obtained from the
Laboratory of the Department of Environmental Sciences at Nanjing University.

TEST PROCEDURES

        The testing procedures followed the methods recommended by the relevant agencies
(APHA-AWWA-WPCF, 1985; China NEPA, 1986).

1. Acute Toxicity Tests

        Renewal tests with fish were conducted in 40 L glass aquaria, while  artificial stream tests were
conducted in artificial laboratory streams measuring 31 L in volume, 11 cm in depth and with a flow
rate of 100 ml/min.  Each vessel contained 20 juvenile fish. Daphnia  were kept in 250 ml glass
beakers, with 10-20 daphnia per beaker. At least 5 test concentrations and a dilution water control
were used in each experiment.  The water  temperature and photoperiod were  maintained as in the
                                           100

-------
holding tanks. Observations were made after 24, 48 and 96 h of exposure. The end-point of testing
was 96 h LC50 for both the fish and daphnids.

2. Chronic Toxicity Tests

       Chronic tests with daphnids were conducted in glass beakers fitted with a side opening covered
with #25 silk screen for overflow of test solution. The test liquid was delivered by a peristaltic pump
and maintained at a volume of 120 ml per beaker.  Each test concentration and control were run in
triplicate.  Five daphnids (12 h age) were placed in each beaker.  The tests were run for 21 days at
22±1  °C and the test animals fed with concentrated green algae in medium.  Mortality, reproduction
rate and growth rate were recorded every day.

STATISTICAL ANALYSIS

       The LC50 and 95% confidence intervals for the acute toxicity data were determined by the
Trimmed Spearman-Karber method (Hamilton et al., 1978). One-way analysis of variance was
performed on the chronic toxicity data.

                                          RESULTS

       The annual amount of discharge :and the chemical characteristics of each effluent were
determined from the Nanjing Municipal Environmental Monitoring Center. It could therefore be
concluded whether or not a treated effluent corresponded to the requirements of the national discharge
limits. Table 1 shows the types,  the over-discharge situation,  acute toxicity and AD-TUa data of the
24 effluents. For the fish tests, the lower LCSOs were selected and presented in the table.

       The degree or grade of chemical toxicity to  aquatic life may be classified in a variety of ways,
but one of the most useful of these was proposed by an international group on the basis  of acute
toxicity threshold values of chemicals  (Cairns and Dickson, 1980).  There is apparently no publication
on the classification of industrial effluent toxicity, but it follows from Table 1 that some effluents  (see
chemical and metallurgical  effluents) even after treatment may be considered to be toxic or very toxic
to aquatic life.  Table 1 also indicates  that most of the effluents ater treatment still exceeded chemical
discharge limits, especially  the chemical effluents C-l, C-5 and C-7 which were also very toxic
according to the biological assay.  Although some effluents with no over-discharge met the discharge
standards, some (such as chemical effluent C-6) were  still highly toxic.  The  concentration-based
toxicity measurements can be translated into toxic units (TUs), i.e. the acute toxic unit (TUa) and  the
chronic toxic unit (TUc). TUs are a simple way to quantify acceptable  in-stream toxicity levels
(Thomas, 1988). The impact of any toxic effluent is of course dependent on the dilution factor and
other conditions.  The Annual Discharge-Acute Toxicity Unit (AD-TUa) could be recommended to
assess the impacts of effluents and to screen the toxic priority of effluents. The lowest LC50 for both
the fish and daphnid species should be used to calculate such values. By using the AD-TUa data
listed in Table. 1, the priority rating of effluents in Nanjing can be established.  The chemical effluents
C-l, C-2, C-3 and C-5,  and the metallurgical effluent  M-l would probably cause significant impacts
on the aquatic ecosystem. Some of the effluents such as F-2, EM, TE-1, TE-2, and TE-3 would have
a lesser impact on the environment.
                                            101

-------
Table 1.  Annual discharge and acute toxicity data from effluents from 24 factories.
Industry Type
CHEMICAL
METALLURGY

GAS
PHARMACY

CHEMICAL
FIBRE
Effluent
C-l
C-2
C-3
C-4
C-5
C-6
C-7
H-l
H-2
G
P-l
P-2
CF
Annual
Discharge
(lOT/yr)
2655.0
1571.54
1400.20
1095.64
34.02
18.42
0.36
72.26
607.0
248.4
600.0
106.3
1,191.25
Over-
discharge
(Yes/No)
Yes
Yes
Yes
No
Yes
No
Yes
Yes
No
Yes
Yes
No
Yes
96h LC50*
for Fish
1.02
(0.95-1.10)
c. carp, FT
13.33
(11.63-15.28)
c. carp, R
26.99
(25.4-28.64)
c. carp, R
19.98
(18.40-21.69)
c. carp, R
0.18
(0.16-0.20)
s. carp, R
0.74
(0.67-0.81)
s. carp, R
NAT
s. carp, FT
1.71
(1.51-1.93)
c. carp, R
10% death
caused by
100% effluent
11.41
(10.38-12.53)
s. carp, R
39.73
(34.90-45.23)
c. carp, R
—
NAT
c. carp, R
48h LC50+ for
Daphnids
0.31
(0.25-0.40)
2.22
(1.60-3.06)
7.50
(6.48-8.76)
21.32
(17.95-25.32)
0.36
39.43
(33.44-46.60)
0.11
(0.09-0.15)
0.38
(0.24-0.59)
34.86
(21.47-56.95)
11.32
(8.56-14.97)
51.50
(45.93-57.84)
30.00
(22.56-39.9)
43.37
(39.45-47.7)
TUa
322.6
45.0
13.3
5.0
555.6
135.1
263.2
263.2
2.9
8.8
2.5
3.3
2.3
AD-TUa
(T.TUa*104
856,503
70,719
18,671
5,484
18,900
2,489
327
19,016
1,741
2,194
1,510
354
2,747
                                                                       Continued..
                                 102

-------
Table 1 (continued)
Industry Type
TEXTILE


TANNING
ELECTRONIC
INSTRUMENT
AND METER
ENGINE
MANUFAC-
TURING
PAPER
MAKING
FOOD
Effluent
TE-1
TE-2
TE-3
TA
E-l
E-2
IN
EN
PM
F:l
F-2
Annual
Discharge "
(10
-------
            Table 2.  Toxic effect of three effluents on the reproductive rate of D. magna.

NOEC*
LOEC*
C-l
0.0125
0.025
C-5
0.0056
0.032
M-l
0.032
0.056
 * % effluent V/V
          Table 3.  Toxicity of chemical effluent C-l downstream from the discharge outlet.
Distance from
the Outlet (M)
Lethal Toxicity
(Lethal Time)
0

LT100
6 min
200

LT100
25 min
400

LT100
8h
1500

LT40
24 h
   Table 4. Main chemical parameters and toxicity of selected effluents before and after treatment.

UNTREATED
TREATED
Chemical
Parameters
(mg/1)
48 h LC50"
for O. magna
96 LC50 for
Silver Carp
Chemical
Parameters
(mg/1)
48h LC50
for O. magna
96h LC50 for
Silver Carp
Chemical Effluent
COD 479
Nitro-Compounds 4.8
Amino-Compounds 13.2
Phenol 2.8
19.84
(16.29-24.17)*
6.13
(5.76-6.54)
COD 25.4
Nitro-Compounds in trace
Amino-Compounds in trace
Phenol in trace
NAT
NAT
Textile Effluent
COD 172
Phenol 0.011
13.43
(7.05-25.54)
44.05
(39.32-49.32)
COD 52.1
Phenol 0.009
48.0
(37.64-61.21)
46.75
(43.93-49.76)
Electronic Effluent
Cr61" 26.0
0.84
(0.68-1.04)
22.39
(20.41-24.55)
Cr6" 0.12
2.40
(1.95-2.96)
14.09
(12.68-15.65)
* Effluent V/V;
* 95% Confidence Interval
NAT = No Acute Toxicity
                                          104

-------
        The results of the acute toxicity tests demonstrated that Daphnia magna was usually the most
 sensitive species to the effluents, which agrees with previous reports (Thomas, 1988). Three of the
 high-priority effluents, chemical effluents C-l and G-5 and metallurgical effluent M-l, were selected in
 the chronic toxicity tests with daphnids. The test concentrations of each effluent were set on the basis
 of the acute toxicity data.  The end-points of chronic toxicity involved survival,  and growth and
 reproduction rates. The results showed that the reproduction rate was the most sensitive indicator of
 toxicity. The NOEC and LOEC of each effluent to D. magna is listed in Table  2.  On the basis of
 acute and chronic toxicity data, the range of acute and chronic ratios (ACR) of the selected effluents
 was found to be from 10 to 100.  This parameter  could also be used to evaluate the acceptable
 in-stream concentrations of other effluents based on acute toxicity data alone.

        The ambient toxicity of chemical effluent C-l was tested at 3 sites downstream from the
 discharge outlet. The results showed that the water downstream from the outlet was lethal to Daphnia
 magna within a short period of time (Table 3).

        In order to examine the efficiency of effluent treatments,  acute toxicity tests using 3 selected
 effluents before and after treatments were conducted with 2 species of aquatic animals. The effluents
 chosen were from chemical, textile and electronic industries, and tests were conducted with D.  magna
 and Silver carp.  According to physico-chemical parameters, biological treatment removed 99% of
 pollutants in the chemical effluent, and there was  no mortality in the acute toxicity tests (Table 4).
 The efficiency of electrolytic treatment on the textile effluent was very good, but the toxicity of the
 effluent to fish was not obviously reduced.  Removal of chromium by ion-exchange from the
 electronic effluent reached 99.2%,  but  while the toxicity to daphnids was decreased, the toxicity to fish
 seemed to be increased (Table 4).

                                         DISCUSSION

        Management of industrial discharges in China today relies on physico-chemical monitoring
 since it reveals the primary constituents of the effluents. However, the results of this study
 demonstrate that such tests do not  detect all chemicals in industrial effluents, nor do they measure
 possible interactions of chemicals or their impact  on the environment.  Ecotoxicological monitoring of
 whole effluents through toxicity testing on aquatic organisms (biological assay) can reflect the  overall
 impact of industrial discharges on  an aquatic system although it does not identify the actual toxic
 pollutants in waste water.  In the present study, the acute toxicity tests were employed to establish a
 toxic order among the effluents tested,  but the impact of an effluent depends not only on the effluent
 toxicity but also on the discharge rate.  Therefore, the AD-TUa of effluents was recommended as  a
 measure for discharge management, and, based on this unit, a toxic priority of industrial effluents  in
 Nanjing was established (see Table 1).

        According to the LC50 and NOEL  values of selected effluents (Tables 1 and 2), the
 Application Factor (AF) for protection  of aquatic life from toxic effluents ranged from 0.01 to  0.10.
 However, the possible interactions between the chemistry of the real receiving water (the Yangtze
; River)  and the effluents were not measured in this study, so the range of safe concentrations for other
 effluents for which chronic testing has  not been done can only  be evaluated by using a proper AF
 value.

        Most of the factories involved  in this study are located along the Yangtze River into which the
 effluents are directly or indirectly discharged. A river can have a large flow and therefore a
 tremendous dilution capacity, but the prevalence of riverside industrial discharge does damage the
                                             105

-------
 ecosystem at the edge of the river. Large-scale fish mortality caused by discharge of industrial
 effluents into small water bodies or the polluted zone of large water bodies does occasionally occur in
 China.  These constitute a warning that must be taken seriously. Despite dilution, the results of this
 study demonstrated that the water downstream from the outlet of chemical effluent C-3 was still lethal
 (Table 3).

        The results of this study showed that the physico-chemical data did not entirely tally with the
 results of the bioassays (Table 1).  The treatment of the effluents was insufficient according to the
 National Discharge Standards, and some of the treated effluents were still very toxic to aquatic life.
 Some of the major constituents of the effluents might include aromatic compounds, benzene and its
 homologues and derivatives, as well as some unknown chemicals.  Ten of the fourteen treated
 effluents were considered to have met the requirements of the National Standards, but one of them,
 chemical effluent C-6, was still highly toxic based on the bioassay results.  Obviously, the results of
 bioassays can not be predicted or substituted for by physico-chemical monitoring.  The effluents may
 contain some toxic substances not detectable at present.  Enforceable discharge limits in China exist
 for only a few of the thousands of chemicals in use, and the routine monitoring of effluents in China
 includes only a few chemical parameters, while other toxins are ignored.  There is undoubtedly
 additivity and/or synergism of toxic effects in multi-chemical effluents. The use of more  chemicals in
 treatment procedures may enhance interactions and actually increase the toxicity of effluents to aquatic
 life. Clearly, the efficiency of effluent treatments can not be effectively examined only by means of
 physico-chemical monitoring, and further studies on the identity of the  toxic substances in industrial
 effluents need to be done to  control toxic discharges.

        In an attempt to control industrial discharge, the China NEPA has developed a Waste Load
 Allocation and created a discharge permit system.  The parameters used in development of permits
 usually include BOD/COD, S2-, oil and other general chemical indicators, but no biological toxicity
 limits.   The system of fines for exceeding discharge  limits is based on Waste Load Allocation and
 enforced by environmental agencies, but appears unable to protect aquatic life against chronic damage.
 The whole-effluent toxicity limitation concept involves using  acute and chronic toxicity tests to
 measure the biological toxicity of waste waters. Toxicity is a useful parameter to protect  water quality
 and resources, and ecotoxicological monitoring of whole effluents is cost effective.  In the United
 States, the toxicity of effluents has been measured since the late  1970s  and discharge permits have
 ncluded toxicity limits since  the early 1980s.  The increased use  of toxicity testing has identified
 substantial numbers of unacceptably toxic effluents (USEPA,  1991). In China, the management of
 toxic effluents must be improved and strengthened.  In order to reduce  toxic chemical release,
 especially from petrochemical factories, biological  assaying must be developed, and effluent toxicity
 limits should be established and incorporated into discharge permits. A national policy for
 development of water quality-based permit limitations for toxic pollutants should be established,  and
 the policy should recommend the use of whole effluent toxicity testing  to achieve a standard that
prohibits the discharge of waste materials in toxic amounts.

                                        REFERENCES

APHA-AWWA-WPCF, 1985.  Toxicity test methods for aquatic organisms. IN: Standard Methods for
        the Examination of Water and Waste-water (16th ed.). American Public Health Association,
        Washington, D.C. pp. 689-826.
Cairns, J. and K.L. Dickson, 1980. The ABCs of biological monitoring. In:  Biological Monitoring of
        Fish. C.H. Hocutt and J.R. Stauffer, Jr., eds.  Lexington Books, Lexington, Massachusetts, pp.
        1-31.
                                            106

-------
China NEPA, 1986.  Standard Techniques of Environmental Monitoring, Sect. 4, Biomonitoring
       (Aquatic Environment).  Beijing, China,  pp. 55-64.
China NEPA, 1989.  Methods for the Monitoring and Analysis of Water and Wastewater. China
       Environmental Sciences Press, Beijing, China.
Hamilton, M.A., R.C. Russo and R.V. Thurston,  1978. Trimmed spearman-Karber method for
       estimating median lethal concentrations in toxicity bioassays. Environ. Sci. Technol. Vol. 11.
       pp. 714-719.
Thomas, N.A., 1988.  Use of biomonitoring to control toxics in Hie United States. Wat. Sci. Tech. 10:
       101-108.   ,
USEPA, 1991. Technical support document for water quality based toxics control. EPA, Office of
       Water, Washington, D.C.
USEPA, 1991. Toxicity identification evaluation: characterization of chronically toxic effluents, phase
       I. EPA, Office of Research and Development. EPA/600/6-91/005, Washington, D.C.
de Zwart, D., 1988.  Biomonitoring of effluent toxicity. In:  Manual on Aquatic Ecotoxicology.
       H.A.M. de Kruijf, D. de Zwart, P.K. Ray and P.N. Viswanathan, eds.  Kluwer Academic,
       Dordrecht, pp. 208-213.
                                          107

-------
      INVESTIGATION INTO THE CAUSE OF THE DEATH OF FISH
                     OCCURRING IN LAKE XUANWU (IV).

     Guoxiang Wang,* Dong Fang1, Gaoquan Shi2, Xiao Lou1, Zhonglin Xia1 and Hongjun Jin2

                                        ABSTRACT

       The cause of recurring incidents of large-scale fish mortality in Lake Xuanwu was
investigated. The pathogenic bacteria, Aeromonas hydrophila, was isolated from silver and variegated
carp recovered from the lake, and is believed to be the responsible pathogen. Population changes of
A. hydrophila in relation to phytoplankton succession and changes in redox shielding level in Lake
Xuanwu are also discussed.

                                     INTRODUCTION

       In recent years, large-scale fish mortality in Lake Xuanwu has been observed to be a recurring
phenomenon (Xu 1991).  Despite numerous previous investigations (Pal and Tripathi 1978, Yu and
Mao 1991, Chen 1992), controversy over the exact cause of death still exists.  Since 1988 there have
been three additional outbreaks of fish mortality, which prompted this study.
The water quality of Lake Xuanwu was tested during one such incident of fish  mortality.  Levels of
un-ionized ammonia (0.12 mg/1), oil (0.41 mg/1), and BOD (14.4 mg/1)  were found to slightly exceed
the Chinese National Standard water quality for fisheries (GB11607-89). Other pollutants, however,
were almost never above the National Standard, and only trace amounts of pesticide were found.

       Despite the measurement of water containment levels during a fish epidemic that exceeded
those of the National Standard, experiments found that these amounts, and their additive effects, were
far below lethal levels. Furthermore, tests performed on water taken from the river feeding into the
lake found no indication of the water being toxic to fish. Therefore, it is reasonable to conclude that
the water quality itself is not directly responsible.  This paper reports our experimental results that
identify Aeromonas hydrophila to be the pathogen involved in the large-scale mortality of fish in Lake
Xuanwu.

                              MATERIALS AND METHODS

SYMPTOMS OF FISH INFECTION

       In August of 1992, an infectious epidemic  was responsible for the death and injury of a large
number of fish in Lake Xuanwu. Numerous  red dots were found on the skin of fish, particularly in
the abdominal area. The mouth, eyes, fins and anus also typically turned red.

ISOLATION AND IDENTIFICATION OF THE PATHOGENIC BACTERIA

       Silver carp and Variegated carp were chosen  for isolation of the bacteria.  Fish used for the
study were all visibly infected and included both dead and living specimens.
   'Nanjing Environmental Monitoring Center, Nanjing 210013, P.R. China
   2Nanjing University, Nanjing 21008, P.R. China

                                         108

-------
Samples of blood and dis-coloured skin were first placed into a salt water solution.  The mixture was
then ground and stirred under sterile conditions. A 1 ml sample of the blood-skin mixture was
extracted and used for a plate culture.  Owing to the identification of Aeromonas hydrophila as a
pathogen common to silver and variegated carp, only those micro-organisms having the following
characteristics were used for subsequent isolation and identification:
       a. Sphere shaped, slightly medially convex
       b. Smooth skin
       c. opaque white
       d. Diameter 1-1.3 mm  (after culture for 24 h)

STANDARD BACTERIA STRAIN USED FOR COMPARISON TESTS

       A standard strain of A. hydrophila was purchased from the Shanghai sanitation and
antiepidemic station.

THE FISH USED FOR GERM VIRULENCY AND COMPARISON TESTS

       The fish used for the tests described below were purchased from a local supplier at the market
(mean length 35.0±2.0 cm; mean weight 0.7+0.06 kg).  The fish appeared healthy and showed no
signs of being infected with any form of disease.

GERM VIRULENCY TESTS

       Once Aeromonas hydrophila was  isolated and identified, it was cultivated in meat soup and
protein peptone.  Standard germs and isolated germs were cultivated in large scale.  The virulency of
the bacteria was tested using both body cavity injection and immersion methods.

       i.  Body cavity injection.

       The body cavity of healthy fish were injected with 10 ml of isolated A. hydrophila solution or
standard germ solution at a concentration of 1.8 x 10/8 germs/ml.  The fish were returned to a tank for
observation (water temperature 27-31 °C, dissolved oxygen content (DO) = 5.0-7.0 mg/1).

       ii.  Immersion method.

       The b odies of healthy fish were scratched at 10 different points with a small needle prior to
being introduced into a tank containing A. hydrophila at a concentration of 107 germs/ml.  The water
in the tank was maintained at 30±2 °C with a DO of 6.6±1.0 mg/1.

                                         RESULTS

THE ISOLATION AND IDENTIFICATION OF THE PATHOGENIC BACTERIA

       i.  Morphological features of the isolated pathogenic bacteria.

       Following 20 hours of cultivation, the isolated bacteria was observed to be composed of short,
rod-shaped cells found singly or in pairs.  The cells were 0.6-0.8 um in diameter and 1.1-2.2 urn in
length.
                                            109

-------
        ii. The physiological and biochemical properties of the isolated bacteria.

        The bacteria isolated from the silver and variegated carp were successfully grown in the
 semi-solid cultivation medium.  Comparison of them to data outlined by Buchonan and Gibbons
 (1984) in "Bergeys Manual of Determinative Bacteriology", the isolated bacteria was identified as A.
 hydrophila.  The bacteria have a positive Voges-Prokauer test, and have the ability to oxidize glycol
 and adversely effect the health of fish.

 VIRULENCY ASSESSMENT OF THE BACTERIA

        A virulency test was performed on healthy carp through comparative exposure to the isolated
 strain and the standard A. hydrophila strain.  Following infection with the bacteria, the following
 symptoms appeared chronologically: red spots appeared around the mouth, then on the anus; discharge
 of blood from the anus was observed; red spots appeared on the skin,  eyes, and operculum.  Upon
 necropsy, blood was found within the intestine.

        The symptoms of infection were found to be greater among fish that had been directly injected
 with the bacteria over those that were exposed to it through immersion.  The virulency of bacteria was
 found to be higher in bacteria isolated from blood samples than from skin samples.

                                         DISCUSSION

        1) The bacteria A. hydrophila was isolated from living and dead infected fish taken from Lake
 Xuanwu.  The results indicate that this bacteria is the pathogen responsible for the periodic large-scale
 fish mortality observed in this lake.  This bacteria acts  to destroy the blood circulation system of fish.

        2) The isolated bacteria were found to readily  infect healthy, Variegated and Silver carp with
 strongly adverse effects.  Specimens injected with the bacteria expired within 48 hours even when
 maintained in water with a high DO (6.5 mg/1).  Although those fish that were immersed in water
 containing A hydrophila did not die after 48 hours, they clearly showed symptoms of disease.  While
 the test fish immersed in water containing A hydrophila and a DO of  6.6±1.0 mg/1 remained alive
 after 96 hours,  fish exposed to the same conditions except for the DO  (which was  controlled at 2.5
 mg/1), expired within 3 hours.  These fish showed symptoms indicative of bacteria infection and not
 low dissolved oxygen.   Healthy fish were found to survive in water with a low DO  (2.5 mg/1) for at
 least 6 hours.

        3) Jiang (1992) indicated that environmental conditions were the underlying cause of the
 mortality of the Lake Xuanwu fish.  Poor water quality, harmful organic chemicals and low dissolved
 oxygen levels have combined with increased temperatures in summer to allow for accelerated growth
 of bacteria. Aeromonas hydrophila acts to destroy the  circulatory system of fish by attacking red
 blood cells.  Ni (1979) reported that this bacteria is conditionally pathogenic in that it is only lethal to
 fish that are stressed by poor water conditions.  Injection of the bacteria directly into the body cavity
 of fish resulted in rapid growth of the bacteria because  of favorable conditions, and led to the death of
 the  fish shortly thereafter. The infection of fish exposed through immersion in bacteria-laden water
was reduced owing to the high quality of the water, which slowed growth of the bacteria.  In this
circumstance, the immune system of the fish was able to successfully combat the bacteria taken in by
the  fish when feeding.  When the immune system of the fish is stressed through poor water quality
 (such as low dissolved oxygen), the rate of infection occurs much more rapidly. During the summer,
the  water  quality is  compromised by organic and other  pollutants and elevated water  temperatures

                                            110

-------
(27-32 °C).  Additionally, there are numerous other germ carriers found in Lake Xuanwu (including
Myxobolus sp., Sinergasilus polywplus, Lernaea polymorpha, and Dactylogyrus sp.).  Finally, the
dissolved oxygen concentration of the lake during summer nights tends to be very low (the mean DO
between 2-7 a.m. is 2.53 mg/1).  These factors combine to facilitate growth of the bacteria and hence,
the infection and death of fish within the lake.

       4)  The dynamic changes in pathogenic bacteria.

       i.  The effect of changes within the aquatic biome upon pathogenic bacteria.
Variations in the population structure of several aquatic species within Lake Xuanwu occurred over the
period involving an outbreak of A. hydrophila. The Crytomonas population was found to decrease
dramatically in contrast to an increase in the numbers of Merismopedia. At the same time, the total
bacteria population in the lake increased sharply.   It should be noted that Crytomonas feed on bacteria,
while Merismopedia is a substrate for numerous pathogenic micro-organisms, including A. hydrophila.
Ingestion of Merismopedia by fish has led to infection by A. hydrophila,

       ii.  The effect of redox shielding level changes upon pathogenic bacteria.
The redox shielding level of Lake Xuanwu is found 5.8 cm below the sediment during the winter
season, and 0.6 cm below the sediment during the summer.  The density of bacteria has been found to
be highest in the redox shielding level  (Saxena et al., 1992, Doran 1982).  Aeromonas sp^ in particular,
are common at this level. Therefore, during the summer when the redox shielding level is close to the
surface of the sediment, the introduction of bacteria from the sediment into the water column occurs
readily.

                                       CONCLUSIONS

       a)  The water quality of Lake Xuanwu has been seriously degraded through eutrophication and
exposure to organic pollutants.  While pollutants  are below lethal concentrations, their presence is
believed to compromise the immuno-response systems of fish within the lake.

       b)  The growth of A. hydrophila is encouraged by the phytoplankton succession resulting from
eutrophication and also by the upward  movement of the redox shielding level.

       -c)  The agent responsible for the recurring incidents of large-scale  fish mortality in Lake
Xuanwu is the conditional bacteria, Aeromonas hydrophila.

                                       REFERENCES

Buchonan,  R.E. and N.E. Gibbons.  1984.  "Berey's manual of determinative bacteriology",  9th ed.
       The Williams and Wilkins Co., Baltimore, p. 545-548.
Chen, H. 1992.  Acta hydrobiological sinica (Chinese).  16(l):40-46.
Doran, J.W. 1982, in "Advances in Microbial Ecology",  K.C. Marshall ed.  6:17-18.
Jiang, L. 1992.  Personal communication.
Ni, D. 1979.  Fish diseases (Chinese), p. 27-28.
Pal, R.N., and S.D. Tripathi. 1978.  J. Inland Fish Soc.  India. 10:166-168.
Saxena, M.P., K. Gopal,  W. Jones,  and P.K. Ray. 1992. Bull. Environ. Contain.  Toxicol. 48:194-201.
Xu,B. 199.1. Researches on fish diseases (Chinese). 13(2): 1-4.
Yu, L., Y. Mao. 1991. Researches  on  fish diseases (Chinese). 13(2):17-19.
                                           Ill

-------
            FLOODPLAIN REHABILITATION ALONG STREAMS
              DEGRADED BY MINING AND MILLING WASTES

                       F.F. Munshower,1 D.R. Neuman,1 and D J. Dollhopf1

                                        ABSTRACT

       Acid generating wastes associated with metal mining and processing have been accumulating
in Butte, Montana (U.S.A.) for over 125 years. In the past, these materials were stored behind
temporary embankments.  Periodic precipitation events breached the dams and flushed the tailings or
other wastes down the local stream (Silver Bow Creek).  The banks of this stream are lined to various
extents and depths with wastes from the erratic floods.  The release of wastes from Butte has been
corrected, but snowmelt or heavy thunderstorms still wash surficial waste deposits back into  Silver
Bow Creek and into its receiving stream, the Clark Fork River. The fishery resource of the upper
Clark Fork River has periodically been devastated by these flushes  of salts and metals.  These metals
also pose a threat to humans living along the stream or river and using ground or surface water
supplies.

       The  Reclamation Research Unit evaluated the use of soil amendments, various incorporation
techniques, and selected species of plants as potential remedial strategies to correct these pollution
problems. Over 35 potential amendments were evaluated for their  ability to correct waste pH and
reduce metal solubility. These amendments ranged from industrial  wastes to  locally mined limestone.
The most effective amendments (CaCO3 and Ca(OH)2)  were utilized in replicated greenhouse trials of
numerous grasses and legumes.  These amendments and selected plant species were combined into a
replicated field trial of different amendment incorporation techniques.

       The  amendments, plant species, and incorporation techniques were evaluated on the basis of
plant cover and production, depth of root penetration, soil neutralization, vadose zone metal levels, and
quality and quantity of surface runoff. The data support coversoils as  most effective.  In  the area of
this study, however, coversoils are not available in quantities adequate for the rehabilitation of all of
the disturbances, therefore, other remedial activities were evaluated. One of these  "other remedial
activities", the pressure injection of lime slurry into the wastes did  not provide an adequate rootzone
for long-term plant growth.  Agricultural tilling provided only 15 cm of adequate rootzone materials.
A plow capable of mixing waste to a depth of 122 cm provided an amended rootzone that varied from
30 to 60  cm deep. Plant cover and production were greater on deep plowed plots than on plots
prepared by the other two incorporation techniques.  Deeper root penetration was also found in the
deep plowed plots than in the plots treated by the agricultural plow or  the injection technique.

       Soil  pore water quality appeared to be slightly improved under each of the treatments, but the
data are not  consistent.  Surface runoff was reduced and surface water quality improved by all
treatments but preliminary analyses indicate that the deep plow improved infiltration more than the
other treatments.
    'Reclamation Research Unit, Montana State University, 106 Linfield Hall, Bozeman, MT 59717-
0290, U.S.A. Reclam. Resch. Unit Publ.  No. 9203.
                                           112

-------
                                       INTRODUCTION

        The Reclamation Research Unit is a division of the Animal and Range Sciences Department of
 Montana State University.  Our research efforts are concerned entirely with the rehabilitation of
 violently disturbed lands.  The Streambank Tailings and Revegetation Studies  ("STARS" as they are
 euphemistically called) are part of an ongoing investigation of the mine waste pollution problems
 associated with toxic Streambank deposits along Silver Bow Creek.

        These Streambank deposits are a direct result of the erosion of tailings and mine wastes
 originally located in and around the city  of Butte, Montana.  Dams on tailings ponds have burst and
 wastes have been carried down Silver Bow Creek by heavy precipitation events or snowmelt numerous
 times over the past  100 years.  The materials have been transported to the Clark Fork River and on
 down the river to the Milltown Dam over 160 km away.

        The area that concerns us extends from Butte to the Warm Springs settling ponds, a distance
 of over 40 km. Several areas along this  stream are locally well known for the depth and extent of
 mine and mill wastes and the severity of pollution.  Various areas of these waste reveal deposits
 ranging from centimeters to meters thick and from less than a hectare to  100's of hectares in extent.
 Waste pH levels range from 2.3 to 7.4; salinity as indicated by electrical conductivity ranges from 0.3
 to 7.5 dS/m or dS nv1 (Table 1). Particle size distributions range from 25 to 95% sand, 6 to 65% silt,
 and 3 to 30% clay.  Elements found in excess at various sites include aluminum, arsenic, cadmium,
 copper, iron, lead, manganese,  mercury, and zinc.

        Because of the large quantities and area! extent of tailings and mine wastes present within the
 Silver Bow Creek drainage, it was decided to explore new and innovative technologies for remediating
 these wastes as opposed to the  traditional method of removal and burial.  The technologies under
 consideration are insitu neutralization, revegetation, and stabilization.  These STARS technologies were
 evaluated for their effectiveness in significantly reducing the toxicity and mobility of contaminants
 present in the wastes, protecting human health, and for cost effectiveness.

                                        OBJECTIVE

       The primary objective of the STARS project was:
          •    to field test and evaluate remedial technologies for the mitigation of
              environmental impacts from Streambank tailings or mine wastes.
              These studies incorporated such concepts as microencapsulation, deep
              incorporation of amendments, coversoil isolation,  and/or enhanced
              evapotranspiration techniques.

       The project team collected performance data on depth of waste neutralization, plant cover and
production, pore water contamination, and surface water runoff quality and quantity.  These data were
necessary to develop alternatives for consideration in the study of waste neutralization.
                                           113

-------
                         Table 1. Range of chemical properties of wastes.
pH
Conductivity
Total Aluminum
Total Arsenic
Total Cadmium
Total Copper
Total Iron
Total Lead
Total Manganese
Total Mercury
Total Zinc
2.3 to 7.4
0.3 to 7.5 dS/m or dS m4
230 to 19,700 mg/kg
20 to 3,100 mg/kg
3 to 110 mg/kg
200 to 11, 200 mg/kg
3,900 to 106,000 mg/kg
83 to 6,500 mg/kg
17 to 13,300 mg/kg
0.1 to 61 mg/kg
19 to 22,000 mg/kg
                       INVESTIGATION APPROACH AND RESULTS

       The STARS investigation is divided into three components:  1) a bench-scale treatability study
(Phase I); 2) a pilot-scale treatability study (Phase II); and 3) the pilot-scale monitoring program
(Phase HI). Phase I was conducted in the laboratory and greenhouse facilities at Montana State
University. Phase n was a field demonstration of the results obtained during Phase I and was
conducted onsite in the Silver Bow Creek drainage.  Phase III was the monitoring and evaluation of
the plots implemented during Phase n.

PHASE I:  BENCH-SCALE NEUTRALIZATION STUDIES: LABORATORY AND
          GREENHOUSE INVESTIGATIONS

       Phase I was designed to develop and test,  in the controlled environment of the laboratory and
greenhouse, materials that would ameliorate streambank tailings sufficiently to allow them to  support
vegetation. In addition, the amendments were selected on their ability to reduce acid and metal
contamination of surface and vadose zone water.

       Existing data concerning tailings and contaminated soils along Silver Bow Creek were
reviewed by the project team.  This information guided the selection of 35 contaminated sites within
the Silver Bow Creek ecosystems. These sites were representative of the wastes along the 40 km of
this stream.  From each of these 35 sites, a 225  kg bulk soil or waste sample was collected.  These
samples were dried, sieved,  mixed, and subsampled for analysis of key chemical and physical
characteristics.  Six primary "kinds" or groups of waste types were identified on the basis of this
chemical and physical characterization. These six sites were representative of the full spectrum of
contamination in the Silver Bow Creek area. A representative example of each group or kind of waste
was selected for further analyses in the laboratory and greenhouse.
                                           114

-------
       Concurrent with the selection of mine wastes or tailings for further study both traditional and
innovative amendment materials were evaluated for possible use in the greenhouse and field studies.
Candidate amendments were evaluated in terms of their ability to raise waste pH,  reduce the solubility
of heavy metals, availability of the amendment, and cost.  Materials evaluated included calcium
carbonate (CaCO3) (agricultural lime), cement kiln dust, calcium chloride (CaCl2), calcium hydroxide
(Ca(OH)2), calcium oxide (CaO), power plant flyash, sugarbeet waste, gypsum, zeolite, triple
superphosphate, fluoro-phospho silicate, phosphogypsum ferric sulfate (FeSO4), phosphate ore, and
various organic amendments.  The chemical and physical characteristics of these amendment materials
were determined. The amendments were mixed with the tailings  or waste materials in specially
designed reaction flasks (Tempe Cells) and their ameliorative properties in controlled and replicated
laboratory column studies determined (Reclamation Research Unit et al. 1989).  Quantities of the
amendments necessary to neutralize various wastes were measured.

       Effluents from the Tempe Cells were evaluated for control of pH and metals by the various
amendments. The leached and amended tailings or wastes were also analyzed.  Results revealed that
the amendments that most significantly reduced acidity were also most effective in limiting the
concentration of soluble heavy metals.  Combinations of CaCO3, Ca(OH)2, phosphogypsum, Fe2SO4,
CaCl2, and triple superphosphate performed best and were subsequently tested in greenhouse plant
growth trials.

       An equation for the calculation of the quantity of lime was developed:

                               t CaCCyiOOO t Soil = 1.25 (A + B)
       where  t = tons
               A = Potential Acidity
               B = Active Acidity
               Potential Acidity = 31.25 (% HNO3 S + Residual S) + 23.44 (% HC1 S)
       where  S = sulfur content extracted from or remaining in the sample
               Active Acidity = SMP Buffer Test

       The  use of calcium carbonate alone did not raise the pH fast enough or high enough to permit
seeding within a reasonable amount of time. We found that portioning the liming rate into 60%
calcium carbonate and 40% calcium oxide rapidly raised the pH and reduced metal toxicity. All
amendments were composed of these two calcium sources in these proportions.

       Plant species used in the greenhouse trials were selected on the basis of their ability to tolerate
acidic soil conditions, nutrient deficiencies and the climatic conditions at the Silver Bow Creek site
(Reclamation Research Unit et al. 1989). In addition, some species were selected for site specific
characteristics such as elevated salinity or textural extremes.  Finally, all species were screened for
commercial availability. Seventeen species and two cultivars of two species were grown in amended
tailings in the greenhouse.

       At the termination of the greenhouse growth period, the plants were evaluated for vigor,
height, color, root growth, and production (Reclamation Research Unit et al. 1989).  Plant tissues were
also analyzed for metal levels.  Livestock and wildlife toxicities from elevated metal levels were
thought to present a concern if the area is revegetated.  In the plant tissues analyzed, concentrations of
boron, calcium, and lead were comparable to tissue levels in nonpolluted areas of Montana. Cadmium,
copper, manganese, and zinc were elevated compared to background levels in Montana but these
concentrations would not be toxic to any grazers and they were not phytotoxic.  Aluminum and  arsenic
levels were higher than most background values in the literature, but they  would not be toxic to
                                            115

-------
 animals. The higher aluminum values were high enough to indicate some plant growth inhibition.
 Metal loading in and on plants grown in the field were much higher. These data are discussed later in
 this report.

 PHASE H:  PILOT-SCALE TREATABILITY STUDIES:
           FIELD DEMONSTRATION INVESTIGATIONS

        Activities involved in what is referred to as Phase II of STARS included the implementation of
 the best amendment options and plant species selections obtained in Phase I into a statistically sound
 field demonstration.  Of the six groups of wastes located along Silver Bow Creek during Phase I, five
 were  selected for further study.  The chemical constituents of these materials were known to be quite
 variable even within an area of relatively uniform deposition. To reduce this variability in the
 demonstration, it was necessary to find homogeneous areas for the field plots. A sampling grid was
 established over each site. At each grid point, pH and electrical conductivity (EC) were measured in
 saturation pastes and total metal levels measured by x-ray fluorescence spectroscopy (XRF).  The
 measured parameters were contour-mapped using geostatistical kriging methods and graphic software.
 The field plots were placed in the most homogeneous locations identified by the  kriged maps.  Other
 factors such as the thickness of the waste materials and the depth to groundwater were also important
 in selecting plot locations.

        After location of the field demonstration sites but before treatments were actually applied, soil
 (waste) samples were collected from each experimental plot to:
          •     measure chemical parameters required to determine amendment rates;
          •     measure plant nutrient status to determine fertilizer rates; and
          •     measure total  metal levels to verify XRF data.

        Amendments incorporated into the various waste materials in the field phase of the STARS
 study included combinations of hydrated lime (Ca(OH)2), agricultural lime (CaCO3), triple
 superphosphate, and iron sulfite.  There were five primary soil treatments replicated four times at each
 STARS site.  These included:
          •     control, no amendment;
          •     agricultural tillage (moldboard plow, discing) incorporation (15 cm depth);
          •    mechanical incorporation (122 cm depth) by deep plow;
          •    Lime slurry pressure injection throughout the 122 cm profile; and
          •     a 0 to 45 cm thick coversoil wedge applied over tailings amended to a depth of 15 cm.

        A dormant fall seeding was planned for 1988.  However, analyses of materials from plots after
 treatment revealed that the pH ranged from 8.5 to 12 in most plots. One month after amendment
 application, some plots still revealed pH values up to 12. Seeding was, therefore, postponed until
 April  1989.

        Seed  mixtures for each site  varied widely.  As in the greenhouse study, plant species selection
 was based upon the characteristics of each individual site, acid and drought tolerance, greenhouse
performance,  and land use.  Unfortunately, most acid tolerant plant species require the equivalent of
over 50 cm of precipitation per year.  The study area receives approximately 35 cm.  Therefore, plant
species were carefully selected that  met the requirements of the sites and were as acid tolerant as
possible. Two seed mixtures were constructed for each site (Schafer and Associates et al. 1989).

       To permit water movement  studies during the monitoring phase of the STARS project, selected
sites were instrumented with suction lysimeters, a neutron access tube, and a piezometer. All sites

                                            116

-------
were fertilized with nitrogen, phosphorus, potassium, and boron.  Rates were based upon nutrient
analyses.  Unfortunately, there were indications from the greenhouse study that the analyses of
nitrogen, phosphorus, and potassium in these severely impacted may not be good indicators of actual
field conditions. Plants showed signs of nitrogen and phosphorus deficiency despite attempts to bring
these nutrients to adequate levels in the growth tubes.  We found similar problems in the field
demonstration.

PHASE HI: MONITORING AND EVALUATING FIELD PLOTS

       The success of the field treatments was determined by evaluating plant species performance,
the quality and quantity of water leaching through the amended materials,  and the quality of surface
water runoff. The plant parameters were monitored in 1989, 1990 and 1991.  Suction lysimeters were
monitored in 1991 and 1992. Surface runoff and root penetration into the wastes were monitored in
1992.  Vegetation evaluation parameters included plant cover, production,  root penetration, and
elemental levels.  Leachate quality and  surface runoff were evaluated in terms of pH, EC, and the
concentration of heavy metals.

       Plant Performance

       Five unique plant rootzone materials were selected for field growth trials. Four different soil
treatments and a control were implemented at each of these sites. It is known that caps or cover soils
will permit rapid revegetation and stabilization of waste materials. This treatment was evaluated at the
study sites to establish a minimum depth necessary for the cover  material and  to provide a comparison
for the other treatments.

       At one  study site, all plants seeded directly into the amended wastes died by the third growing
season. This site was considered the most difficult of the revegetation study sites.  Pre-treatment metal
levels in the waste material were very high (1,750 mg As/kg soil, 11,000 mg Cu/kg soil, and 22,000
mg Zn/kg soil), salinity was elevated at 6.7 dS/m,  and  the site exhibited poor drainage.  This resulted
in vegetation being submerged on several occasions during the study. Waste material at this site will
be excavated and transported to a storage area because of unacceptable risk to humans living in close
proximity.  Response, in terms of vegetation performance at the other four sites,  was variable but
generally indicated that the toxic properties of the  soils had been  ameliorated sufficiently to permit
plant growth and soil stabilization.  Results are summarized for two of the sites in this paper. These
two are Site 7,  which has soil, and Site 21 with a silty clay soil.  Information gathered at the other two
sites is comparable to that itemized for  Sites 7 and 21 in this report.

       Site 7

       Root zone materials  at Site 7 were high in metals and very sandy (Reclamation Research Unit
et al. 1989).  Despite these plant growth limitations the Deep Plow (DP), the Agricultural (A) treat-
ment, and the Coversoil Wedge (CW) produced significantly greater total plant cover than the Control
plots (Table 2).  At this particular site the DP treatment produced almost as  much plant cover as the
CW.

       Production data (Table 3) revealed numerous significant differences  among the various
treatments in grass and total production. The most numerous differences were between the CW's and
the Controls or between the DP treatment and the Control. On the basis of plant cover and production
data at this site, the DP treatment must  be designated as the most desirable soil incorporation method.
                                            117

-------
        Importance Values (TV's) were calculated (relative % cover + relative % production) for the
most common seeded species (Table 4).  These values provide a relative estimate of how well a
species performed in this soil under the treatment and amendment conditions of this study. The
Control plots generally were devoid of vegetation and data from them reinforces the plant inhibitory
information previously inferred from cover and production values, therefore, TV's were not calculated
for these treatments. The IV's from the other treatments indicate that two of the wildryes and one of
the wheatgrasses performed well in this sandy soil.  Observations revealed that these grasses  were
established and reproducing within three growing seasons. Mammoth wildrye (Elymus giganteus)
established well and was spreading by rhizomes (Mean IV =112.3).  Altai wildrye (E. angustus) was
also growing very aggressively (Mean IV = 92.7).  Thickspike wheatgrass (Agropyron dasystachyutri)
was spreading vegetatively on the plots (Mean IV = 35.8). Basin wildrye (Elymus cinereus)  performed
better on the coversoiled plots (IV = 53.7) than on the other  treatments (Mean IV = 20.7).  Legumes
did not perform well on this site (IV = 0.0 to 22.5), but they can be expected to increase in cover and
production throughout the next few growing seasons.  The fescues (Festuca ovina and Festuca
longifolia) will probably decrease in cover as the taller wildryes shade them out of the community.

        On this type of material and this method of soil acid neutralization mammoth wildrye, altai
wildrye, and thickspike wheatgrass constitute an aggressive seed mix.  However, a legume should be
added to the vegetation on the site to lend a nitrogen fixing component and diversity to the new plant
community.

       Metal levels in or  on two of the dominant grasses at  this site are shown in Table 5.  For
comparison, metal analyses of related wheatgrasses (Agropyron smithii) collected in southeastern
Montana are presented in this Table.   Although elevated, the metal levels in or on the vegetation from
this site will not pose a threat to grazing wildlife or the occasional livestock that might pass through
the area.
                                            118

-------




1— I
ON
ON
i— <
b
S
CO
ri
C-
i
£
1?
O
>
8
13

wQ
^
£
8
8

V
&A
«y
£
1
«
1
H










il
!§ "^
£







cs
3
^

s
CO





"3
£
c *
CO
•e
o
ft

CO
2^









1-H
1
CO




3
o
P


1
£
CO
CO
I
o





Treatment


oo
CO




u
m
^H

<
O
vd



o
co ON
CO TT






 oo «o
OCO-

-
g
1
1
d
**H
M
i— i
r^
1
>.
f>
I
_o .
§

c

8

o
to
•a
•if
8
Q
3
§
•rt
i
f>
a
0

-------








T-<
8
T— (
tt
rS
CO
t>
£
CO
S
^•^
c
bb
a
JO
o
1
o<
[elation
tuj
«y
^,
t^
1
CO
4>
1












a
to
**-< O
S o
/P
o
£






cs
i
•o
CO



1


CO
f
MH
CO

1







t— <
nj3
S
CO



1



1
£


1







g
i

e



pq
oo
o
F-




§
OO
oo


0
ON

0
pq
ON
oo

n
^ti
00


"^
o
cn

*
*
oo
CN






•a
u<
1
1



<
ON
»— i




<
06


^
c3


^
00


^^<
T— 1
cn





•n
o.

<5
<,
r-

a
"^^
cn


.
X>
1
O
S
£
!§

CD
: — i
CD
•s
13

^
8
^
_o
•5?

S
Q
.2
ON £*
oo o
^^ T3
£§
^O
*~H J7J
.& o
S ft
etf S
I8

2 1


£3
* *
*
120

-------
8
a «
3 43
fl
^ g

* §
S 'o
'i-g
•S 2
Pi PH
%
'1 °


|'|
(D ^_^
WI


CM «— i en r- CM
^ en O ^H

t-^ oo oo \o in
Qx CM OO CM '— *
CM '^T


o\ \o oo , — i i— i
w ^H m r-H


•S
2 w
a  t~- OO O CM
OS -^t ^4 o ^^
T— ( f-

oo m «n o vo
cO ON *w> ^^ t***
v-H IO T~^


'^ C«
O §
elSSill
Q

CM OO CM O O
•^ vo od c5 o
r— (

en in oo o O
i-H OO

Os en "^ o O
"" °°
c
0
0
QJ V3
*I^H ^
rt --H 0 ^
C W) ^H ^H d> 43
3 
\o o


f^ <— i o ^ CM
en 9i ^ ^ ^

CM en o O r--
CM OO O O l>
en 'vf
fi
o
•43
O
— . ***
Cl K. w ^ ~ 00
S 5P —i o
«— i CM T— I
Oil
*g
^


O "5J td td CL . 1
U

in in <— i o oo
CM ^n OS O \6
•^ VO «— i CM


t~- t> in o oo
r~ in t-: o en

oo oo r- o o
•<*' in <-3 o en
1-1 CM i-l CM
W
fli
^

> JJP i — i 
-------
         Table 5. Comparison of mean elemental levels (mg/kg) in dominant vegetation
                               collected (July 1991) from Site 7.
	 	 	 Treatment 	 	 SE Montana
Agricultural Deep Slurry Western Wheatgrass**
Element Tillage Control How Injection Topsoil (Agropyron Smithii)
Thickspike wheatgrass (Agropyron dasystachyuni)
Aluminum
Arsenic
Cadmium
Copper
Lead
Manganese
Zinc

Aluminum
Arsenic
Cadmium
Copper
Lead
Manganese
Zinc
90.2 B*
3.1 A
0.27 B
14.9 B
12.3 A
78.1 A
69.7 A

66.5 AB
7.1 A
0.39 A
13.4 A
14.8 A
75.1 A
63.8 A
119.0CB
2.2 A
0.41 B
26.6 C
3.90 A
58.4 A
72.0 A

***






86.1 B
2.1 A
0.29 B
15.7 CB
8.8 A
53.3 A
56.4 A
165.0 C
2.5 A
0.23 AB
21.7 CB
6.4 A
76.4 A
47.1 A
Altai wildrye (Elymus angustus)
62.9 AB
3.3 A
0.36 A
12.2 A
3.3 A
53.0 A
57.2 A
134.0 B
7.6 A
0.26 A
15.0 A
6.7 A
60.8 A
44.4 A
33.4 A
. 3.9 A
0.12 A
6.4 A
20.9 A
59.4 A
35.6 A

36.4 A
1.8 A
0.09 A
6.0 A
1.2 A
71.0 A
42.4 A
5 to 50
<1 to 2
0.01 to 0.1
2.5 to 4.5
1.0 to 2.0
25 to 65
12 to 21








  *  Multiple mean comparison based on LSD at significance level of 0.05.  Means
     followed by same letter in rows are not different.
 **  Al and As estimated from the literature, Cd, Cu, Pb, Mn, and Zn from Munshower
     et al. 1987).
 *** Species did not grow on control plot.
       Two soil pits were opened at Site 7 to examine root penetration and distribution.  All soil pits
were placed in DP treatments because previous excavations had provided evidence that penetration was
deepest in this treatment  In both pits, roots penetrated to 25 cm. There did not appear to be any
impediment to deeper penetration and the plant roots may still be developing.  It is doubtful that the
roots have reached their point of maximum penetration.

Site 21

       Table 6 summarizes plant cover at Site 21.  Both grass and total cover data revealed
significant differences among the treatments.  Cover in these two categories was higher on all treat-
ments than on the Control plots. As at Site 7, the average total cover on the A and DP treatments
were similar.
                                          122

-------
t-1

§ o
S|


es
£2
5

T5
Cfl
3
O

CO
•B
o
fi
CO




t-H
•'' i§
T3
1

"3
o
H
C
CO
•e
o
CO
CO

C
n

<
0
d


PQ
U
en
8
en
>n


•n
d

*
oo



•a
3
1

^
o
d
0
d

<
o
d


o
d
<
0
d


o
d

<
o
d




"o
1
U

U
p
vo
U
1— H
VO

<
O
CN


U
ON
8
en
cs


en
c^**

CO
p
>n
>n



^
S
I

PQ
^°.

PQ
p
en

<
vo
d


• PQ
oo
en
PQ
S


<
CM

PQ
o
55?

e
o
1
S1
t-H
00

8°

. VO
U U
VO ,-H
en d

«
en -*
d cs


U U
Tf 00
en t-^
8°
C~; •


^ ON
. ^ en

PQ r )
U vo

 oo «n
o en i-<
U
123

-------
        The production data (Table 7) also revealed differences between the Control plots and
 the treatments, but there were almost no differences between the treatments and the CW. As
 at Site 7, production and cover on the DP plots appeared to be very successful and at this site
 production on the DP plots were comparable to the CW.

        Importance Values suggest that the crested wheatgrasses performed very well on the
 CW's but not on the other treatments (Table 8).  On Seed Mix 1 plots, Tall wheatgrass
 (Agropyron elongatum) dominated the landscape.  Intermediate wheatgrass (Agropyron
 intermedium) revealed an identical pattern on Seed Mix 2 plots. Both of these wheatgrasses
 shared dominance with the crested wheatgrass species (Agropyron cristatum and Agropyron
 desertorum) on the CW's.

        Metal  levels in or on the plant tissues from these plots (Table 9) were generally ten
 times those found in similar vegetation at Site 7.  The differences between metal levels in the
 vegetation growing on the CW's at the two sites were much lower, however.  Lower metal
 levels in this vegetation may be due to lower metal levels in the root zone or reduced dust
 accumulation  on the elevated wedges.  The possibility of trace element imbalances in
 livestock and  wildlife forages must be considered. Copper and zinc levels in  or on vegetation
 are high enough to produce problems to animals grazing these sites throughout the year.
 Cadmium levels are elevated above levels recommended for livestock consumption (NRC
 1980).

       Root penetration at Site 21 was quite variable. In two soil pits opened in one plot
 amended by slurry injection, roots penetrated to 12 cm and 20 cm.  In a replicate  of the slurry
 injection treatment, root penetration was to 17.5 cm. In a third replication of this  treatment,
 roots had penetrated to 27.5 cm.  This latter example was noted as the best revegetated slurry
 injection plot In the only A plot examined, roots stopped growing  vertically  and  branched
 laterally at 17.5  cm.  This was the limit of lime incorporation on this plot. In a DP treatment,
 roots had penetrated to 38 cm.  Even at this depth there was no indication that the roots were
 at their limit of penetration.

       The increase in production (and cover) of the crested wheatgrasses  on  the CW's
 indicates that these species will do well when a coversoil is available, but are  not  capable of
 successfully competing on finer textured, acid producing wastes even if amended.  On this
 type of materials, Tall wheatgrass, Intermediate wheatgrass, and Russian wildrye (Elymus
junceus) appear to be most successful.  Their use  with a legume must be recommended on
 this type of amended waste.

       Water Quality

       One of the  major impacts  of mining related wastes is upon water.  Both ground and
surface water are affected by these materials with detrimental effects upon  water for
                                         124

-------





i — i
o\
OS
T— 1
8
1
|
T— (
CN
S
*rH
CO
/_v
*
5
1
a
O
1
I.
0
1
2
0
1






0
•J3
•8 J
S|
O
H



CN
X
*jpH
"S
CO






T 	 1
1
CO


1

co
•e
0
PL,

CO
1
O
3
£

CO
I

i
1



I
1



ffl
U
i — i


ffl
00



VO

ffl
vo
(N

ffl
V— 1
O
en
CN

'<
O
55

ffl
vj
^•n
cs



1
oversoil
U

^J
1
o
.a
• ^
1
1
>-.
1
1
£
S
1

0
i— t
o
13
8
1
.S?
^
Q
CO
oo °
O 13
^ d
rt §
S O i
cS Q
^ ° *J
2CQ £D
1 s|
fl [ft 4J
O *cj O
H ^ a
* *
*
125

-------
o

a|
1 1
II •*•* ,R
it ii"M >^»
££
^
1^
R
•s 1
*I
(D
O
o 13
||

?45
II
^33 r )
o
' »^
CO 43
g

vo ^f vo CN CD
«> K co' CN CD
CO "O
<— i
0 CN CO
1—1 £D O CO O
>*^

VO CN vo ON O
^-t ^ CO VO CD
P- *— <

13 to
S^J
•3 -8 .g S * 1
O •?-> 5jQ

T-f CO' 3 O co ••* VO ^ O ^ O CO fc CO ^ O ^f *~^ O\ O f~. T- i CD 00 VO ° CN ° £& oo o 2 S ^ a 'd s ^^ ^O »IH i^ n3 [3 oo co co r- vd CD CD O i-H o vq co o co' ON CD CD oo r- o p- co 0 CD CD ° * ~ 1 ^ a 13 "d 3 o . ft M_ too 4) <£ >. O n CN o o CN ON CN O CD ON VO CN r-H ^ CN VO r- i CD O vd oo CN* CD* CD T CO i— 1 CO ON »— i O O VO CD CD CD CD "^T CO * — 1 r-H •§ > 52 T3 .5 6 3 3 S < < <; S £ t~^ "^t ^i" ^ Tf CN VO CD ON CN f~- O **«<« OO CN [~- O 8| ^ 1 oo 173 g 1 a -S 'C! 1 g too too too ?f § <3 « J 126


-------
       Table 9. Comparison of mean elemental levels (mg/kg) in dominant vegetation
                             collected (July 1991) from Site 21.
Treatment SE Montana
Agricultural Deep Slurry Western Wheatgrass**
Element Tillage Control Plow Injection Topsoil (Agropyron Smithii)
Tall wheatgrass (Agropyron elongatuni)
Aluminum
Arsenic
Cadmium
Copper
Lead
Manganese
Zinc
316.0AB*
26.1 AB
2.0 B
278.0 B
30.5 AB
917.0 B
750.0 B
**






384.0 B
33.2 B
1.7 AB
247.0 AB
60.6 B
672.0 AB
541.0 AB
302.0 AB
21.9 AB
1.5 AB
216.0 AB
25.1 AB
512.0 A
393.0 A

74.0 A
4.3 A
0.61 A
63.9 A
5.4 A
320.0 A
249.0 A

5 to 50
<1 to 2
0.01 to 0.1
2.5 to 4.5
1.0 to 2.0
25 to 65
12 to 21
Intermediate wheatgrass (Agropyron intermedium)
Aluminum
Arsenic
Cadmium
Copper
Lead.
Manganese
Zinc
311.0 B
21.7 B
1.9 B
228.0 B
28.3 B
790.0 B
589.0 B
***






266.0 B
22.3 B
1.6 AB
196.0 AB
38.2 B
622.0 B
491.0 AB
286.0 B
26.6 B
2.1 B
268.0 B
29.7 B
677.0 B
575.0 B
48.4 A
3.1 A
0.57 A
59.5 A
3.5 A
243.0 A
195.0 A







 *  Multiple mean comparison based on LSD at significance level of 0.05.  Means
    followed by same letter in rows are not different.
 .** Al and As estimated from the literature, Cd, Cu, Pb, Mn, and Zn from Munshower
    et al. 1987).
***  Species did not grow on control plot.
human consumption, agricultural uses, and fishery resources.  Toxic elemental concentrations in these
two water sources were measured to clarify the changes produced by the amendments and treatments
used in this study.

       Pore Water Quality

       The impact of amendment and treatment upon soil pore water chemistry was measured in
water pulled from suction lysimeters.  Pore water quality at Site 7 appears to indicate that there was an
effect of the addition of lime amendment to the soil at this site (Table 10).  Most of the metals in
water pulled from the treated plots show a decrease in soluble concentrations and calcium an increase
in concentration compared to the levels reported in the Control plot.

       Data from Site 21  do not show these differences.  Control plot data are very limited because
the soil zones at 40 and 150 cm were too dry to provide a pore water sample with these lysimeters.
Furthermore, concentrations of metals in pore waters from the various treatments at this site (Table 11)
were higher than the  values recorded at Site 7.
                                            127

-------
         Runoff Water
             study area has a history of fish kills resulting from the erosion of metals, salts, and acid
 into the receiving nver after precipitation events of sufficient magnitude to wash mining wastes into
 £™aVef'- r ,  !•    f aSlUmed ^ amendment of ^ materials and vegetation growing on them will
  ncrease infiltration, leach salts deeper in the profile and reduce erosion of surface salts  During the
 tiurd growing season simulated rainfall was applied to the plots to evaluate this phenomenon
 Precipitation  rates  were 5 cm/hr for two hours. Runoff was collected, measured, and elemental levels
 in tne water determined.

        The coarse textured material at Site 7 had relatively little runoff, especially after
 amendment application and plant establishment (Table 12). The DP and Ag treatments yielded
 insufficient runoff for chemical analyses for the elements in question.  Runoff at Site 21 was adequate
 from all treatments for the analytical determination of the elemental levels. Results of the  analyses
 support the hypothesis that amendment application will reduce pollutant concentrations in the water
 Table 13 shows the remarkable decrease in heavy metal levels in runoff from all treatments at Site 21.

                                        CONCLUSIONS

  f A   L!m®  f™endment of acid generating mining wastes has been shown to reduce the phytotoxicity
 of the materials but incorporation of the amendment into the wastes has been  and continues to be a
 major problem.  This study evaluated amendment and incorporation of wastes into several unique
 types of mining wastes. Amendments consisted of calcium carbonate and calcium hydroxide but the
 amount of the amendment added to each study site varied with the properties of the individual sites
 Incorporation  techniques included slurry pressure injection, agricultural plowing, deep plowing  and'
 coveting the waste with a soil cap. Several parameters were measured to determine the effectiveness
 of each incorporation technique.

        While capping any  waste with soil is a preferred rehabilitation technique, expense'and lack of
 suitable materials often prevents the use of this technique.  Of the alternative incorporation  techniques
the DP appears to produce the most satisfactory response.  Plant cover and production were higher and
runoff was reduced  and improved by this technique.  Nontoxic plant rootzones were also deepest under
this method of amendment incorporation.
                                           128

-------
i
fN
i
i
o















x— >
H-1
s
g
•^
£
tJ
g
o


13
a
1
tti
s
















03



£4



a
^



00
s

X>




O
£


a



a



. -o
u

I

l^fl
_1
§ ^
fl r\
03 }^
£* &
i*
00
88
^^ r*H
, ON ON
' "* ^
0
^^ ^^
o o^
, oo >o
! vo i-i
0 O
o CN
CO OO
! 
O O
o o
C-f V£5
! CO CN

08
CO CO^
! co CN
>o
C^
i -H 10



00
! £8
80
°. ^
C""* ^*.
i CO CN
08
o o.

. OO CO
! i— i i— i

vo ^h
1 i-i «r>
, vo CN
o





-s
U o
PS o o n
! «n CN
0 0

^ CN
i co" i>
O O

, i-? ON


08
co vo^
! K S

• *
i -rh 1-1


O
* s
, CO i-?
i i— 1 CO
0
"1,
r^~ vo
! i^ co
O
""> cO^
, ON >0
! VO i-H

1"H
i '-' *
! CN C4
! t> ^

5" i-l CO
Q vo oo
a
o
8
"5
a
i 00°
55 ^^^
0 O

O^ ^ Q
CN i— i >O
O O O

•<*" ON ci
oo co CN

000
co i — i O
co i— i o
co c~~ c~~

o
00 ^
I-H' 00 CN

ooo
oo^ ^r ^4"
CN" 1-1 CN

* * *
i — i ^ i — i




^- vo ON
1-1 CO CN

CN
000
^H VO r-H
gll
"T vo co
r- ^ •*
T— t T— * r-H

O vo *
^H ^H CN
VO vo
l-H CN 1—*

ON VO i-l
>0 »-i VO


£
0
S
O O
« ^. ON ^
                                                                     o
                                                                     •a
                                                                     o
                                                                     I
                                                                     I
                                                                     CQ
                                             129

-------
cs
2
S3

I

I"

I
I











2
a
c
1

f.
p
CJ
c
E
S















ol



W

-


00
s

g



*

a


<3


u


£


1 "^ t^
'
||

w O
[•••4 f™^
CO
o
en

§
ON
£"•*«
1 *— * |
O
o
! So !
0
i OO
1 1— * 1
8
&
i »— i i

! 3 .


! £ i

ON
i CM ,
§
£;
• m !


i «n i

• 3 i

ik OO £&
H vo «


-3
o o o »n
0 ¥ON~

en
, f~ en
! en cs
§|
ON Tj*
, *r> r-i
' *° T*
O 0
oo oo
i I-H O i-?
0
, r- cs
! O ON
Q
o"00.
! t-H ,-H

i -H 3


i 5SS

t» «0
I ON Tt
So'
O O
. 00 VO
! cs en


i en cs
, ss
1 VO Tt
^*-
J? T"H f"^"
O ^-H \r>
§
'f?
8
M
& o
a o o >o
,3 •* ON I-H
CO
o o
VO T^ oo
«O Q OO
vo ^F en
sis
ON vo H.
^- O vo

0 O 0
en o
en  vo en
CO OO ^

^
a
 O O 
^-1 r-H
80
0 0
•/"> e^f


en

-------
          s
1
1
CO
          t
I
1
<&
•*
I
Tre
           GO
1
T— (
oo
-
r-
vo
T-I
T— I
i — i
2
r— <
a
o
s
t>
-
CO
*— 1
.
CO
t— (
00
a
'If
0
d
1
1
o<
o
d
I
Agricultural
-






co
3

"o
GO
CO
«
„
cs
o
ON
T— 1
0?
o







o
T— t



o

Agricultural
- T 	 1
                                              131

-------
                                  LITERATURE CITED

National Research Council.  1980.  Mineral Tolerance of Domestic Animals.  Sub-
       committee on Mineral Toxicity in Animals. National Academy of Sciences, Washington, DC
       577 p.
Reclamation Research Unit (MSU), Schafer and Associates and CH2M HILL, Inc.
       1989.  Final Summary Report. STARS Phase I: Bench-Scale Soil Column and Greenhouse
       Treatability Studies, and Tailings Ranking System. Streambank Tailings and Revegetation
       Studies.  Silver Bow Creek RI/FS. Document No.:
       SBC-STARS-PHASE I-F-R1-102589. Montana State University, Bozeman, MT.
Schafer and Associates and Reclamation Research Unit (MSU).  1989. Final Summary
       Report. STARS Phase II: Field-Scale Treatability Study Plot Construction.  Streambank
       Tailings and Revegetation Studies. Silver Bow Creek RI/FS.  Document No.: SBC-STARS-
       PHASE H-F-R1-051789.  Montana State University, Bozeman, MT.
                                        132

-------
                NONPOINT SOURCES AND WATER QUALITY

                 Lee A. Mulkey1, Robert R. Swank Jr.1, and Rosemarie C. Russo1

                                      INTRODUCTION

       In the twenty years since passage of the Clean Water Act in the United States, much progress
has been made in decreasing pollution by sewage and industrial wastes of the nation's rivers, lakes,
and coastal waters. Water pollution problems from point sources, such as municipal and industrial  .
outfalls and other sources that are identified as coming from a clearly defined location, still exist and
work is continuing in this area.  However, it is important to concentrate more efforts on reducing and
controlling nonpoint source pollution. Nonpoint source (NFS) pollution comes from diffuse sources
and enters surface and ground waters in many ways, often in surges or storm events and often in large
quantities.  NFS is the largest remaining category of contamination threatening water quality in the
United States.  Effectively dealing with NFS pollution is a much more difficult task than correcting
point source problems and requires different approaches.

         Fisheries, wildlife, and recreation are the uses most affected by NFS pollution.  Severe
damage has been caused to aquatic communities nationwide, and the aesthetic values of many of the
nation's recreational waters have been destroyed. Current estimates assign 45% of impaired estuarine
areas, 76% of impaired lake areas, and 65% of impaired river miles to NFS causes.

       The United States has lost billions of dollars of benefits from aquatic resources since the NFS
problem was documented almost 20 years ago.  Sediment causes decreased light transmission through
water, which results in decreased plant reproduction, interference with feeding and mating patterns  of
aquatic animals, decreased viability of aquatic life, decreased recreational and commercial values, and
increased drinking water costs. Increased siltation also destroys critical habitat for certain sport fishes
and prematurely fills storage capacity of reservoirs. Nutrients promote premature aging of lakes and
estuaries.  Pesticides and herbicides hinder photosynthesis in aquatic plants, affect aquatic
reproduction, increase the susceptibility of biota to stress, accumulate in the tissues of fish and other
aquatic organisms, and are hazardous to both human health and wildlife through drinking water and
the food chain.  NFS pollution has resulted in beach closures, fishing bans, decreased property value,
erosion, loss of aesthetic appeal of lakes and streams, and contamination of drinking water supplies.

                              SOURCES OF NFS POLLUTION

       NFS pollution is contributed by agriculture, urban runoff, atmospheric deposition, highway
drainage, hydromodification (stream channelization, flood prevention, lake drainage), silviculture,
livestock grazing, construction practices, land disposal activities, septic systems, landfills/spills, and
mining.  Some NFS problems are generated from more than one kind of source and only recently have
been recognized as requiring multimedia control programs. For example, nitrogen has diverse sources,
plays an essential role in biogeochemical cycling, and can create ecological and human health
problems.  Pesticides and pesticide transformation products represent a major NFS concern that has
potential environmental and health effects, and that can adversely impact both groundwater and surface
water resources.
    Environmental Research Laboratoiy, U.S. Environmental Protection Agency, Athens, Georgia,
U.S.A.

                                          133

-------
 AGRICULTURE

        Agriculture is the largest NFS, affecting 50-70% of impaired surface waters.  Siltation of
 surface waters is a major pollution problem arising from eroded croplands and overgrazed pastures and
 rangeland. Fertilizers and manures are major pollutants causing impacts in surface waters.  Nitrates
 and pesticides are increasingly found in groundwater, arising from their use on crops and from
 livestock wastes.  Animal waste from confined livestock facilities contains, in addition to nutrients and
 oxygen-demanding organic matter, bacteria that can cause drinking water contamination, fish kills, and
 shellfish bed closures.  It also contains phosphorus that accelerates eutrophication of freshwater lakes.
 Agricultural activities often dramatically alter the landscape, increasing the vulnerability of sensitive
 ecological habitats including wetlands and riparian zones.

 URBAN RUNOFF

        Urban runoff NFS problems arise from streets and parking lots, and from industrial  sites.
 Such runoff carries salts  and oily residues  from road surfaces and may also carry nutrients and toxic
 organic chemicals and metals.

 ATMOSPHERIC DEPOSITION

        Short- and long-range transport in the atmosphere and subsequent deposition on land and
 surface waters of a variety of pollutants, notably ammonia, nitrate, heavy metals, phosphate, and
 pesticides, are significant sources of NFS pollution in surface waters, both fresh and marine.

 HIGHWAY DRAINAGE

        Runoff from highway surfaces results in salts and oils being transported as pollutants to
 waters. Nutrients  and combustion product organics may  also be in highway runoff.  Pollutants from
 highway construction and maintenance may find their way into surface waters.

 HYDROMODIFICATION

        Habitat modification, polluting chemicals, and increased sedimentation adversely affect water
 quality; these problems result from construction and other engineering projects such as dam  building
 and stream channelization.

 SILVICULTURE

        Forestry operations, such as road construction and timber cutting, result in erosion, causing
 siltation problems in  streams and lakes. Debris and pollutants from logging roads and habitat
 modification from logging cause NFS problems in surface waters and for forest wildlife.

CONSTRUCTION PRACTICES

        Development activities such as clearing land, building new structures, and modifying existing
structures produce sediment and toxic substances that can result in  adverse effects in receiving surface
waters.
                                            134

-------
LAND DISPOSAL ACTIVITIES

       Toxic pollutants from hazardous and nonhazardous waste disposal sites can contaminate
ground and surface waters.  Sewage sludge spread on land and leakage from household and
commercial septic systems can also adversely affect both surface and ground water systems.

SPILLS

       Inadvertent spills during transport of chemicals or chemical wastes may cause pollution of
surface waters with hazardous chemicals, sometimes resulting in kills of aquatic plants and fishes.

MINING

       Runoff from mining activities can result in sedimentation problems  and the addition of
surfactants, acids, metals, and toxic organics to surface waters.  NFS problems can also arise from
abandoned mines, waste piles, and oil and gas wells.

                                 REGULATORY MANDATE

       Legislative mandates for the U.S. Environmental Protection Agency (EPA) to address NFS
pollution are primarily the Clean Water Act of 1972, as amended in 1987, and the Coastal Zone
Management Act (reauthorization amendments of 1990).

     The Clean Water Act as amended establishes a national program to control nonpoint sources of
water pollution.  Two major new requirements were established:  (1) reports from the 50 state
governments describing water impairment due to nonpoint sources, the types of sources causing the
problems, and state and local control programs; and (2) state programs for controlling nonpoint source
pollution including methods and a time frame for remedying problems.  The U.S. Environmental
Protection Agency began a three-year grants program in 1990 through which EPA is awarding $140
million to states  to support approved state nonpoint source programs.

       The Coastal Zone Management Act Reauthorization Amendments of 1990 require that states
develop a Coastal Nonpoint Source Control Program. The. Act establishes a "technology-based"
approach to NFS management with the associated presumption that best management practices will
meet water quality goals.

       Multimedia issues will require coordination across legislative mandates; for example, "trading"
permits for nitrogen may be issued across the Clean Air Act and Clean Water Act. Solutions to NPS
pollution problems will also require close cooperation among many local, state, and federal (U.S. EPA,
U.S. Department of Agriculture, U.S. Geological Survey, U.S. Fish and Wildlife Service, Federal
Highway  Administration) agencies.

       In the United States much of the regulatory activity directed toward controlling NFS pollution
is voluntary.  That is, landowners and operators like farmers, ranchers, and foresters, are encouraged to
adopt Best Management Practices (BMPs) that help control NFS pollution while sustaining production
and profitability. Federal and state government institutions provide technical  and financial assistance
in this activity.
                                            135

-------
                                    RESEARCH APPROACH

         The logical assessment, management, and restoration unit for research addressing NPS
 problems is the watershed, the surrounding land and associated land uses that contribute to the aquatic
 system.  Problems are identified in a watershed context; long-term solutions require watershed or
 geographic area management plans that address the multiple sources of, and solutions to, the NPS
 problem.

         The U.S. Environmental Protection Agency's research strategy is to develop the data,
 assessment tools, and technical outreach to answer four key questions: (1) What are the specific
 causes of the observed problems within impacted watersheds?  (2) How can major nonpoint sources be
 reduced by prevention, restoration, and watershed management?  (3) How do we determine the
 effectiveness of prevention, restoration, and management efforts? (4) What policies and mix of
 technologies are most effective in solving problems at different geographic scales?

        The focus is on watershed scale processes and responses that will enable an integrated,
 multimedia approach to problem identification, human health and ecological assessment, pollutant
 runoff and leaching, control, landscape and stream  system restoration, and pollution prevention.  The
 research embodies geographic targeting of watersheds and problem areas within watersheds, integration
 Of the effects of all stresses (physical, chemical, and biological), incorporation of designated use
 considerations, and trading of "pollution rights" between point and nonpoint sources within watersheds.

        EPA's current NPS research plan has five major scientific components:

        (1) Watershed Assessment Methods. Methods are needed that integrate geographic
 information systems, remote sensing, modeling, monitoring, and diagnostic  bioassessments to
 characterize the sources, stressors, controls, and management methods needed to improve, protect, or
 restore watershed ecology and water resources.  Currently, such methods  do not exist. Assessments
 are typically limited in scope, address single media impacts, apply to limited scales, and often lack a
 comprehensive basis for action.  Priority is now being given to  developing a systems  analysis
 capability for geographic targeting, policy analysis, watershed management  planning, and relating
 source controls to impacts.

        (2) Watershed Management and Control Technology. Best management practices applied to
 individual sources and watershed practices applied to mitigate multiple sources are the currently
 mandated approaches for NPS management within developed watersheds. Data documenting their
 effectiveness are limited, cause-and-effect relationships are uncertain, and field programs that measure
 the cumulative impacts of watershed level controls are urgently  needed.

        (3) Watershed Restoration.  Substantial public investments are proposed for restoration of
 severely damaged watersheds. NPS management compatible with restoring  the ecological integrity of
 watersheds is a major design and implementation challenge for such projects.  Priority is being given
 to developing data on the costs and effectiveness of various NPS management approaches for
 watersheds under restoration.

       (4) Pollution Prevention.  Largely undeveloped but developing watersheds should be managed
where possible to sustain natural  conditions.  Innovative approaches that limit pollutant inputs to
watersheds and replace existing NPS-generating practices with sustainable development concepts are
emerging and must be verified, documented, and promoted.  Priority is being given to research on
integrated farm management systems such as farming by soils; integrated pest management; chemical

                                           136

-------
and waste input reductions (variable application technologies); innovative forestry practices; new
mining methods; and sustainable development in urban, suburban, and rural environments.

       (5) Multimedia Assessment for Nitrogen Management.  Gaps in knowledge of biogeochemical
cycling processes are being filled by conducting process studies of forest response to wet/dry fall
nitrogen loads; by developing better understanding of nitrogen movement and utilization in
agricultural, urban, and suburban settings; and by developing methods for quantifying the effects of
anthropogenic nitrogen inputs to estuarine systems in terms of measurable changes in estuarine
community structure and function.

       The research begun in 1992 emphasizes a concerted,  systematic approach to solving the NFS
pollution problem, beginning with the primary source of many of these problems--agriculture~and also
focusing on the most pervasive pollutants-nitrogen and pesticides.

                            THE MASTER RESEARCH PROJECT

       In 1992 the U.S. EPA, U.S. Department of Agriculture, and U.S. Geological  Survey initiated a
major research project to begin to implement the long-term NFS research strategy. This project is a
five-year, $35 million study on the  environmental effectiveness of midwestern agricultural NFS
controls.  The EPA component of the tri-agency project is called  MASTER (Midwest Agrichemical
Surface/Subsurface Transport and Effects Research).

       Through the research being conducted in the MASTER project, EPA is working to understand
the ecological impacts of agrichemicals and to separate out and distinguish these impacts from those of
other types of agricultural practices such as habitat and hydrologic modification.  At the Walnut Creek
Watershed field site in the state of Iowa (Figure  1), MASTER studies are underway to determine the
effects of agrichemicals on aquatic  ecosystems'  structure and function (Figure  2), to quantify the
subsurface assimilative capacity for agrichemicals, and to quantify interactions of groundwater and
surface water in this watershed.  The effectiveness and longevity  of alternative agricultural
management practices are being projected and monitored to enable the prevention of ecological
degradation and derive potential ecosystem restoration methodologies.  Diagnostic and predictive
ecological and hydrological modeling tools are being developed (Figure 3) to assess and minimize the
ecological risk from agricultural management practices in this watershed. This involves linking
existing fate and transport  models and ecosystem effects models to predict chemical movement and
resulting ecological impacts at different spatial and temporal scales as a function of control/restoration
strategy.

       The research project  is still in its early stages. However,  an example of the kind of
information being obtained is shown in Figure 4.  The current distribution of stream atrazine
concentrations for the western corn beit plains can be used to determine the proportion of streams in
this region with atrazine concentrations that exceed the proposed  atrazine water quality criterion of 3
micrograms/liter.  If alternative agricultural practices were implemented on the highly vulnerable
watersheds within this region, it might be possible to reduce the proportion of streams exceeding the
criterion from 70% to 30%.  This reduction would reduce the risk to stream resources.

       By the end of 1992 a preliminary assessment of the ecological effects  of agricultural
management practices, including ecological restoration techniques, will be completed for the Walnut
Creek Watershed (15 square  miles). By the end of 1996 the  MASTER project will provide the tools,
criteria, guidelines, and scientific information for the development of regional  and national agricultural
                                           137

-------
policies and regulations that will improve the quality of the nation's water resources while maintaining
economic sustainability of the nation's agricultural industry.

               FEDERAL, STATE, AND LOCAL COOPERATION IS NEEDED

       U.S. EPA and other federal agencies can provide scientific information through research,
technical expertise, education, and some funding to help solve water quality problems  arising from
NPS pollution.  But state and local governments must address NFS pollution through careful land-use
and development planning that protects aquatic resources and aquatic and terrestrial habitats. Farmers
and foresters must employ production practices that protect nearby streams and groundwater.
Individual citizens must help by such measures as taking used oil to collection centers, applying lawn
and garden chemicals carefully, and disposing of chemicals and debris in a safe manner.  Only if all
sectors of the populace do their part can the major problem of nonpoint source pollution be  effectively
solved.

                                      BIBLIOGRAPHY

Griffin, R.J. Jr.  1991. Introducing NPS Water Pollution.  EPA Journal 17(5): 6-9.
Mulkey, L.A. 1992.  Nonpoint Source Issue  Strategy. Environmental Research Laboratory, U.S.
       Environmental Protection  Agency, Athens, GA.  3 p.
Reilly, W.K. 1992. The Issues and the Policy.  EPA Journal  17(5): 20-24.
Russo, R.C., L.A.  Mulkey, R. Carlson, and A. Fairbrother. 1992. Nonpoint Sources Research Plan
       FY93-FY97.  Environmental Research Laboratory, U.S.  Environmental Protection Agency,
       Athens,  GA. 55 p.
Swank, R.R. Jr. et al. 1992. Research Plan for Midwest Agrichemical Surface/Subsurface Transport
       and Effects Research (MASTER). Environmental Research Laboratory, U.S. Environmental
       Protection Agency, Athens, GA.  87  p.
                                          138

-------
 ECOLOGICAL ENGINEERING SYSTEM FOR THE CONVERSION OF
    MUNICIPAL WASTEWATER INTO RESOURCES IN ZHENJIANG

                          Yu-hui Zhuo1,2 Xu Shen1 and Xiao-liu Geng1

                                        ABSTRACT

       Xhenjiang (E 119°58', N 32°19') is a mid-sized industrial and tourism city with a population
of 350,000, located in the south of Jiangsu Province and closely situated to the Yangtze River.  With
the rapid development of industry and increase in population in the past two decades, the problem of
water pollution has become very serious, resulting in industrial water shortages and restraints on
further economic development.  To solve these problems, a research program (198601991) was
undertaken cooperatively by the Environmental Science Department of Nanjing University and the
local Environmental Protection Authority.  After technical and cost-efficiency evaluation of a number
of wastewater treatment plans, ecological engineering for the conversion of municipal wastewater into
resources (oxidation ponds + land treatment systems) was selected as a cost-effective and
environmentally sound strategy.  This plan was accepted by the local government and listed on the
World Bank's loan program for environmental protection in the southern part of Jiangsu Province.  In
June of 1992, the design plan successfully passed the comprehensive assessment of the World Bank.
The final wastewater treatment capacity of the project will be 300,000 tons per day and the total
investment will be 160,000,000 RMB (including a $20 millon US loan from the World bank). The
construction period will cover 5 years (1993-1997).  The  two key parts of the project are the three
municipal sewage intercepting systems which will be built along the main water bodies in the urban
district (Harbour, Old Grand Canal and Yueliang River) and the ecological engineering systems which
will be constructed on approximately 13 square kilometers in a suburban region 2 kilometers away
from the urban district of Zhenjiang.  The ecological engineering system will be comprised of
oxidation ponds, land treatment systems, and several types of ecological farms. The treatment process
involves collection of all the municipal wastewater by the three sewage intercepting systems, which
will then be pumped into the ecological engineering system to be treated until discharge standards are
met. Any wastewater that can be diverted or converted to resources (crops, fisheries, poultry and
wood) will be. Treated water will eventually by discharged into the main stream of the Yangtze
River.  This paper will describe the project in detail and analyze its feasibility. It is expected that this
project will solve the dual problems of water pollution and industrial water shortage in Zhenjiang and
that the city will benefit socially, economically, and environmentally. And of equal importance, it will
become a demonstration project and a major field experimental base for studying and developing
low-cost ecotechnology suitable for the treatment and utilization of wastewater in cities and towns of
southern China.
                               ENVIRONMENTAL SURVEY

GEOGRAPHICAL LOCATION

       Zhenjiang city (E 119°58', N 32°19') is located on the southern shore of the Yangtze river in
the center of the Jiangsu province. The cities of Nanjing and Shanghai lie 70 km upstream and 503
km downstream from Zhenjiang, respectively.
    'Environmental Science Department of Nanjing University, P.R. China
    2Director, Water Pollution Control and Conversion of Wastewater into Resources, Nanjing University,
P.R. China
                                           139

-------
 CLIMATE

        Zhenjiang experiences a northern, subtropical monsoonal climate that is typically warm and
 humid.  The annual mean air temperature is 15.4°C and the frost-free period averages 238 days out of
 the year. During the summer, the prevailing wind direction is from the southeast with an annual mean
 wind speed of 3.6 m/s and an annual maximum of 16 m/s.  The average annual rainfall is 1060 mm
 with 45% of the total rainfall occurring during the summer.

 SOIL

        The predominant soil type of the delta area near the Yangtze river is a fine, silty, sandy soil
 over clay, while in the floodland formed near the Yangtze are sand clays, slime sand clays, and silty
 fine sandy soil.

 VEGETATION

        Mixed broad-leaf forests of both evergreen and deciduous tree species comprise the dominant
 vegetative cover.  On undeveloped river banks (such as the Zhenrun and Jiaobei shoals) there are
 helophytes such as reeds, medical zizanta caducifloria, acorus calamus, and sparganium stoloniferum.

 LAND AREA AND POPULATION

        The urban district of Zhenjiang covers 26 km2, with a 1990 population estimated at 350 000.
 Therefore the population density is 13562 people per square kilometer.

 ECONOMY

        Zhenjiang is a mid-sized industrial city with a diversified economic structure, of which
 electronics, machinery, textiles, paper, construction materials, and chemical production are the
 predominant industries.  The gross industrial production of the urban district of Zhenjiang in 1988 was
 3.4 billion RMB.  Zhenjiang is also an important port city and a major, international tourist
 destination, plays a vital role in the economy of the city.

                  AQUATIC ENVIRONMENTAL  QUALITY ASSESSMENT

 AQUATIC ENVIRONMENTAL BACKGROUND

       Within the urban district of Zhenjiang, the primary water bodies are the Yangtze river, the
 Great (Old) Grand Canal, the Harbour, and the Yueliang river (Fig. 1).

 BSTVESTIGATION OF WATER POLLUTIONS SOURCES

       In 1990, the total wastewater discharge from the city reached 200 000 m3 per day, of which
 most was untreated (Table 1).  Over the years, this practice has resulted in serious water pollution
problems in all of the main water bodies in Zhenjiang.

COMPREHENSIVE ASSESSMENT OF WATER QUALITY

       Evaluation of water quality was made using the P.R. China national surface water quality
standard (GB3838-88) and 8 water quality indices (chemical oxygen demand (COD), biological

                                          140

-------
oxygen demand (BOD5), ammonia nitrogen, volatile phenol, total As, Pb and Cr6+), selected from
water quality data gathered from 1985 to 1989 (Table 2). The results indicate that the water quality in
the Yangtze river is at an acceptable level while the other water bodies within the region are polluted
to varying degrees.  The most serious indicators are COD, BODS, ammonia nitrogen, and volatile
phenol.

ECONOMIC ANALYSIS  OF WATER POLLUTION

        Water pollution has not only restrained further development of the local economy, but has also
threatened the beauty of the region and therefore the associated tourism.  The highest density of
factories in the urban district is found along the harbour  and the Old Grand Canal. Of these, the
harbour has become the most important industrial water source.  However, deterioration of the Harbor
water has resulted in unacceptable water quality for industrial use, and the consequent water shortage
has both decreased industry production (for example, the Dadon Paper Mill showed a loss in profit of
5 million RMB in 1989) and increased factory demand upon the municipal tap water supply.  In 1989,
the total municipal water supply was 150 000 m3 per day of which the industrial demand accounted
for 80 000 m3. The demand upon the municipal water supply by further industrial development in the
area combined with the  decreasing water quality in the Harbour will  likely lead to a shortage of
drinking water in the city. The negative impact of water pollution has already been felt by the tourism
industry of the Zhenjiang area, and if no steps are  taken  to improve the water quality, the long term
viability of the tourist trade may be  seriously jeopardized.
                             F—icmlc w«Uf


                               U*«ltl«>I*»l"
                                          ZHENJIANG rX\
Note: F, G respectively stand for function of water body and water quality goal
           Figure 1. Map of functional regionalization of main water bodies in Zhenjiang.

                                             141

-------
     Table 1.  Characteristics of wastewater discharged from
                 the urban district of Zhenjiang.
Receiving Water
Harbour
Old Grand Canal
Yuenliang River
Wastewater
Discharge
97,800
84,600
20,300
SS
56.77


COD
75.57
29.26
8.17
Phenol
0.08
0.01
0.05
Cyanide
0.06
0.01

Oil
1.12

0.05
As

0.03

Pb

0.01

Ammonia
1.31


Table 2.  Comprehensive assessment of water quality in Zhenjiang.
Cross-section
Dry
Period
Normal
Period
Wet
Period
1985
1986
1987
1988
1989
1985
1986
1987
1988
1989
1985
1986
1987
1988
1989
Yangtze River
1
+
+
+
+
+
H-
+
+
+
+
+
+
+
+
+
2
+
-A
-A
-A
+
+
+
+
-A
+
+
+
+
+
+
3
+
+
+
+
+
+
+
+
+
H-
+
+
+
+
+
Old Grand Canal
4
-B,C,N,A
-B,C,N,A
-A,N
+
-A
-A
-A
+
+
-A,N
+
+
+
+
+
5
-B,C,A,N
-B.C.A.N
-A
-B,A,N
-B,A,N
+
-A,N
+
•t-
-A,N
+
+
+
-A
-A,N
Harbour
6
-B,C,A,N
-B,C,A,C
-B,A,N
-B,A,N
-B,C,A,N
-B,C,A,N
-B,A,N
-B,A,N
-B,A,N
-B,N
+
+
-N
-A
-N
7
-B,C,A,N
-B,C,A,N
-B,C,A,N
-B,C,A,N
-B,C,A,N
-B,C,A,N
-B,C,A,N
-B,A,N
-B,A,N
-B,A,N
-B,A,N
-N
-B,A,N
-B,A,N
-N
8
-B,A,N
-B,C,A,N
-B,A,N
-B,A,N
-B,A,N
-N
-B,A,N
-B,A,N
-B,A,N
-B,A,N
-A,N
-N
-B,A,N
-A,N
+
Yueliang River
9



-B,N
-B,C,N



-N
+



+
+
10



-B,N
-B,N



-N
+



+
+
                           142

-------
                    PRIMARY FEASIBILITY ANALYSIS OF INTEGRATED
                     WATER POLLUTION TREATMENT IN ZHENJIANG

FUNCTIONAL REGIONALIZATION OF MAIN WATER BODIES

        Classification by water use of the main water bodies in the Zhenjiang region is illustrated in
Figure 3-1.

WATER QUALITY GOALS

        The short-term water quality goals for the main water bodies within the Zhenjiang region, are
for the harbour and those portions of the Yangtze river and Great Grand Canal within the Zhenjiang
district to reach the 2nd, 3rd, and 4th levels of the national surface water quality  standards
(GB3838-88) by the year 2000.
       expansion  of municipal  water  supply  capacity by

       traniralttlng clean water from distant water  bodlcx
                           municipal water  supply capacity

                           oeed lo  be  further    expanded
     urjjancy of water pollution treatment

     falls+.Investment on treatment lack*
                                                                 deterioration of water

                                                                pollution more  serious
u	
                           more factories are forced to

                            use   municipal  tape  water
       treatment of water pollution

       In  nearby  water   budles
      llnrertmeDt I
Improvement  of

 water   quality
                                                         1 1
   factories directly use water In  nearby

   water bodies  as   Industrial   water
         :                     '
                                                          occupied percentage of Industrial water

                                                           In  municipal   water  supply falls
               costs on water supply falls + get more economic

                benefits by converting wsntewater Into resources
   occupied percentage of domestic

   water   falls
                                             Figure 2.


                                              143

-------
ANALYSIS OF THE PROJECT DESIGN SCHEME
        Reliance of Zhenjiang's industrial core upon the harbour as a primary water source means that
the harbour is strongly linked to the economic health of the Zhenjiang region. In recent years,
continued deterioration of the harbour water has created a shortage of water for industrial  needs.
There are two possible solutions to address the shortage of water.  One is to expand the municipal
water supply capacity by diverting water to the system from the Yangtze river.  The other solution is
to treat industrial waste water and allow the water quality of the Harbour to recover to a level where it
would again be suitable for industrial use.  The two paths lead to  very different situations, and the
resulting event flow diagrams are  shown in Figure 2.

        The flow diagrams indicate that treatment of the waste water will ultimately provide the better
solution, both economically and environmentally.  This has in fact been established in other urban
areas of southern China, such as Shanghai, where the total costs of municipal water supply expansion
have generally exceeded that of methods undertaken to improve the water quality.
Just to the north of Zhenjiang, 2 km from the urban core, are two large deltas referred to as the
Zhenrun and Jiaobei Shoals. These deltas occupy a flat area of 13 km2, the soil is soft floodland
sediment, and they are currently undeveloped.  The  deltas may be a possible location for the
establishment of low-cost oxidation ponds and land  treatment systems to treat the industrial pollution
from Zhenjiang.  Part of the pollution problem in the Harbour is a lack of adequate water mixing or
exchange from the Harbour back out to1 the Yangtze river.  By pumping the sewage to an area closer
to the Yangtze river and treating it before discharge, water pollution should be effectively controlled.
            Eculuaical EnEtnceriniis

           • uC Zhenrun Shual
 Setraze Intercepting

 system of Yuenlians

 Rlrcr
                                         Sewage Intercepting

                                         system of Habor
                        Sewaqe Intercepting:

                        system of Old Grand

                        Canal
\ Integra led Treatment

; Enislneerlns  of Old

 Grand Canal
                                                   The engineerings of

                                                   transmission    of

                                                   conling water  frnrn

                                            ^       putror  ataliun   to

                                                   Old Grand Canal

Figure 3. Integrated treatment project of water pollution in Zhenjiang.


                            144

-------
OUTLINE OF PRIMARY DESIGN SCHEME

       Three municipal sewage intercepting systems will be constructed along each of the three main
water bodies in Zhenjiang; the Harbour, the Old Grand Canal, and the Yueliang river, all of which
currently collect municipal waste water (Fig. 3). The Jiangbin pumping station will direct sewage onto
the Zhenrun and Jiaobei Shoals through pipelines crossing over the harbour. Thirteen km2 of the
deltas will be developed into waste water treatment facilities where the waste water will  be treated and
diverted as a supply for other industries (such as agriculture, fisheries, poultry, and forestry), before
being discharged into the Yangtze river on the condition that water quality standards are met.
                    MUNICIPAL WASTE WATER TREATMENT PLAN
OUTLINE
       The water pollution treatment plan for Zhenjiang is comprised of two components. The
primary system is the treatment facilities for general treatment of city waste water as a means to
improve water quality and provide a water source for industries.  An auxiliary system will be provided
to improve the water quality of the Old Grand Canal as a means of enhancing the tourism potential of
the area.  Only the primary treatment plan is discussed below.  Table 3 lists the major components
included in the design of the water treatment facilities.  The complete project has been divided into
three stages of construction as outlined in Table 4.

FIRST STAGE OF DEVELOPMENT

       The treatment facility is being designed to handle a daily flow of 124 500 m3 of waste water,
of which the industrial contribution is 91 000 m3 and the domestic portion 33 500 m3. The primary
goals of the first stage of development are to establish the  sewage interception system in the harbour,
and intermediate-testing experimental oxidation ponds on the Zhenrun Shoal.

(1)  Sewage Interception System of the Harbour.

       An overview of the sewage interception system for first stage development is illustrated in
Fig. 4.  The total land area to be serviced is 449.25 hectares with a population of 221 800. Systematic
analysis suggested that the interception ratio of the system would be optimal at 3.  The  1990
estimation of the total cost of this stage was 46 722 900 RMB.

(2)  Intermediate-Testing Experimental Oxidation Ponds.

       As previously discussed, the Zhenrun Shoal is located to the north of Zhenjiang city between
the Harbour and the Yangtze river. This delta is composed of floodland sediment and covers an area
of more than 2 000 hectares.  The main soil types on the delta are sand clay and slime sand clay.  The
soil texture is very soft and the land is flat, making it readily suitable as a site for the design,
construction, and maintenance of biological oxidation ponds and other treatment facilities.

       In June of 1992, the North China Civil Engineering Design Institute made a primary design of
the experimental  oxidation ponds on Zhenrun Shoal based  on the report "Primary design scheme of
first-stage  project of ecological engineerings for the conversion of wastewater into resources in
Zhenjiang", and the results of an environmental impact assessment of the area. The location chosen
occupies a 10 hectare area in the northeastern corner of the Zhenrun Shoal (Fig. 3).
                                           145

-------
        Once in operation, the facility is designed to treat 5 000 m3 of waste water per day.  The
predicted intake water levels of BODS at 30 mg/1 and SS at 60 mg/1 will be treated to give output
levels of BODS at 200-250 mg/1 and SS at 200-300 mg/1.  A flow chart of the treatment processing
steps, a table outlining some of projected operating data, and a schematic of the layout of the
collecting ponds are all given in Figure 5.  The total investment into Hie project was estimated in 1992
as 3 470 000 RMB.

              Table 3. The major components of tfie principal sewage treatment facilities
                                       planned for Zhenjiang city.
            Sewage Intercepting
             Engineerings (SIE)
                                Engineering of Sewage
                                   Pipeline Crossing
                                    Over Harbour
                                                     Sewage Treatment and
                                                    Utilization Engineerings
  • SIE in south bank of Harbour
  • SIE in both banks of Old Grand Canal
  • SIE in both banks of Yueliang River
  • Engineerings of remaking municipal
      water discharge system
                              • Depositing reservoir

                              • Jiangbin Pumping Station

                              • Sewage mainlines
                                crossing over Harbour
                                             • Depositing tanks
                                             • North and south man-made banks of
                                                Zhenrun Shoal
                                             • Intermediate-testing experimental
                                                oxidation ponds
                                             • Pumping stations of Zhenran Shoal
                                                and Jiaobei Shoal
                                             • Ecological farms of Zhenrun Shoal
                                                and Jiaobei Shoal
                                             • Scientific Experimental Base for
                                                ecological engineerings
              Table 4.  The three stages of construction of the principal sewage treatment
                                       facilities for Zhenjiang city.
   Stage
 Duration
             Engineerings
            Expected Benefits
    First
    Stage
1993-1995
• SIE of Harbour
• Transmission engineerings of acidic
   wastewater
• Jiangbin Pumping station
• Main Pipeline over Harbour
• Zhenrun Shoal Pumping station
• Experimental Oxidation ponds
• Obvious improvement of water quality in
   Harbour
• Pollution load reduction: 50%
• Industrial water shortage of factories along
   Harbour can be almost solved
   Second
   Stage
1994-1996
• Immediate testing base of ecological
   engineerings
• Ecological farms on Zhenrun Shoal
• Expansion) Engineering of Zhenrun
   Shoal Pumping station
• SIE of Old Grand Canal
• Water quality of Harbour reaches 3rd level
• Pollution load reduction: 80%
• Economic benefits: 500.000RMB in 1996
   Third
   Stage
1995-1997
• SIE of Yueliang River
• Construction of ecological farms
   completed
• developing Jiaobei shoaly land
• Engineerings of transmitting water from
   Power Station to Old Grand Canal
• Water quality of Yueliang River reaches
   5th standard
• Pollution load reduction: 99%
• Economic benefits: 1 million RMB per year
                                                 146

-------
                                                  Shoal Pumping Slujun
    .  Y:in;«tri! River
      ,  \\

                     Zhenrua Shnal

IN'T-fl,r.<,59. ii.-.
I -t.iiijcitv2)9 i
                                                                            SN-. i;     I  V—^£.i
                                                                      \   ..^^w«==T-
                                                                                          anil now:l/«  un|t po^h^
                                                                                                     pantlf

                                                                                               ua paadi

                                                                                               « mtcrcepttnf mufnMnc

                                                                                               liber rdnforced  ptjjite ptpct

                                                                                                 drain
                                                        i\=^-^x^.
fJOTE:I-MMrilT.INr-no» topcclirclr >l»j for Ja.lallcJ flo. and :^V===rP,====^Xl"^
                   ..._                                ^^?^^'       \\ 1\
      Inlerreilllil)! fluw;WET-nui»tDRY-flui» rcvprcllv I     I        \

      naif la 
-------
flume flow
meter
•*
laiikK

»
anaerobic
ponds
-»
L faculta-
tive punds
-1
II faculta-
Itvc ponds
f
oxidation
ponds
- _J
WBslcwater
from Jiang
Bin Pumping
Staltcn
     * technical parameters  *
diiicharee


tyoc



anaerobic

ponds
facultative
ponds
oxidation

ponds
DOD5
inlet

water

(tng/L)
200

-250
120


60

outlet

water


-------
ENVIRONMENTAL IMPACT ASSESSMENTFOR THE ZHENJIANG HARBOUR

       The Environmental Science Department of the Nanjing University in China and the Athens
National Environmental Research Laboratory in the U.S.A. worked jointly on a research program in
which the DYNHYD4 hydrodynamic and the WASP4 entrophication models were used to develop a
water quality model of the Zhenjiang harbour.  The results indicate that completion of the first-stage of
the project will permit the water quality of the  harbour to reach or even surpass the 3rd level of the
national surface water quality standard (GB3838-88).  The model also predicts that the water quality
will reach the desired standard within the first month of operation of the first-stage facilities.

ENVIRONMENTAL IMPACT ASSESSMENT FOR THE YANGTZE RIVER

       Fully treated water from the oxidation ponds and other land treatment systems will be
discharged into the Yangtze river at a zone near some man-made banks known as the Xisa banks.  The
distance from the discharge  site to Dantu city, the nearest developed center downstream, is 9.8 km. It
is assumed that the high volume of the Yangtze will dilute and further biodegrade the treated waste
water, which will further reduce any negative impact on downstream water bodies.

ENVIRONMENTAL IMPACT ASSESSMENT FOR THE RESIDENTIAL
SCENIC AREAS OF ZHENJIANG

       The oxidation ponds on the Zhenrun Shoal will be 2 200 m from Jiao Mountain, an important
natural resource,  and 2 550 m from the Jiangbi residential  area.   During most of the year, the
prevailing wind is from the southeast and such nearby scenic and residential areas are upwind from the
water treatment site.  During the winter, however, the prevailing wind is from the northwest and could
carry odours from the oxidation ponds to these areas.  An  attempt is being made to reduce this effect
by adjusting the layout and design of the oxidation ponds,  establishing green belts, or using other
technological methods.

TREATMENT OF TOXIC WASTES

       In order to ensure that the treatment facility is able to maintain an acceptable quality of
discharged treated water, it will be necessary to minimize the toxic pollutants entering the system.  To
do this requires integrated utilization of new management methods and technologies. This is viewed
as one of the major difficulties facing the successful operation of the project. Ultimately, the project
will include measures to ensure that the discharge of toxic wastes into the sewage collection system is
only done when the  waste meets a certain standard. This may require pre-treatment facilities in some
factories.  Part of the long-term environmental management of the city must include conditions that
will prevent the establishment of new factories  that produce unacceptable toxic waste, and measures
that can force the closure of those factories unable to meet the minimum standard requested.

LATER STAGE DEVELOPMENT PROSPECTS

       The oxidation pond treatment system on the Zehrun Shoal is intended to serve only as a
transitional method of water treatment.  The ultimate goal of the project is to establish a large-scale
facility that will not  only be able to treat waste water, but will also be efficient enough to allow for
the  recovery and  reuse of treated water. For this  later stage of development, the Zhenrun and Jiaobei
Shoals are able to offer several site advantages, including:
                                           149

-------
        (1)     Ample space for expansion - 13 km2, equal to half of the existing urban district.
        (2)     Fertile topsoil - subclay/sand, rich in organic materials and suitable for vegetation.
        (3)     A warm and damp climate - average annual air temperature of 15.4°C with a 238 day
               frost-free period.
        (4)     Close proximity to Hie industrialized area - permits inexpensive transport of sewage.
        (5)     Flat topography.
        (6)     Close proximity to the Yangtze river - allows water from the Yangtze to be used to
               adjust waste water treatment and discharge concentrations.

        The Environmental Science department of Nanjiang University will construction an
experimental field station on the Zhenrun Shoal.  In addition, a major national research laboratory
entitled "Water Pollution Control and Conversion of Wastewater into Resources" has been recently
established within the Environmental Science department.  This new laboratory will allow for the
study and development of models relating to waste water treatment. These developments together with
technological assistance from the World Bank,  should ensure that the final stages of the project will be
realized.

                                          SUMMARY

        The outlined plan was developed with preliminary scientific investigation, and consideration of
the available technology and the rich land resources available in Zhenjiang. The economic,
environmental, and  social benefits derived from the project for Zhenjiang include:
        (1)  The majority of domestic and industrial waste water of the urban district of Zhenjiang will
be treated and reused, effectively controlling the water pollution problem.
        (2)  The factories near the harbour and adjacent rivers will once again be able to use those
water bodies for source water, solving the industrial water supply problems. The costs associated with
municipal water supply expansion can thus be avoided.
        (3)  The Zhenrun and Jiaobei Shoals will be developed for light industry such as rice fields
and fisheries.

        Finally, and importantly, the facility will serve not only as a demonstration project, but also as
a major field experimental base for the study and development of low-cost waste water treatment and
post-treatment water uses for other regions of southern China.

                                        REFERENCES

Environmental science  department of Nanjing University "Primary feasibility analysis of water
        pollution treatment scheme and project" 1990/11.
North China Civil Design Institute "Primary design scheme of intermediate-testing experimental
        oxidation ponds on Zhenrun Shoal of Zhenjiang" 1992/6.
                                            150

-------
WATER QUALITY MODELING OF TOXIC ORGANIC CHEMICALS IN
              THE YANGTZE RIVER AND ITS APPLICATIONS

                                       .      by

                         Ouyong Xu, Guangyao Sheng and Haixan Zou1

                                        ABSTRACT

       The maximum permissive loads of five organic pollutants (phenol, nitrobenzene,
chlorobenzene, bis-2-chloroethyl ether, and di-n-butyl phthalate), were calculated in sewage effluence
being discharged into the Yangtze River at a site up-stream from a newly constructed water pumping
station.  A two-dimensional, non-steady state, differential equation was used to simulate chemical flow
patterns within the river in order to account for width and tidal variations.  The first-order attenuation
coefficient (K) in the differential equation represents the  sum of the pseudo first-order rate constants of
environmental processes such as volatilization, direct photolysis, hydrolysis, oxidation, and microbial
transformation.  Rate constants were computed using the Exposure Analysis Modeling System
(EXAMS) developed by Burns et al. (1982), and also from corresponding parameters measured in the
field. The results indicate that attenuation of the organic compounds is significant compared  with the
transport process provided that the coefficient of attenuation exceeds  10-2 /hr.

                                     INTRODUCTION

       The site of sewage discharge from Changzhou city into the Yangtze River is located
approximately 8.5 km up-stream of the Ligang Water Works pumping station (currently under
construction). The water source of the new pumping station will be the Yangtze River.  The close
proximity of the sewage discharge to the pumping station raises questions about the impact of organic
pollutants from sewage effluents upon the water quality at the Ligang pumping station, and suggests a
need for assessment of the maximum permissible load of these organic pollutants.  The maximum
permissive loads of five organic pollutants (phenol, nitrobenzene, chlorobenzene, bis-2-chloroethyl
ether, and di-n-butyl phthalate) were calculated using a two-dimensional non-steady state mass transfer
equation and the mixed coefficients for the Changzhou section of the Yangtze River (Institute of
Environmental Water Conservancy, He Hai University, unpublished), and the Exposure Analysis
Modeling System (EXAMS) (Burns et al. 1982).  The  effects of changes in the magnitude of the
first-order attenuation coefficients of organic compounds  upon their maximum concentration levels
within the source water at Ligang  are discussed.

                               MODEL AND PARAMETERS

       The section of the Yangtze River under study,  from the sewage discharge site at the mouth of
the Zaojiang stream to the pumping station at Ligang, is wide and subject to tidal bores occurring
twice daily.  The direction of water flow changes during  the dry season, but during the rainy  season,
these tidal effects are reduced and only flow velocity is affected.  Toxic organic compounds
discharged into the Yangtze River are both diluted and attenuated, resulting in a cline in the
    'Department of Environmental Science, Nanjing University, Nanjing, China
                                          151

-------
 concentration of pollutants over the study area.  The distribution of these pollutants can be simulated
 using a two-dimensional dynamic equation that includes the effects of flood and ebb tides:

 8 (HC) + 8 (UHC) + 8 (VHC) = 8 (HDx SC) + 8 (HDy 8C) + H(S-KC)
 8t     8x     8y    8x     8x            Sy     8y

 where:
 t = thne
 H = depth (m)
 x and y = plane coordinates where x is positive in the direction of water flow, and y is perpendicular
 to x and positive from right to left
 U, V = average water velocities (m3/sec) in X and Y planes respectively; their magnitude and
 direction are variable with the tide.
 C = concentrations of toxic organic compounds (ppb)
 Dx,  Dy = mixing coefficients in the X and Y planes (m2/sec)
 S = load of toxic organic compounds (g/m3.sec)
 K = first order attenuation coefficient of the toxic organic compounds (hr-1)

        The equation shown was solved to provide a numerical solution by splitting the equation using
 the control volume method and then solving by the implicit difference approach.  A rectangular grid
 was imposed over the study area of the river using 200 m divisions in the Y-axis, while the X-axis
 was divided unequally into 200, 400 and 600 m divisions. A total time period of 8 days divided into
 30 minute time steps was used in the computation.  The mixed coefficients (diffusion and dispersion)
 in the X and Y planes were calculated by the following formulas which were the results of field study:

 Dx = 6.00 HU*
 Dy = 0.52 HU*,  where U* is the shear velocity.

       EXAMS (Burns et al., 1982a, b) is an essential model for studying the transport and dispersion
 of organic compounds in an aquatic environment. Using the physical and chemical properties of an
 organic compound combined with some aquatic environment parameters the first-order attenuation
 coefficient, K, can be computed for use in the two-dimensional dynamic equation given above.
 Theoretical and applied studies using EXAMS have been reported in the literature (Wolf et al., 1980;
 Zepp, 1980; Honeycutt and Ballantine, 1983). In a previous study, the transport and dispersion of
 nitrobenzene, bis-2-chloroethyl ether, quinoline, and phenol were examined in the Changzhou  section
 of the Great Canal, which is connected to the Yangtze River and within 20 km of the current study
 area.  Using the data collected from the Great Canal on nitrobenzene and bis-2-chloroethyl ether, we
 were able to verify the reliability of EXAMS as well as the precision of the parameters selected (Xu
 and Sheng, 1987). Therefore, it seems reasonable to expect that the application of EXAMS to the
 current study section of the Yangtze River will provide reliable, reproducible results. The first-order
 attenuation coefficient, K, and the rate constant were estimated from environmental parameters such as
 temperature, bacterial population, etc.

       The important kinetic processes to be considered when evaluating the behavior of organic
pollutants  in a water body are hydraulic transport, volatilization, direct photolysis, hydrolysis,
oxidation, and microbial transformation.  The main function of EXAMS is to calculate the pseudo
first-order rate constants of these processes.  The first-order attenuation coefficient in the differential
equation described above is calculated as the sum of the pseudo first-order rate constants of
volatilization (KV), direct photolysis (KP), chemical oxidation (KO), hydrolysis (KH), and microbial
transformation (KB).  The physico-chemical properties (from Mabey et al. 1981) and rate constants of

                                           152

-------
the five organic compounds examined are shown in Table 1.  The effect of season on the measured or
estimated environmental parameters for the section of the Yangtze river under study are given in
Table 2. The pseudo first-order rate constants (K) calculated by EXAMS of the five organic
compounds examined are listed by season in Table 3.  The calculated values of K show that phenol
had the fastest rate of degradation while di-n-butyl phtalate had the  slowest.

        Following completion of a new sewage pipe line, the load of organic pollutants discharged into
the Yangtze will increase. A review of the literature indicates that bacteria counts can be expected to
increase up to 20 times their current population at the discharge site and directly downstream.  The
increased pollutant load requires  modification of the bacterial degradation rate constants (KB) and the
corresponding first-order attenuation coefficients (K) for the compounds examined (Table 4).
   Table 1. Physico-chemical properties and process rate constants of selected organic compounds.
Compounds

mol.wt.
solubility (ppm)
sediment/water
distribution coef.
(mg/kg)/(mg/l)
Kerry constant
(atnvm3/mol)
vapour pressure (Torr)
acidic hydrolysis rate
constant (/h/M)
neutral hydrolysis rate
constant (/h)
alkaline hydrolysis
rate constant (/h/M)
oxidation rate
constant (/h/M)
Bis-2-chloro
ethyl ether
143
1.02xl04

13.9

1.3xlCr5

0.71

0.0000

4xlO'6

0.0000

24.0
Nitro-
benzene
123.11
1.9xl03

36

1.31xlO'5

0.15

0.0000

0.0000

0.0000

0.0000
Phenol

94.11
9.3xl03

14.2

4.54xlO'7

0.341

0.0000

0.0000

0.0000

l.OxlO'7
Chloro-
benzene
112.56
4.88xl02

330

3.58xlO-3

11.7

0.0000

0.0000

0.0000

0.0000
Di-n-butyl
phthalate
278.3
13

1.7xlOs

4.50X10-6

1.6X10"4

: 7.92xlO'3

0.0000

79.2

1-4
microbial transformation
rate constant
(mg/cell/h)
3xlO'9

3xlO'9

3xlO'6

3xlO'9

3xlO'8

                                             153

-------
Table 2. Some environmental parameters of the Yangtze River study site
                       in three different seasons.
                       Dry
Intermittent
                   Rainy

wind velocity

near surface (m/s) 3.5




water temperature
oxidant
concentration (M)
bacterial cell
(°C) 15
Ix

200

3.5
20
10-9 1 x 10-9

2000

3.5
25
1 x 10-9

4000






population (cells/ml)





aeration
coefficient (cm/h)
suspended solids
(mg/L)
PH
3.80

100

8.0
3.80

300

8.0
3.80

600

8.0





latitude 31.77'; average cloud cover (%) 69.0

Table
3. Pseudo first-order rate constants
attenuation coefficients of selected organic

Compounds

chlorobenzene
nitrobenzene
bix-2-chloro
ethyl ether
phenol
di-n-butyl
phthalate

chlorobenzene
nitrobenzene
bix-2-chloro
ethyl ether
phenol
di-n-butyl
phthalate

chlorobenzene
nitrobenzene
bix-2-chloro
ethyl ether
phenol
di-n-butyl
phthalate
KP


0
0

0
0

0

0
0

0
0

0

0
0

0
0

0
KH


0
0

3.99x10-*
0

4.4X10-6

0
0

3.98xlO's
0

1.52xlO's

0
0

3.97x10-*
0

7.69xlO'7
KO KS

dry season
0 5.87xlO-7
0 e.OOxlO'7

0 5.99xlO-7
0 5.99x10-"

0 . 3.3xlO-7
medium period
0 5.46xlO's
0 5.94xlO-6

0 .,, 5.98xlO'6
0 5.97xlO-3

0 LlSxlO'6
raining season
0 l.OOxlO'5
0 1.17xlO'5

0 1.19X10'5
0 1.19xlO-2

0 1.17X10-6
(hr-1) and
compounds.
Kv


8.79x10-"
3.79x10-"

3.51x10-"
2.55xlO'5

6.77xlO;6

7.34x10-"
3.54x10-"

3.29x10-"
2.50xlO'5

2.25X10'6

5.99x10-"
3.28x10""

3.07x10""
2.44xlQ-5

1.09x10-*


K


8.80x10-"
3.80x10-"

3.56x10-"
6.25x10-"

USxlO'5

7.40x10-"
3.60x10-"

3.39x10""
6.00xlO-3

4.93X10'6

6.09x10-"
3.40x10-"

3.23x10""
1.19xlO'2

3.03xlO-6
                              154

-------
              Table 4.  Modified microbial degradation rate constants and first-order
                         attenuation coefficients of five organic toxins.
Rate Constants


chlorobenzene
nitrobenzene
bix-2-chloro-
ethyl ether
phenol
di-n-butyl
phthalate

dry
season
1.17xlO-5
1.20xlQ-5

1.20xlO'5
1.20xlO-2

6.66xlO'6
KS
intermittent
season
1.09xlO'4
1.19X10-4

1.20X10-4
0.119

2.30xlO'5

rainy
season
2.00X10-4
2.34X10"4

2.38x10^
0.238

2.34xlO-5
K
dry
season
8.91X10-4
3.91X1Q-4

3.67x10^
1.20xlO-2

USxlO'5

intermittent
season
8.43X10"4
4.73X1Q-4

4.53X10-4
0.119

2.68xlO'5

rainy
season
7.99X10"1
5.62x10-"

5.49X10"4
0.238

2-53X10'5
     Table 5. Maximum permissive concentrations of selected compounds in source water (ppb).
compounds
max. permissive
dissolved cone.
total
max.
permissive
cone.
dry
season
intermittent
season
rainy
season
chloro-
benzene
20
20.007
20.020
20.040
nitro-
benzene
20
30.001
30.003
30.006
bis-2-chloro
ethyl ether
0.3
0.300
0.300
0.300
phenol
1.0
1.000
1.000
1.000
di-n-butyl
phthalate
200
234
302
404
                                RESULTS AND DISCUSSION

       Before the permissive loads of the organic toxins selected can be calculated, several conditions
need to be defined.

       (1)  Flow rates of 8000, 50000, and 30000 m3/sec were used to represent river flow during the
dry, rainy, and intermediate seasons, respectively.  These flow rates were obtained from data from the
Institute of Environmental Conservancy, He Hai University.

       (2)  The submerged discharge pipe (200 m in length) at the mouth of the Zaojiang stream is to
be installed at a depth of 26 m below the surface of the river.

       (3)  The organic compounds studied are assumed to remain bound to suspended solids and not
to redissolve, and therefore will be removed in the water treatment processes at the Ligang  Water
                                           155

-------
 Works pump station. Therefore, the maximum permissive concentrations of chemicals in drinking
 water (Sitting,  1985; Verschueren, 1983) may be treated as the maximum permissive dissolved
 concentrations  in source water at the intake of the pump station (as shown in Table 5). Based on
 estimates of the seasonal concentrations of suspended solids, the formula for the derivation of the total
 maximum permissive concentration of an organic compound at the Ligang station is as follows;

 Cd=    1    C
     1 + Kp[p]
 where
 Cd = the maximum permissive dissolved concentration (ppb)
 C = the total maximum permissive concentration (ppb)
 [p] = the concentration of suspended solids (ppm)
 Kp = the sediment/water partition coefficient of an organic compound.

        (4) Only the organic toxicants released from the mouth of the Zaojiang stream into the
 Yangtze river are considered in this study, and other sources of pollutants further upstream are
 ignored.

        As previously indicated,  the change in the direction of the flow of the Yangtze River from
 tidal effects results in an accumulation of pollutants and greatly complicates the concentration cline of
 the pollutants.  At the end of ebb tide, the river flow  velocity approaches a minimum and a cloud of
 highly concentrated sewage forms about the discharge site. During flood tide, the cloud of
 concentrated sewage along with sewage from further downstream is carried upstream above the site of
 discharge.  At the cessation of flood tide, the river flow velocity again approaches a minimum and a
 cloud of concentrated sewage reforms. When ebb tide begins again,  sewage being discharged
 combines with  sewage being carried back downstream to form yet another type of concentration
 pattern.  Tidal changes cause a dynamic pollutant concentration distribution that results in variability
 of the pollutant concentration at the pump station in Ligang.  In order to protect the quality of
 drinking water the maximum concentration of a given organic pollutant at Ligang should not exceed
 the maximum permissive concentration calculated for that pollutant.

        The maximum permissive pollutant concentrations were calculated for the three different river
 flow rates corresponding to the three different seasons identified.  The maximum permissive
 concentrations were also calculated for the years 2000, 2010, and 2020 during which the sewage
 discharge volumes are projected to be 320, 500, and 700 thousand tons per day.  The results indicate
 that the maximum permissive loads are lowest during the dry season  due to the minimal flow at this
 time. Therefore, during this season the discharge of organic toxins into the river must be closely
 monitored.

        In order to compare the relative importance of transportation  and degradation of organic
 compounds, the maximum permissive loads of "conservative" organic toxins can be calculated if it is
 assumed that the first-order attenuation coefficient of each organic compound  is equal to zero.
 Comparison of the ratios of permissive loads of "degrading" organic compounds to "conservative"
 compounds indicates that only the ratios  for phenol are quite high (1.26, 2.52, and 3.04) and hence,
 only for phenol is the degradation process significant compared to that of transport. Low ratios for the
 other four compounds suggest that the degradation process is  of little importance  and can be ignored.
Therefore, these compounds should be treated as "conservative" ones.

        As  previously discussed , tidal effects can cause the concentrations of organic compounds at
the pump station in Ligang to vary considerably.  The ratios of minimum  to maximum concentrations

                                            156

-------
for each of the organic pollutants examined is given in Table 6.  Further calculations indicate that
during the dry season (8000 m3/sec flow rate) the maximum concentration of organic pollutants at the
Ligang pump station occurs 1.5 hr after the beginning of the highest water velocity. During the
intermediate season (flow rate of 30000 m3/sec), the highest pollution concentration at the Ligang
pump station occurs  1 hr before the beginning of the highest water velocity, while during the rainy
season (50000 m3/sec flow rate) the highest concentration is reached 2.5 hr before high water velocity.
In the event that the concentration of organic pollutants at these times exceeds the maximum
permissive concentration, the pumping station should be stopped for 1-2 hr to avoid contamination of
the water supply.

       When the first-order attenuation coefficient of an organic pollutant becomes large enough, the
attenuation ratio (the ratio of concentration of a degrading compound to a conservative compound)
becomes larger than unity.  A sample set of attenuation ratios derived with a first-order attenuation
coefficient (K) of 0.1 hr-1 is given in Table 7 for different flow rates and contaminant loads.  The
results indicate that attenuation ratios are related to hydraulic factors and are independent of the
organic pollutant load. At a given flow rate and pollutant load, the attenuation ratio and pollutant
concentration at the pump station varies with the first-order attenuation coefficient.   This is illustrated
in Fig. 1 using  a flow rate of 8000 m3/sec and a pollutant load of 200 g/sec.  In this study, when the
first order attenuation coefficient, K, exceeds 0.01, the attenuation process becomes significant
compared with the transport process.  When K is smaller than 0.01, the attenuation process can be
ignored and the pollutant can be treated as a conservative compound.  The first-order attenuation
coefficient is a critical value and should be used to decide if attenuation is important in the
computation of models for river studies.

            Table 6.  Variations in the concentration of organic toxicants in the river water
                           at the Ligang Water Works pumping station.
        Compound
Flow (m3/sec)
Ratio of minimum
to maximum concentration
        phenol
        chlorobenzene
        nitrobenzene
        bis-2-chloroethyl ether
        di-n-tutyl phthalate
       8000
       0.631
       0.637
       0.637
       0.637
       0.637
        phenol
        chlorobenzene
        nitrobenzene
        bis-2-chloroethyl ether
        di-n-tutyl phthalate
       30000
       0.379
       0.508
       0.508
       0.508
       0.508
        phenol
        chlorobenzene
        nitrobenzene
        bis-2-chloroethyl ether
        di-n-tutyl phthalate
       50000
       0.302
       0.409
       0.409
       0.409
       0.409
                                             157

-------
            Table 7. Relationship between attenuation ratio, flow, and load.
Flow
(m3/sec)
8000
30000
50000
Load
(g/sec)
5.0
80.0
5.0
80.0
5.0
80.0
Concentration
of organic
compound
0.171
2.743
0.146
2.179
0.108
1.720
Concentration
of conservative
compound
0.772
12.348
0.302
4.830
0.178
2.855
Attenuation
ratio
'4.51
4.50
2.22
2.22
1.65
1.66
    34



    29



    214-
a
      -6
-5
-4
-3
-2
-1
                                 LogK (hr")
0
                                                                        ••10
                                                               D
                                                              01

                                                               c
                                                               o
                                                                              
-------
                                      REFERENCES

Burns, L.A., D.M. Cline, R.R. Lassiter. 1982. Exposure Analysis Modeling System.  (EXAMS): User
       manual and system documentation. EPA 600/3-023.
Burns, L.A., D.M. Cline, R.R. Lassiter. 1982. Exposure Analysis Modeling System.  (EXAMS): User
       manual and system documentation, (Chinese transl.).  Translated by Guangyao Sheng and
       Ouyong Xu. 1988.  China Environmental Science Press.
Honeycutt, R.C., L.G. Ballantine.  1983,. Mathematical modeling application to environmental risk
       assessments. In: Fate of Chemicals in the Environment, R.L. Swann and A. Exchenroeder
       (eds)  Am. Chem. Soc. Washington D.C. pp. 249-262.
Mabey, W.R., J.H. Smith, R.T. Podoll, et al.  1981.  Aquatic fate process data for organic priority
       pollutants. EPA-440/4-81-014.
Marshall S.  1985.  Handbook of Toxic and Hazardous Chemicals and Carcinogens, 2nd. Ed. Noyes
       Publications. Park Ridge, New Jersey USA.
Verschueren, K.  1983.  Handbook of Environmental Data on Organic Chemicals, 2nd Ed.  Van
       Nostrand Reinhold Co. 1310 pp.
Wolf, N.L., L.A. Burns, V.C. Steen. 1980. Use of linear free energy relationship and an evaluative
       model to assess the fate and transport of phthalate esters in the aquatic environment.
       Chemosphere 9: 393-402.
Xu, O., and G. Sheng.   1987. The transport and fate of some organic toxicants in Changzhou section
       of Great Canal - Application of EXAMS, (in Chinese) Environ. Chem. 6: 1-13.
Zepp, R.G.  1980. Assessing the photochemistry of organic pollutants in aquatic environments.  In:
       Dynamics, Exposure and Hazard Assessment of Toxic Chemical, R. Haque (ed). Ann Arbor
       Science Publishers, Ann Arbor, pp. 69-110.
                                           159'

-------
          A PLAN OF REGIONAL WATER POLLUTION CONTROL
                            IN RURAL AREAS OF CHINA

                                        Yongchun Zhang1


                                         ABSTRACT

        A plan for controlling regional water pollution caused by town and village industries (TVI's)
 in rural areas of China where stream networks intersect, was developed using the county of Wuxi as
 an example. A number of techniques and methods were used in the formation of the pollution control
 plan, including analysis of stream network structure, examination of stream flow and relative
 hydraulics, analysis of the relationship between waste discharge and water quality, and field and
 laboratory experiments for the derivation of appropriate model parameters.  This led to formation of a
 mathematical model for assessing stream network water quality and utilization of the model to
 calculate the total permissible levels of TVI waste discharge in the region.  The maximum allowable
 amount of pollutant discharge from each town or plant was determined using an optimized systematic
 analysis method based on a stratified structure.

                                      INTRODUCTION

        In the past decade,  the development of town  and village industries (TVI's) in China has
 increased dramatically, especially in coastal regions and the downstream regions of large rivers such as
 the Yangtze, Yellow, and Pearl rivers. The expansion of TVI's in some areas has been so rapid that
 industrial output has increased 88% per year. Outside of the more developed centers, TVI's tend to be
 scattered and small in scale, with limited funding, few trained technicians, and only crude, inefficient
 equipment.  The limited supply of electric power is scarcely enough for the TVI's, which means that
 little, if any, power is diverted to waste treatment.  Despite the low concentration of TVI's in rural
 areas, the relative output of pollutants is high.  The environmental  monitoring capacity in rural areas is
 very limited, and as a result, the discharge of untreated waste water into the water systems surrounding
 rural TVI's has led to serious water pollution problems.

        Located south of the Yangtze river at the Taihu catchment beside Taihu  Lake, Wuxi county is
 a typical district marked by a network of intersecting streams.  Within the county's 1235 km2 area
 there are more than 3000 streams of varying size and approximately 1 million people. Increasing
 industrial  output within Wuxi county has been an governmental priority that has  resulted in the
 establishment of 6000 TVI's within the county.  Correspondingly, water pollution has increased to the
 point where the waste discharge has reached 15 million tons annually. This has  led to serious
 problems that include a shortage of drinking water, the destruction of fish breeding habitat, a shortage
 of clean water for agriculture (and vegetable irrigation in particular), an increase  in the number of
 water pollution-linked accidents, and a lowering of the water table.  Wuxi county was therefore
 selected as an excellent case study site.

        The aims of this study were to provide a water pollution control plan for the Wuxi county
government, and to develop a systematic approach that could be used to form regional water pollution
controls in other areas of rural China.  The focus of the plan was to provide feasible and realistic
    Nanjing Institute of Environmental Science, NEPA, Nanjing, (210042), P.R. China

                                           160

-------
guidelines to limit waste discharge. By scientifically determining the maximum allowable levels of
waste discharge that still protected water quality, environmental standards could be met, TVTs could
operate and discharge pollutants at this maximum level, and financial investment in waste treatment
would be minimized.

                                         METHODS

       The pollution load of any given water body is the result of discharge from sources directly
along that water body, both within the county and further upstream, and also, due to the high level of
stream integration, from neighbouring water bodies. As a result, all pollution sources communicating
with the water body,  even at a distance, must be controlled.                                  ,

       The following factors were considered while producing this water pollution control plan:

(a) User demands
       - geographical distribution of users
       - water quality desired by users
       - water quantity desired by users
       - future users and their demands regarding water quality and quantity
       - variation of future water demands (quality and quantity)
       - basic and maximum demand levels (quality and quantity) at different periods of time  ;

(b) Pollution sources
       - geographical distribution of main sources of pollution
       - main categories of pollution
       - amounts of pollution from various  sources
       - capacity of  pollution control and waste water treatment
       - probable future status of pollution  sources
       - probable amount and categories of pollution in the future
       - probable control levels and treatment capacity

(c) Water resources
       - structure of the channels in a stream network
       - hydrological and aquatic conditions in streams
       - present water quality
       - probable unfavourable hydrological conditions in the future
       - probable water quality under the most unfavourable hydrological and aquatic conditions
       - required treatment of waste influxes to meet the desires of the users under different
               hydrological and aquatic conditions
       - sewage treatment needed to meet the basic requirements of users under the most
               unfavourable hydrological and aquatic conditions

       Ideally, the appropriate mathematical approach to evaluate the stream network pollution load
would be to take every stream within the network and every TVI pollution source, and treat it as one
big system using the  method of systematic optimum planning.  However, the complexity  of the
necessary calculations and the power of the  computer that would be needed precludes use of such an
approach in a rural situation.  The method used, therefore, was a simplification of the above  approach
and involved the following changes:
                                           161

-------
        (a)  The complicated stream network was simplified by considering only the principle streams
 within it. While the smaller streams were not considered directly, their role in pollutant transport was
 taken into account.
        (b)  Each town with a relatively high concentration of TVI's was considered to represent the
 principle pollution source for the immediate surrounding area.  It was assumed that waste water from
 different industries in the district was concentrated in the town.
        (c)  Since 90% of the total pollutant load could be attributed to 30% of all the TVI's, and
 since 36% of these TVIs discharged the most pollutants, only those industries  were considered.  In
 Wuxi county, and within the area of the Taihu catchment in general, the main pollution discharging
 industries are textile, chemical, construction materials, pulp and paper, beverage, metal processing,
 non-metal processing, food processing, and machinery industries. Therefore, only these plants were
 selected for pollution control, and so the complicated system was generalized into a relatively simple
 one.
        (d)  Streams were classified into stratification categories according  to their relative importance,
 position, and function within the stream network.
        (e)  Calculation of the allowable TVI waste discharge amount into  each stream depended upon
 the order of stream stratification.  The towns and associated nodes where streams intersected were
 considered as pollution control sections. In order to optimally assess the allowable waste discharge
 from each pollution source so that the permissible water pollution load was maximized, a mathematical
 model of water quality and systematic methods were used.

    The objective function (Z) of the systematic analysis was expressed as:
  Z - J) Lea  *  Qei - Max
      i-I
where:

Lei = the concentration of a pollutant type discharged at different pollution sources (mg/1)
Qei = the flow of waste water discharged from different pollution sources (1/min).

The limiting condition which ensured the water quality of each segment of the stream met the goal of
water resource protection was given by:
Lf1 * Bj * Lei
                           g - Le2
where:
    = the vector of water quality standards of each stream segment (mg/1)
Alt B! = the coefficient matrix related to the waste water discharge and the discharge
         of water flow in streams.
                                           162

-------
/I                     \
  I  1 0  	  0 0
  I  al 1	 0  0
  I           /
  I         /
  I  	  = Al
  I       /
  I     /
  1    /
  |  00  	  an 1
 \                      /

/                      \
  I  bl 0  	  00
  I  .0 b2	  0 0
  I          /
  I         /          ,
  I  	-	   =  Bl
  I       /
  I     /
  I    /
  j  0 0  	 0 bn
\ '                      /
 g = (gl, 0.	0)T


 gl = a, * L2



              -Q3i)
    =
  1
 bi =
Qli - Q3i + Q2i


	Qpj	
Qli - Q3i + Q2i
  at = eF
                   4Ex(Kl  + K3)/U2)
                                       163

-------
    After running the systematic analysis model, Hie optimal waste discharge amounts for each
 pollution control spot was calculated. For streams on the lower level of stratification, these calculated
 discharge amounts for the corresponding nodes were treated as the limiting factors.  By repeating the
 above calculations, the total limiting amount of waste discharge for the given area, and the maximum
 allowable pollutant discharge amount for each pollution source was calculated by systematic
 optimization.

                              DISCUSSION OF TECHNIQUES

     In order to conduct the optimal systematic analysis of water pollution control within a stream
 network, some preliminary calculations must be made.

 WATER QUALITY MODEL

    When modelling stream networks, the assessment of water quality includes models that consider
hydraulics, water quality, and conditions linking the nodes of a stream network. The hydraulics model
is composed of the so-called Saint-Venant equations, while the water quality model used the
convection-diffusion equation.  The model form was altered for different pollution and water current
conditions. Using the biochemical oxygen demand (BOD) and dissolved oxygen (DO) models me
equations are as follows:
  at    B ax
                          U:U:
  at
          ax
                  ax
                                _
                        «x
                                   ax
                                       - 2 Kio.AL. + s
                                         te-i
                                                                       i = il
 a(AuD) +  a(AUD)
   at         ax
                                 ax
                                                         Kio AX. + SI
 PI
 2 QJ141 = 0
il-l
Hjl.l
                                = Hjl,pl
-    . Vjl
                                         164

-------
  — . yjl =
Marginal condition, i = f
H (i,j) = hi (f,j)
L (ij) = 11 (fj)
D 0 j) = dl (f,j)

Initial condition, j = 0
H (i,0) = h (i)
U (i,0) = u 0)
L (i,0) = 1  (i)
D 0,0) = do  0)

Where,
Eq. () = so-called continuity equation
Eq. ( ) = momentum equation
Eq. () = equation of pollution conservation
Eq. ( ) = equation of DO conservation
B = average  width of cross sections, for shallow and wide channels.  It approximately equals the width
of the water  surface (m).
A = area of the cross sectional water flow (m2)
Q = water  discharge (m3/s)
c = coefficient expressing the roughness of the stream bed (Chezy coefficient)
d = hydraulic radius, for wide and shallow rivers, it usually approximates the average depth of the
water in the channel
Kio  = coefficients defining the decay rate of pollutants; for organic pollutants, these coefficients define
the oxidation rate, (Kl), settling/resuspension rate, (K3), etc.
s = other sources or sinks, including branch influxes, etc.
Pp = amount of coefficients
io = ordinal number
il = ordinal number of cross sections located in the areas of stream network nodes
jl = ordinal numbers of the nodes of a stream network
Vjl  = water  volume at node jl of the stream network
tiflFjl = net flux of pollutants at node jl
hl(f,j) = water level at marginal sections
ll(f,j) = concentration of pollutants in marginal sections
h(i)  = original water level of a section
u(i)  = original flow velocity in a section
10) = original concentration of pollutants in  a section
f = ordinal number of marginal sections
i = ordinal number of segments of the stream network
j = ordinal number of time steps
m = total number of time  steps
n = total number  of segments of the  stream network
dl(f j) = concentration  of dissolved oxygen in marginal sections
do(i) = original concentration of DO of a section
                                            165

-------
 Dji = net flux of dissolved oxygen at node jl
 si = other sources or sinks of dissolved oxygen
 K2 = coefficient defining the rate of re-aeration of water in the stream network

    When the marginal and original conditions are known, the model equations can be solved using
 the implicit difference method.

 PARAMETER ESTIMATION

 In order to be able to mathematically assess water quality variation in a stream network, the model
 parameters needed to be estimated from data collected in the field. The parameters which typically
 need to be estimated in a water quality model the longitudinal dispersion coefficients, the coefficient
 of river bed roughness, the pollution degradation rate, the net rate of air/water oxygen movement, and
 the rate of pollutant setfling/resuspension.

 (a) Dispersion Coefficient

 Several representative stream segments of the stream network were selected for a dye tracing
 experiment Dye was mixed homogenously in the river at one site, and then cross-sectional samples
 were taken at two or three downstream sites.  The longitudinal coefficient was estimated with the
 following formula:
              tl - t2
where;
ctj, otj are the standard deviation of the dye concentration distribution over time
tit tj represent the time of the dye cloud passing through the sampling sites

(b) Degradation Rate

Water samples were collected from different segments of the stream network.  Water samples were
maintained in the laboratory at a steady temperature and the BOD was analyzed daily for a 20 day
period. The degradation rate of organic pollutants was estimated using the minimum square regression
method:
 fl - 1 - e-k/t
      t * e'
                                           166

-------
          2fjy -
(c)  Multiple Parameters Estimation

    Using a systematic analysis method, several parameters could be estimated
simultaneously. The procedures was as follows:

    Water quality testing was performed at both up and downstream locations of different
stream segments within the network. .Optionally, several groups of parameters could be taken
from the respective valid range. The mathematical model was applied to each group of
parameters separately, and following comparison of the mathematically derived results with
the field observations, the data set that showed  the largest deviation was discarded.  A new
group of parameters was formed and the process repeated until finally a group of the most
ideal parameters was estimated.

    The following model equations were used to estimate multiple parameters simultaneously.
The objective funciton provides minimum deviation and is given  by:

Z - :: Lc - Lm ::2 -» min

where:    Lc, Lm are the calculated and monitored results, respectively.

    The following limiting conditions were used to ensure the parameters and the calcualted
results were within the valid range.

ai <- Ki <- bi

Ai <- Li 
-------
                    Table 1.  Model parameters estimated from field data.

                   Parameter               Unit            Symbol         Value
1.
2.
3.
4.
dispersion
degradation
settlement
re-aeration
m/s
1/d
1/d
1/d
Ex
Kl
K3
K2
33.35
0.26
0.51
0.75
                                      SUMMARY

    A regional water pollution control plan for Wuxi county was derived for the town and
 village industries and the associated rural stream network. The pollutant limits for 36 towns
 within the county were assessed.  Economically and technically, the plan is usable. By using
 the mathematical model of water quality, it is possible to protect water quality.  If th eplan is
 followed in Wuxi county, the water quality should recover to a level suitable for fisheries and
 agriculture, and some streams should be suitable as a source of drinking water.

    (a) A water pollution control plan is important to initiate in areas of stream networks and
 wehre TVI's are densely distributed. The crux of regional pollution control is to make clear
 the relationship between the pollutants discharged in different areas, by different mechanisms
 and in varying amounts, with the changes in water quality at different locations. The
 mathematical model of water quality and the systematic  analysis method are powerful tools in
 achieving an udnerstanding of this relationship.

    (b) The relatively complex association of parameters within the model can be simplified
 for practical use.  The  systematic optimum plan in the stratified structure is one such  mehtod
 of simplification.

    (c) Data collected in the field is important to maintain th eaccuracy of the model.
 Reliable and consistent field experimentation and sampling is  therefore an important
 component of the overall control plan.

    (d) The water current proved to be the determining  factor in assessing the concentration
 and distribution of pollutants within the stream network.  Changing flow characteristics,
 therefore,  are a potentially important mechanism towards improving water quality.  Further
research in this area is  needed.
                                        168

-------
  ADSORPTION OF NEUTRAL AND IONIC ORGANIC POLLUTANTS
                           ON SOILS AND SEDIMENTS

                                      John C. Westall1

                                     INTRODUCTION

    Sorption is one of the most fundamental processes in controlling the transport, fate, and
environmental effect of organic pollutants in aquatic environments.  Sorption processes directly
control the distribution of these compounds, and thereby indirectly affect transport, bioavailability,
and  degradability.  Whereas the sorption of neutral organic compounds has been  studied extensively
and is understood comparatively well for an environmental process, the sorption of ionic organic
compounds has not been researched  nearly as thoroughly.  In this paper we review the data and
models for sorption  of neutral compounds and summarize some of our recent work with ionic and
ionizable compounds. The literature cited simply provides  a key to more  detailed descriptions of the
data presented; no attempt has been made to provide a comprehensive review of the literature.

CLASSES OF ORGANIC COMPOUNDS

       For this discussion, all organic compounds will be assigned into three classes: neutral,
ionogenic, and ionic, as shown in Figure 1. The neutral compounds, as exemplified by the
chlorobenzenes and methylbenzenes, generally interact weakly with water and sorb strongly onto
environmental organic matter.  The predominant driving force for sorption is "the hydrophobic effect",
which arises from relatively weak interactions between the solute and the solvent, the solute and the
sorbent, and the solvent with itself.  The extent of sorption is relatively independent of solution
chemistry (e.g., pH and major ion composition).  Due to the weak interaction of  these compounds
with both the solvent and the sorbent, sorption is relatively easy to describe for a wide range of
compounds and a wide range of sorbents with relatively simple models.

       The ionic surfactants, as exemplified by linear alkylbenzene sulfonate (LAS) and
dodecylpyridinium (DP) possess both a hydrophobic group and an ionic group. The hydrophobic
group is expected to undergo the same weak, nonspecific  interactions  as observed for the neutral
organic compounds. However, the ionic groups interact much more strongly with both the solvent
and the sorbent; the negatively charged LAS is repelled by the predominantly negatively charged
environmental surfaces, while the positively charged DP is  attracted. The question with respect to the
sorption of these compounds is the extent to which both the hydrophobic and the ionic terms
contribute to sorption - does one predominate, or are they roughly equivalent and additive.

       The ionogenic compounds, as exemplified by the weak acid chlorophenols, and the weak base
methylanilines, exhibit both neutral and ionic forms in solution, depending on the pH. Like the ionic
compounds, both hydrophobic and electrostatic interactions are anticipated for the ionic forms, while
for the neutral form, the  ionic interaction is absent.  However, characterization of the sorption of
ionogenic compounds as a function of pH is generally difficult;  since many environmental surfaces
have pH dependent charge, it is difficult  to separate the effect of pH on the solute from  the effect of
pH on the sorbent.  In contrast, the charge of the ionic compounds is fixed and not pH  dependent.
    Department of Chemistry, Oregon State University, Corvallis, OR 97331-4003, U.S.A.

                                          169

-------
          CLASSES OF  ORGANIC COMPOUNDS
                          Neutral
                    qi            [o
                      Cln
              chlorobenzenes     aikylbenzenes
OH
x*x
O
      Acids
           0"
  ci
    n
ci,
chlorophenols
                                       Bases
                                               (CH3)n
                                  methylanilines
    Anions
    so
    LAS
                                      Cations
                               N-alkylpyrldinium
               Figure 1. Classes of organic compounds.


                          170

-------
       DISTRIBUTION  EXPERIMENT
suspension
 i ' -.'  • ,





  • ' ' (0)
            centrifuge
analysis
     -> C(w)
                                          C(s)
                ISOTHERM
 CP
_*:


~o
o
 (L
         C(w) (mol L')
           Figure 2. Experimental procedure.
                   171

-------
        The remainder of this discussion will be focused on the compounds in  Figure 1, grouped
 according to charge: (i) neutral, (ii) anions and anionogenic, (iii) cations and cationogenic.

 EXPERIMENTAL METHODS

        Essentially all of the data presented in this study  were obtained in simple batch experiments
 (1, 2), as shown in Figure 2.  A suspension containing the solvent, the sorbent, and Hie solute were
 mixed until  apparent equilibrium distribution was reached, the phases were separated by
 centrifugation, and the concentration of the target organic compound was  determined both in the
 solution and on the sorbent.  The data are generally plotted as isotherms: C(s), the amount on the
 sorbent per mass of sorbent (mol kg'1) vs C(w), the equilibrium concentration in solution (mol I/1).
 For linear isotherms, the slope is the distribution constant, Kj (L kg"1):
                        _C(s)
                      Kd=C(w)

 For the neutral compounds, linear isotherms are the rule, while for ionic  compounds, nonlinear
 isotherms are the rule. In the case of nonlinear sorption, the distribution is represented as the
 concentration-dependent  distribution ratio, D:
(1)
                         C(w)
                                                                                            (2)
               DISTRIBUTION OF NEUTRAL ORGANIC COMPOUNDS (3,4)

        It has been shown that the solution-sorbent distribution of the neutral  (particularly
nonhydrogenbonding) organic compounds can be explained very well  based on a property of the
compound and a property of the sorbent.  The  property of the compound is the octanol-water partition
constant, and the  property of the sorbent is the organic carbon content of the  sorbent, foc, the fraction
organic carbon:
       logKd=alogKow
(3)
The correlation between log Kj and log foc is illustrated in Figure 3 for the compound
1,4-dichlorobenzene and a collection of sorbents including lake sediments, river sediments, aquifer
materials, and activated sludge. The  correlation between log K^ and log K^ is illustrated in Figure 4
for the  methyl- and chloro- benzenes for sorption on an aquifer material with 0.25 %  organic carbon.

        The correlation expressed by Equation 3 is among the most widely used and widely useful in
the field of environmental organic chemistry. The reason for the success of this simple correlation is
the weak, nonspecific interaction of the neutral organic compounds with the solvent and the sorbent.
As the  specificity of the interaction of the solute with the solvent and the sorbent  increases (e.g.,
hydrogen bond formation, strong dipole interactions), the  success of the simple model decreases.
                                            172

-------
 OJ 1,4-DICHLOROBENZENE
 Y
 Cl
0.001
                0.01
                 «oc
  Figure 3. Kj vs property of sorbent,
   log Kp =0.73 log Kow-1.47
                       Data For
                     Substituted
                       Benzenes
2.5    3.0
 Figure 4.
 3.5     4.0

  log Kow

vs property of sorbent, Kow.


    173
                              4.5     5.0

-------
                        ORGANIC ANIONS AND ORGANIC ACIDS

ORGANIC ANIONS (5)

       The organic anions that are considered are homologs of 4-alkybenzenesulfonates containing  10,-
12, and 14 carbons in the alkyl chain, referred to as C-10, C-12, and C-14 LAS
(4-(l-methylnonyl)benzenesulfonate, 4-(l-methylundecyl)benzenesulfonate, and
4-(l-methyltridecyl)benzenesulfonate, Figure 5).  The questions posed about the  sorption of these
compounds are, (i) the linearity of the isotherm,  (ii) the  effect of pH, and (iii) the effect of the
number of -CH2- groups.
       The isotherms for adsorption of LAS on sorbent EPA-12 in 0.01  M monovalent  salt solution
are shown in Figure 6 on a log-log plot. The isotherms  are  nearly linear: the slope of log-log plots
for all three compounds is approximately 0.9.  (The slope would be 1.00 for linear isotherms on a
log-log  plot)  Furthermore it is seen that the curves for the homologs are displaced along the x-axis
approximately 0.4 log units per -CH2- group, as expected for  the hydrophobic effect.
       The effect of pH on the adsorption of LAS is shown in Figure 7. The change in distribution
ratio with pH was d log Dc/d log  [H+] ? 0.17.  (These data were obtained from the very low
concentration range of the isotherm  in which the isotherm was effectively linear, yielding a constant
value of Ka = D.)  This effect of pH may be due to (i) the influence of H+ as a surface potential
determining ion or (ii) specific pH-dependent  surface-complexation or ligand-exchange reactions (e.g.,
ligand exchange with -OH of =FeOH or =A1OH groups).  Competition of SO42'  and HP042' with
adsorption of LAS has been observed.
       The correlation of distribution ratio with sorbent properties is shown in  Figure 8. A weak
correlation with organic carbon content is seen and attributed to the hydrophobic effect. The apparent
correlation with clay  content is attributed to the covariability of clay content with organic carbon
content for these sorbents.
                 S0
                LAS
C-IO
C-12
C-14
chlorophenols
                    Linearity of  isotherm?
                    Effect of pH?
                    Effect of n_CH2-?
                        Figure 5.  Organic anions and acids in this stuuy.
                                            174

-------

 O
     0
                                     -10 LAS
                                     -12 LAS
                                 A  C-14 LAS
-11  -10  -9   -
                        8   -7  -6   -5   -4  -3
                        log  H  / M ................

              Figure 7. Effect of pH on sorption of LAS.
                             175

-------
HYDROPHOBIC IONIZABLE ORGANIC COMPOUNDS (6-9)

       Hydrophobia ionizable  organic compounds exist in the aqueous phase as a neutral species and
as an ionized species. The sorption of the neutral species can be characterized by the same paradigm
as that used for neutral organic compounds, while adsorption  of the ionic form requires consideration
of the charge, as outlined above for sorption of LAS.

       The organic acids with pH dependent charge are the chlorophenols, shown in Table I.  These
weak acids ionize according to the reaction:

       AH = A- +  H+          K,,                                                      (3)

and the neutral species is distributed between octanol and water  according to  the reaction

       AH =  AH            K™                                                        (4)

where the overbar represents a species in the nonaqueous phase. The pH-dependent concentration
distribution ratio  is formulated as

       D =   FA-1 +  FAHI                                                               (5)
              [A-] +  [AH]


where the total concentration of the acid in both phases is taken into account.


                              Table I.  Properties of chlorophenols.
                     Compound
logK,
2-chlorophenol
2,4-dichlorophenol
2,4,6-trichlorophenol
pentachlorophenol
-8.52
-7.85
-5.99
-4.74
2.17
2.75
3.38
5.01
                     Reactions

                            AH =  A- +  H+

                            AH =  AH

                     Distribution ratio

                        D = FAH1 + FA-1
                             [AH] + [A-]
                                          176

-------
                         of C-12 LAS on
                     Different Sediments
   Q
       300
       200
       100
               °— -o f
                       oe
                      % clay

                                    ,-/--°
         0.00

           I	
0.01   f()c    0.02
0.03
                        20.            40
                            % clay
                             60
      Figure 8. Effect of sorbent properties on sorption of LAS.
              PCP
    - o--o  TeCP
              TCP
Figure 9.  Effect of pH on octanol-water distribution of chlorphenols.
                          177

-------
       The distribution of the chlorophenols between octanol and water is shown  as a function of pH
in Figure 9.  Three regions are seen in the figure.  At high H* concentration, the neutral species is
dominant in both phases, partitioning is independent of pH and given by Equation 4. At intermediate
H+ concentration, A- is the dominant species in the organic phase and AH is the dominant species in
the organic phase; the distribution is pH dependent with a "knee" in the curve at pH = pK,.  At low
H* concentration, the ionic species is the dominant form in both phases, and partitioning is again pH
independent  Often the presence of the anionic species (A-) in the organic  phase has been ignored,
but as shown in Figure 9, it is significant.

       The partitioning of pentachlorophenol with a variety of sorbents reported in the literature (9)
is depicted in Figure 10; the pH dependence is similar to  that observed in Figure 9 for water-octanol
distribution, indicating that octanol is a good "model sorbent" for these compounds. However, note
that the ionic species of the compound in both phases must be taken into account, that  the relation of
Kj to K
-------
ORGANIC CATIONS (5)
       The organic cation used in this study was  N-dodecylpyridinium (DP) (Figure 11).  Adsorption
of organic cations is predicted to be stronger than adsorption of organic anions of comparable
hydrophobicity due to the electrostatic attraction of the cations to the  negatively charged   '
environmental sorbents.

       The isotherms for adsorption of DP on three environmental sorbents and two  specimen
minerals is shown in Figure 12 on a log-log plot. Highly nonlinear isotherms are observed, with the
log-log slope  of approximately 0.6 in the  lower concentration range.  The aqueous concentrations
cover seven orders of magnitude in one case. Although these isotherms show an almost step-function
appearance when plotted on a linear scale (not shown), it is clear that the  isotherm is continuous if a
wide range of concentrations is considered.

       The effect of pH is shown in Figure  13 and the isotherms obtained in  different salt
concentrations are shown in Figure 14. Adsorption is shown to be virtually independent of pH but
relatively dependent on salt concentration.   The independence of pH and dependence on salt
concentration is consistent with  a classical ion-exchange mechanism. Two features of the  isotherms at
different salt concentrations are noted: at low concentrations, the  isotherms are shifted along the
x-axis by about 0.8 log  units; at higher concentrations, at about the exchange capacity of the sorbent,
the isotherms cross over, with DP adsorption becoming favored at high salt concentrations. This
behaviour is  consistent with the reactions:
       NaX  +  R+  =  RX + Na+
and
       Y + R+ +  Cl-  =  YRC1
                                  (6)

                                  (7)
where:
NaX is the exchanger in the Na+ form, R+ is the DP cation, and

Y is a site for "hydrophobic" adsorption;

one might expect that Y is related to natural organic matter or even to RX groups.
                                 ORGANIC CATIONS
             dodecylpyridinium
                     DP
                                          quinoline
                   Linearity of  isotherm?
                   Effect of pH?
                   Effect of salt?
  pyridine
   (o)
                                                             N
B+H+ ^ Bf-T
                                                  Figure 11. Organic cations used in this study.
                                             179

-------
        Adsorption  Isotherms
     Dodecylpyridinium in 0.01 M Na+
0
\
o 0
E -2
i — i
S ~3
en
.2 -4
5

1 ' ' ' i. i0 oo
o Montmorillonite °
. • EPA-12 ° ,**
• Lula N6 - • ^^m
A Kaolinite *W** AA
- A Borden Sand B»**^ * AA **
g • A
• A
• A.
m A
~ «•* A ^
* A-^
J - * (^ 1 1 1 i
-9 -8 -7 -6 -5 -4 -3 -.
              log C(w)  [M]
Figure 12. Isotherms of DP on different sorbents.
            Effect, of pH  on
     Distribution of Alkylpyridinium
t
3
0
Q
cn
o
2

1
• • • •' . •
""" " "• .
A A A A A
A
• EPA-12 A A
• Lula N6
A Kaolinite
11 -7
log [H+]
 Figure 13. Effect of pH on distribution of DP.
               180

-------
               Dodecylpyridinium on Lula N6
                              2.5 g/L
               a  0.01 M NaCI
               •  0.1 M Nad
           -9    -8   -7    -6    -5    -4    -3    -2
                           log C(w)  [M]
       Figure 14. Effect of [NaCI] on adsorption isotherms of
                    DP on aquifer material.
       100
                                       Subsoil
                                     O Acidic
                                     A Basic
                                 (Zachara et al. 1986)
                                J	1
Figure 15. Effect of pH on sorption of the organic base, quinoline (10).
                          181

-------
 ORGANIC BASES (2,10)

        The adsorption of the organic base quinoline on two  subsoils as a function of pH (10) is
 depicted in Figure 15. A pH dependency similar to that seen in Figure 10 (pentachlorphenol) is seen,
 except that in this case, it is the ionic species that is bound more strongly than the neutral  species.
 (Note also that Figure 15 is on a linear-log scale, and thus not quantitatively comparable to Figure 10.)

 SUMMARY FOR ORGANIC CATION SORPTION

        The isotherms of ttie organic  cations are highly nonlinear, but not just step functions, although
 they may  appear to be so if plotted on a linear scale.  The adsorption of the cation is virtually
 independent of pH but dependent on salt concentration. The isotherms can be explained by a cation
 exchange reaction and a  "hydrophobic" adsorption  reaction.

                                          CONCLUSION

        These data and the accompanying explanations present a very simple picture  of sorption of
 neutral and ionic organic compounds.  As a first step in providing an overview, this approach is
 justifiable; however, the reader should  be aware that even relations such as Equations 3, which I have
 cited as one of the most generally applicable in all of  organic environmental chemistry, are  still
 active areas of research, as details such as type of organic carbon,  polarity of solute molecules etc. are
 further debated.  Thus, this discussion should be considered a beginning, not an end.

                                      LITERATURE CITED

 B. Brownawell, H. Chen, J. Collier, J. Westell, "Adsorption of Organic Cations to Natural Materials," Environ.
        Sci. Technol., 24, 1234-1241, 1990.
 C. Annette Johnson, John C. Westall, "Effect of pH and potassium chloride concentration on the octanol-water
        distribution of methylanilines," Environ. Sci. Technol., 24, 1869-75, (1990).
 R. Schwarzenbach, and J. Westall, "Transport of Nonpolar Organic Compounds from Surface Water to
        Groundwater:  Laboratory Sorption Studies," Environ. Sci. Technol., 15, 1360-1367, 1981.
 J. Westall, "Properties of Organic Compounds in Relation to binding by Natural Materials," in Transport and
        Transformation of Organic Compounds in Groundwater, G. Bengtsson, Ed., Swedish Natural Science
        Research  Council, Stockholm, 1984, pp. 65-90.
 B J. Brownawell, H. Chen, W. Zhang, J. C. Westall, "The Adsorption of Surfactants,"  in Organic Substances and
        Sediments in Water - Processes and Analytical, R.A. Baker, Ed. Lewis Publishers, Boca Raton, 1991,
        pp.  127-147.
 John C. Westall, Christian Leuenberger, Rene P. Schwarzenbach,  "Influence of pH and ionic strength on the
        aqueous-nonaqueous distribution of chlorinated phenols," Environ. Sci. Technol., 19, 193-8 (1985).
 John C. Westall, C. Annette Johnson, Wanjia Zhang, "Distribution of lithium chloride, sodium chloride,
        potassium chloride, hydrochloric acid, magnesium chloride, and calcium chloride between octanol and
        water," Environ. Sci. Technol., 24, 1803-10  (1990).
 Chad T. Jafvert, John C. Westall, Erwin Grieder, Rene P. Schwarzenbach, "Distribution of hydrophobic
        ionogenic organic compounds between octanol and water: organic acids," Environ. Sci.  Technol., 24,
        1795-803, (1990).
 Linda S. Lee, P.S.C. Rao, P. Nkedi-Kizza, JJ. Delfino, "Influence of Solvent and Sorbent Characteristics on
        Distribution of Pentachlorophenol in Octanol-Water  and Soil-Water Systems," Environ.  Sci. Technol. 24,
        654-661 (1990).
John M. Zachara, Calvin C. Ainsworth, Larry J. Felice, Charles T. Resch, "Quinoline Sorption and Retention to
        Subsurvface Minerals: Role of pH," Environ. Sci. Technol., 20, 620-627, (1986).
                                              182

-------
        A STUDY  OF HEAVY METAL SPECIATION MODELLING
                               ON THE LE AN RIVER

                                    Yuhuan Lin and Qi Li1

                                        ABSTRACT

       The toxicity of metal pollutants in aquatic systems varies from one metal species to another
depending upon the water quality and the types and concentrations of heavy metals within the system.
In this study, heavy metals within the water and sediment of a site along the Le An River, China, were
examined and classified as dissolved or suspended. The suspended paniculate metals were further
divided into active and inert classifications.  The active portion in water included a soluble part and a
part comprised of elements in suspended paniculate matter. The active portion in sediment included
soluble, exchangeable, carbonate-bound, and ferric-manganese oxide- bound parts, while the inert
portion included organic sulfide-bound and residual parts.  A model was proposed to simulate the
aquatic transport and transformation of dissolved and paniculate heavy metal species. The model
consisted of two components: the River model, and the MINTEQA2 Thermodynamic Equilibrium
model (U.S.EPA, ERL, Athens, Georgia). The model was tested on field-collected water quality data
from a section of the Le An river, and on hydrological and meteorological data from the main
hydrological station along the Le An river (up-, mid- and downstream sites).

       Using the River model, calculations of the  concentration and speciation of copper in both
water and sediment indicated that the impact of acidic mine drainage on the water quality of the Le
An river was considerable. Near the discharge site, the concentration of copper in sediment reached
levels of 3-5 g/kg, while the concentration of dissolved copper in the water at this site exceeded the
national standard of 0.01 mg/1.  The length of river polluted by the acidic mine drainage was about 50
km during the dry season, and up to 200 km (to Poyang Lake) during the flood season.  The
model-derived results for copper were in agreement with the values measured in the field for both
water and sediment.

       The species of heavy metals present within the water column of the river were calculated using
the MINTEQA2 model, and recorded values of water quality and dissolved metal concentrations from
the Le An river. The results indicated mat large amounts of Al .and Fe were introduced into the river
from the mine drainage.  These metals flocculated  and formed a sediment on the river bottom during
the dry season which was then carried downstream to the Le An river estuary during the flood season.
A prediction of the effects of these metals on aquatic organisms was made by comparison of the actual
metal concentrations with standard allowable levels.

                                     INTRODUCTION

       The study was conducted on a section of the  Le An river system in the northeastern part of the
Jiangxi province (Fig. 1).  The Le An river forms from the drainage of a 9616 km2 water shed around
the Huiyueshan mountain (lat. 29'11")  at an elevation of 860 m, and travels 279 km through Dexing
and Leping Counties before reaching Poyang Lake (elevation 32 m) in Poyang County Gat. 29"2").
Within Dexing County, the Le An river is fed by two tributaries, the Dawu and Jishui rivers, which
    'Research Center for Eco-Ehvironmental Sciences, Chinese Academy of Sciences, Beijing 100085,
China

                                       ,   183

-------
                                         Snmpling s i Les
Lo.13
Lo.l
Lo.2
Lo.3
Lo.4
Lo.5
Haikou
CUkou

.\inngl_un
Daicun
Lo.6
Lo.7
Lo.8
Lo.9
Hush.-in
Jicdu
llan.i i ndu
Sliir.hcnjic
                                                            Lo . 1 0 Cn i j i Aw.-i
                                                            I.O.11 Hunng 1 onsmi .TO
                                                            I.o . 1 2 Shunn*;i;;\n<
                                                            Lo.13 Longkou
                    l.o.ll
                         Figure 1. Sketch on Le An River, Poyang Lake area.
  are surrounded by four different mining operations. The largest open-cast copper mine in China is
  located along the Dawu river, while the Fujiawu and Damaoshan copper mines and the Yinshan lead
  and zinc mine are found along the Jishui river. A large amount of acidic mine waste is discharged
  from these mines, which imposes a serious heavy metal pollution load (especially of copper) upon the
  Le An river system.

         The toxicity of metal pollutants in aquatic systems varies from one metal species to another
  depending upon the water quality, and the types and concentrations of heavy metals within the system.
  In this paper, a model of heavy metal species in river systems is proposed.  The model consists of two
  components:
         (1)  The River model, which treats the transportation of heavy metals within the river,
  including dissolved metals and suspended particles. Metals defined as suspended particles are further
  divided into active and inert parts.
         (2)  The geochemical, thermodynamic equilibrium model, MINTEQA2 (from U.S. EPA,
  Athens, Lab.), which can simulate metal species within an aquatic system.

  RIVER MODEL FOR METAL SPECIATION

         Suspended paniculate metals can be classified as being active or inert.  The active portion
  includes metal species that are water soluble, can exchange cations, or are either carbonate or
  ferric-manganese bound. These  active portions are easily altered with changes in water quality, and so
  may take part in the movement and transport of elements in water.

         The sulfide-bound and residual metals form the non-active or inert portion.  It is necessary to
  simulate the movement and transport of the active and inert portions of metals in suspended sediment
                                              184

-------
separately.  The grain size of metals transported in suspended sediment is generally less than 20 jjm in
diameter.  A model describing the transport of dissolved and paniculate metals is proposed below:

dC = D 82C + V8C - Ks • C + R                                                               (1)
dt       8x2       8X

where;
C is the metal concentration (active or inert portions) in suspended matter (mg/1)
V is the average flow velocity (m/s)
D is the dispersion coefficient of metal (m2/s)
Ks is the sedimentation and resuspension coefficient of metal (1/s)
R is the point or nonpoint source load (mg/s).

       The expression can also be used to calculate both the active and inert portions of metals in
suspended sediment. In order to calculate the concentration of dissolved metal, it is necessary to
define the partition coefficient representing the equilibrium relationship between the dissolved species
in the water and the active portion of metals in suspended matter.  The partition coefficient is
expressed as follows;

Kp = Cse/Cdw                                                                               (2)

where;
Kp is the partition coefficient of dissolved and suspended metals (I/kg)
Cse is the concentration of the active part of metal in suspended matter (mg/kg)
Cdw is the concentration of metal dissolved in the water (mg/1).

The total concentration of metal in the water is:

Ct = Cdw + Cac - Cin                                                                        (3)

where;
Cac and Cin are the concentrations of the active and  inert metal portions, respectively (mg/1)

The active metal concentration can be represented by:

Cac = Cse * SS * 10-6                                                                       (4)

where;
SS is the concentration of suspended matter (mg/1).

In the model, the balance of total suspended matter (SSt) in the  water is:

SSt = SSi + SSp - SSout - SSd                                                                (5)

where;
SSi  and SSp are the flux of suspended matter from the upstream and lateral boundaries,
respectively (g).
SSout is the downstream output flux of suspended matter (g)
SSd is the flux of suspended matter as bottom sediment (g).
                                             185

-------
 Finally, the total concentration of suspended matter is expressed as:

 SS = SStfQ                                                                                  (g)

 where;
 Q is the volume of water (m3) within a river segment (defined below).

        Expressions (1) to (6) can be used to represent the physical and chemical transport processes
 of metal within a river, including dispersion, precipitation, dissolution, adsorption, desorption, and
 sedimentation.

        a)  Segmentation of the River and Hydrological Data

        In order to simulate the transport and transformation of metals within a river, it is necessary to
 divide the river into several segments based on hydrological data and river moqjhology.  In this study,
 the section of the Le An River being examined was roughly divided into three sections; upstream,
 midstream, and downstream, corresponding to the locations of three hydrological  stations (Table 1).
 The hydrological data used was collected from 1975 to 1986.  In order to increase the accuracy of the
 simulation, the studied section of the Le An river (from the point of heavy metal  discharge to a point
 159 km downstream) was then subdivided into ten segments (Fig. 2). Water velocity, runoff, and
 runout from the boundary were estimated from the meteorological and hydrological  data (Table 2).
 The results indicated that there was  considerable variation between the upstream and downstream flow
 volumes and velocities. The seasonal runoff of the river also  varied considerably, with flow rates of
 between ten to a few hundred cubic meters per second in the dry season, increasing to rates of several
 thousand cubic meters per second in the wet season.
               Table 1. Characteristics of the three main sections of the Le An river.
Sections
Length (km)    Gradient (%o) Width (m)     Depth (m)
Upstream
Midstream
Downstream
61.4
83.6
72.5
1.0
0.34
0.065
50-100
100-150
>200 5-15
1-5
2-8

         Table 2.  Flow velocity, runoff, and runout from the boundary in the normal season.
Segment
                                                          10
11
Q (m3/s)
V(m/s)
Qb (m3/s)
116
0.92
10
129
0.80
3
134
0.75
5
149
0.69
15
184
0.42
35
199
0.38
15
204
0.32
5
219
0.28
15
224
0.27
15
224
0.26
20

0.20

                                            186

-------
                0  19.520.5 40   60      84
                                               96     106    125    13b     159 Ion
                   123    4
                                                                     10  n
                         Figure 2.  Scheme of segments on Le An River.
       b)  Calculation of Partition, Sedimentation, and Dispersion Coefficients

       The partition coefficient for copper, Kp, was calculated for each river segment based upon the
concentrations of dissolved metal in the water and of the active metal in suspended matter.  The
partition coefficient varied considerably between the first five segments (Table 3). The concentration
of the active metal portion in segment 2 was so high that the dissolved concentration in the water may
have reached levels up to several hundred parts per billion.  Below segment 5, Kp stabilizes.
Segment
           Table 3.  Partition coefficient for copper, Kp, at different segments of the river.
10
Kp(l/kg)       5800   5099   1166  2825   500    500    500    500    500    500
The sedimentation coefficient, Kd, was calculated as:

Kd = dp/St

where,
dp is the average depth of water in each segment (m)
St is the sedimentation velocity of suspended particles (m/s).
The results are illustrated in Table 4.
                                            187
            (7)

-------
 Segment
              Table 4. Sedimentation coefficient, Ks, at different segments of the river.
1
6
8
10
 Ks(*10-3)     0.92    1.98    0.51    0.16   0.16   0.15   0.05   0.05   0.05   0.05
 (s-D
The dispersion coefficient, Dl, was calculated according to Pick's Law:
N = D, * A * dC                                                                        (8)
        X
where;
Dl is the dispersion coefficient (m2/s)
A is the cross-sectional area of the segment (m2)
X is the distance between midpoints of adjacent segments (m)
£>C is the differential concentration of segments (g/m3)
N is the mass flux of the heavy metal within the cross section (g/s).  The value of N was assumed to
be 100 m2/s for calculation and had little effect upon the final results.

THE THERMODYNAMIC EQUILIBRIUM MODEL - MINTEQA2

        The MINTEQ family of programs, of which the latest version is MINTEQA2, was developed
by the USEPA. These programs employ the chemical equilibrium constant method to relate metal
species within a chemical system and can calculate the equilibrium concentration of individual species
in an aquatic environment.

       The results from the river model for metal speciation were entered into the MINTEQA2
program and the species distribution of heavy metals throughout the water column were calculated for
the individual river segments.

       a) The Water Quality of the Le An River

       Cation and anion concentrations  and pH were measured from river samples from each segment
during the field investigation portion of the study, and are shown with the results of the MINTEQA2
program in Table 5. The pH analysis indicated that pH tends to be higher during the dry season and
lower during the wet season (Fig.  3), implying that the impact of acidic drainage on water quality was
greater in the wet season. Dissolved oxygen and Eh (approximately 190 mv) were stable throughout
the river.
                                           188

-------
Lo.
Lo.
                Table 5. Water quality and ionic composition of the Le An river.
pH
Ca
Mg
 K
Na
Cu
Pb
Cd
Fe
Mn
Al
As
        S07
1
2
3
4
5
7

8
9
10
11
12
13
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
o.oo .
0.24
13.41
3.20
3.45
1.10
0.64
0.38
0.39
0.193
0.36
0.097
0.035
0.014
0.030
0.719
0.112
0.125
0.078
0.096
0.071
0.070
0.061
0.054
0.001
0.001
0.001
0.098
4.582
0.299
1.061
0.462
0.289
0.164
0.186
0.135
0.225
0.016
0.010
0.0023
-0
-0
-0.003
-0
-0
-0
-0
-0
-0
-0
-0
0.0007
0.0007
4.315
58.82
7.79
7.82
3.69
4.18
3.54
2.09
3.08
2.54
4.62
3.69
3.72'
3.27
152.08
25.68
28.35
8.48
12.62
9.62
4.79
4.80
24.40
6.60
5.90
5.80
Zn
1
2
3
4
5
7

8
9
10
11
12
13
6.56
4.79
6.46
7.17
7.20
7.25
7.25
7.25
7.25
7.25
7.25
7.25
7.25
7.068
20.59
13.76
9.33
5.82
9.05
9.05
7.45
7.82
8.51
9.21
9.02
8.89
1.54
11.07
2.48
2.33
1.43
2.18
2.145
1.81
1.81
1.73
2.02
1.98
1.82
1.73
1.05
1.51
0.87
0.93
1.09
0.64
1.22
3.63
1.01
3.20
3.49
3.39
1.84
2.55
2.11
2.10
1.61
1.81
1.91
2.11
3.13
1.91
2.95
2.66
2.53
0.003
1.94
0.10
0.18
0.07
0.04
0.02
0.01
0.001
0.001
0.001
0.001
0.001
0.021
0.021
0.001
0.001
0.001
0.01
0.003
0.004
0.001
0.004
0.001
0.001
0.001
0.031
0.053
0.010
0.013
0.021
0.057
0.052
0.034
0.007
0.059
0.001
0.001
0.001
1. 1989, 4; 2. 1989, 12; 3. 1990, 4.
                          Figure 3. pH changes along Le An River.



                                         189

-------
                              DISCUSSION AND CONCLUSION

 THE CONCENTRATION AND SPECIATION OF HEAVY METALS IN RIVER SEDIMENT AND WATER

       In this study, copper was used as a representative heavy metal to assess the validity of the
 model in predicting heavy metal transport and deposition.

       a)  Total Sedimentary Copper Distribution

       The calculated copper concentrations in sediment over the study area during three seasons
 (flood, dry, and normal) are depicted in Fig. 4. Copper concentrations reached highs of 3000-4000
 mg/kg near the Dawu river mouth (location 2) which then decreased to 700-800 mg/kg at the
 downstream end of the study area (location 10). There was strong  agreement between the calculated
 results (Fig. 4) and the field-collected data (Fig. 5) regarding the copper concentration distribution.
 The copper concentration in the segment near the point of pollutant discharge changed dramatically
 with the season and, according to the calculations, with different flow rates.  The concentration of
 copper in the river segment near the source of discharge was typically higher while the transport
 distance of copper was shorter, in the normal season compared with the flood season.  During the dry
 season the copper concentration in the segment near the Dawu river mouth reached a maximum but
 the transport distance was much shorter than in other seasons.

       During the flood season, the concentration of copper along the river changed slowly, with the
 attenuation process dominating. Accordingly, the rate of sedimentation of pollutants along the river
 bottom diminished, and the distance that pollutants could be carried by the river exceeded 200 km (to
 Poyang Lake, location 13).  The river  velocity during the flood season also increased, and thus lateral
 transport of particles was greater than the sedimentation rate.  During the normal and dry seasons, the
 results indicated that sedimentation was the principle transport process.

        4.5
      -3.0
      X
      -15
      |i.s
      8 1.0
      3 0.5
         0


• f
G2 Flood
GJ Dry
C3 Normal
19.5 20.5  40  60  84  96  106 125  138 159
            Distance, km
                                                                      1  Dec.  1989
                                                                      2  Sept. 1989
                                                                      3  May 1990
                                                   19.5 20.5
                                                              60  84  96  106  125 138  159
                                                               Distance, km
Figure 4. The calculated Cu concentration in sediments.    Figure 5. The concentrations of Cu in sediment.


       b) Speciation of Copper in Sediment

       The total, inert and active copper concentrations within sediment during the normal  season are
illustrated in Figure 6. The inert and active portions of the sediment-bound copper can be liberated
into the water.  Analysis of the speciation partition of the sediment-bound copper by sequential
extraction indicated that 50-90% of the copper was inert, while the remainder, 10-50%, was active.
This percentage calculated as inert was in agreement with the experimental results.  Any difference
                                           190

-------
may be accounted for by loads coming from points other than the monitored discharge site, and
temporal differences during estimation of the drainage composition.

       The inert proportion of copper in sediment downstream was greater during the flood season
(Fig. 7) than in the normal season (Fig. 6) because of the transport of suspended ore residue and
tailings.  During the normal season most of the copper particles were deposited in the river segment
immediately below the discharge site.  The proportion of active component at downstream sites was
greater during the normal season than in the flood season.
        19.5 20.5  40  60  84   96  106 125  138 159
                     Distance, km
Figure 6. The speciation of Cu in sediments
at normal season (Sept 1989).

       c) Total Aquatic Cu Distribution
                                                 19.520.5
            84  96  106
            Distance, km
Figure 7.  Speciation of Cu in sediments
at flood season (May 1990)
       The calculated concentration distribution of copper in the river water (Fig. 8) was the reverse
 of what was in the sediment. During the dry season, the concentration in the river segment near the
 discharge site was less than in the dry season, and pollutants were generally carried no further than 50
 km downstream.  The concentration gradient along the river was steeper in the dry season.
 The concentration of copper in the water at locations 2 and 3 varied seasonally from 4.5-25 ppm
 depending upon the initial load of pollutant. The concentration dropped to 0.5-0.05 ppm in river
 segments 50 km downstream, while at location 10 the level was 0.01-0.005 ppm.

       A comparison of the calculated and field data indicated that the aquatic copper concentration
 was dependent upon the ratio of volumes of runout to drainage water.  Higher flow rates and greater
 drainage led to higher levels of copper at downstream segments.  Conversely, decreased flow rates and
 low drainage resulted in high copper concentrations  only in the segment near the discharge site.

       d)  Aquatic Speciation of Cu

       The total, dissolved, and paniculate (inert and active portions) concentrations of copper were
 calculated from the model for the flood  season (Fig.  8).  Over 90% of the copper was found to exist in
 suspension, with 80% inert  and less than 10% (53 to 100 ppb) dissolved.  During the normal season,
 the concentration of the dissolved fraction can increase to 300 ppb (Fig. 9). The ratio of particulate to
 dissolved components remains constant  (about 9:1) independent of the season.
 The particulate component of the total aquatic Cu concentration estimated by the model was 10 to
 20% higher than that found from the field data, while the value for the dissolved component was
 approximately the same. This difference may be the result of the model input assumption that 90% of
 the total concentration of copper in the drainage  was suspended particulate copper.
                                             191

-------
 THE CHEMICAL SPECIES OF HEAVY METALS THROUGHOUT THE WATER COLUMN

        The thermodynamic equilibrium model (MINTEQA2) can be used to calculate the species of
 metals within a water column from water quality data and dissolved metal concentrations calculated
 from the river model. The results indicated that the species of metals found within the water column
 are dependent upon the pH, Eh and anion concentration.

        The concentration of metals in the water at location 1 was low, with simple hydroxide
 compounds and Mn as the dominant metal species (Fig. 10).  In contrast, the dominant metal species
 at the pollutant discharge site at the mouth of the Dawu river were very complex (Fig. 11).  Copper
 was the main species found here, while ferric hydroxides, Fe(OH)2+ and Fe(OH)2+, and A13+ were
 also predominant. The amount of free ferric iron and aluminum hydroxide was very low. The
 composition of metal species found in river segment 2 was similar to that found at the mouth of the
 Dawu river (Fig. 12).

        There was a  noticeable difference between the metal species  composition of the pollutant plume
 found at location 2 and that found 10 km downstream at location 3 (Fig. 13). Large amounts of Al
 and Fe were initially found in a number of different forms, including A1(OH)3, Fe(OH)2+, A1(OH)2+,
 and A1(OH)4-, while copper was initially found as Cu(OH)2 or CuCO3.  There was then a dramatic
 increase in the flocculation of Al and Fe species and precipitation from the water column into the
 sediment, so that the concentration of metals in the water rapidly decreased downstream.  The aquatic
 metal species composition from location 3 to location 5  was similar.
                             1 Total
                             2 Inert  ,•*
                             3 Equilibrium
                             4 Dissolve
        19.5 20.5 40
                   60   84  96  106  125 138  159
                     Distance, km
Figure 8. The total, dissolved Cu in water,
active and inert species in suspension at
flood season.
0.28
,J °-24
£ 0.20
.1 0.16
s
gO.12
J0.08
0.04
0
-
-

.
7m

i
YS
y^
'/
"A
I





19.5 20.5 40
           60  84  96  106  125  138 159
            Distance,  km
Figure 9.  The dissolve Cu concentration in
water at dry season.
       Further downstream between locations 5 and 10, the concentration of metals was low (Fig. 14).
Simple hydroxides predominated, and the water quality was essentially restored to normal levels.

BIOASSESSMENT OF METAL SPECIES

       The toxicity of metal ions varies between species depending upon environmental conditions and
species composition.

       a)  The formation of Al species in an aquatic environment is pH dependent, with free Al
predominating at pH levels below 5.0.  Free Al ions formed 52.8% of the total concentration at the
                                           192

-------
Dawu river mouth and 64.2% at location 2. The principle aluminum mineral formed at Gukou and at
the Dawu river mouth may have been alunite (3A12O3*K2O*4SO3*6H2O). At location 3, the pH
varied from 4.79 to 6.64 and A1(OH)3 was the predominant species, forming 58.3% of the total
concentration.  The free Al ion concentration was very low from the middle river segments
downstream to Poyang Lake. Dissolved Al concentrations above 1.5 mg/1 can cause pronounced
physiological stress on warm-water fish; 0.5 mg/1 is the standard maximum concentration (Moore,
1984).  The concentration of Al introduced into  the Le An river through acidic mine drainage exceeds
the national permissible level established for fish and other aquatic organisms, particularly during the
flood season when pH is low.  Locations 2 and  5 would be the most toxic with respect to Al, and
form a serious environmental hazard.  The impact of Al on the Le An river ecosystem needs to be
closely monitored.

       b) Copper is highly toxic to aquatic plants, invertebrates, and fish.  The copper concentration
at the mouth of the Dawu river and at location 2 reached 1.944 mg/1, which exceeds the established
fisheries standard.  The total downstream concentration almost exceeds the fisheries standard of 0.01
mg/1. Free Cu2+ is the dominant form below pH 5.0.  Locations 2 (Gukou) and 3 (Zhongzhou) would
be the  most toxic.

       c) Other elements:  The concentrations of cadmium, zinc, and lead were very close to the
established fisheries standards, particularly during the dry season. At location 3, free  Cd was the
predominate form of that element below pH 5.65 while CdCOS predominated above pH 7.25.  The
relatively high solubility of ZnCO3 and Zn(OH)2 resulted in high levels of dissolved  Zn being found
in the river.

       There is no fisheries standard established for Fe in China, but the standard level in drinking
water has been set at 0.3 mg/1. It is impossible to predict the impact of Fe upon the aquatic
ecosystem. While the toxic effects of manganese on aquatic species are thought to be limited, the
maximum allowable Mn concentration has been reported to be 0.77 mg/1 for rainbow trout and 1.0
mg/1 for drinking water. The large amounts of  Fe and Mn found in the Le An river will cause a water
treatment problem with regards to drinking purposes.
                                             193

-------

        Figure 10. Speciations of metals at Location 1.
    Figure 11.  Speciations of metals at Dawu River mouth.
       Figure 12. Speciations of metals at Location 2.
       Figure 13.  Speciations of metals at Location 3.
J       /  F.WIO;
"1       I   or -~ ~=
                          Ml<0,
                          '•7>
     Figure 14. Speciations of metals at Poyang Lake.
                        194

-------
                                       SUMMARY

      The ongoing drainage of acidic mining waste into the Le An river has significantly affected
water pH and the concentrations of aluminum, iron, and copper. The River model and the
MINTEQA2 model were used successfully to estimate heavy metal speciation over three seasons. The
calculated data was found to agree well with the monitored data. Use of the River transport model
and the Equilibrium model in combination allows for the description and prediction of water quality.

                                ACKNOWLEDGEMENTS

      Many thanks to Dr. W. Salomon (Institute of Soil and Fertility RA, Haren, Holland) and Dr. R.
Russo (Director, ERL, Athens, USEPA) for support of the river model and MINTEQA2 program.

                                      REFERENCES

Campell, T.A. 1979. PGC Bisson. M. Anal. Chem. 51:844.
Fostner, D. and W. Salomons. 1983. Proc. NATO adv. Res. Workshop.  Nerri/Italy, New York,
      Plenum Press, pp. 245.
Klenkel, P.K. 1981. Modeling of river (Hsien Wen Shen ed.), New York, John Wiley, pp. 17.
Moore, J.W. and S. Ramamoorthy. 1984.  Heavy metals in the natural water. New York,
      Springer-varlog, pp. 268.
Ramsey D.L. and D.G. Branoon. 1988. Water,  air, and soil pollution.  39:1
USEPA. MINTEQA2 computer program for calculating aqueous geochemical equilibria.  Envir. Res.
      Lab, R. and D. office, Athens, Georgia. 90613.
                                          195

-------
                         UNCERTAINTY ANALYSIS WITH
                     CORRELATED INPUTS USING QUAL2E

                                    Linfield C. Brown, Ph.D.1

                                      INTRODUCTION

        Quantitative uncertainty analysis is assuming increasing importance in surface water quality
 management Decisions affecting the appropriate use of stream assimilative capacity, the permitting of
 waste water discharges, and the prediction of toxicant concentrations are increasingly subject to debate
 because of questions arising from the quality of the field data and the validity of model assumptions
 used to make forecasts.  Conventional approaches to uncertainty analysis have generally assumed that
 all model inputs act independently. The significance of correlation among model inputs on errors in
 model predictions, though usually unknown, has always been presumed to be important. Only recently
 have investigators developed procedures for estimating the magnitude of its impact in model
 forecasting.

        This study is a continuation of previous work (Brown, 1990), which reported on the effect of
 correlation among input parameters and variables on the output uncertainty of predicted dissolved
 oxygen (DO) concentrations for a single reach  Streeter-Phelps model. The current study presents
 preliminary results from applying the input correlation computational strategy developed for the
 Streeter-Phelps model (analytical solution) to the overall uncertainty analysis framework of the multi-
 reach QUAL2E-UNCAS model (numerical solution).  The QUAL2E-UNCAS model was selected for
 this application because  it is a general purpose  computer code, widely used by consultants and state
 regulatory agencies in waste load allocation and other water quality management activities.  The
 QUAL2E-UNCAS model also has a substantial international user community.

                                       PRIOR WORK

       Burgess and Lettenmaier (1975) and Chadderton et al. (1982) performed first order  error
 analysis on the  DO deficit from Streeter-Phelps models, without including the effects of correlation
 among the input parameters.  Song and Brown  (1990) showed that the effect of correlation  among
 input parameters in Streeter-Phelps type models could increase the variance of the predicted dissolved
 oxygen (DO) deficit by 20 to 40%, depending on the conditions of the simulations and the travel time
 at which the correlation effects  are assessed.  A monte carlo analysis showed that the effect of model
 non-linearities on the validity of the first order  analysis was small compared to the correlation effects.

       Brown (1990) demonstrated that the effect of correlation among  input parameters and variables
 was also  markedly dependent on the structure of the correlation matrix used in the uncertainty
 analysis.  Working with hypothetical data sets for a single reach Streeter-Phelps model, he reported
 that input correlation could either increase or decrease the standard deviation of the simulated DO
 deficit, depending  on  whether the correlations represented model use in a calibration  or a forecast
mode. Typical forecast mode correlations tended to increase (25-50%), while calibration mode
correlations tended to decrease (20-35%), the DO deficit standard deviations when compared to those
obtained when assuming  independence among input parameters and variables.
    'Department of Civil and Environmental Engineering, Tufts University, Medford, MA 02155, U.S.A.

                                          196

-------
       The objective of the current study is to assess whether the effect of these two different
correlation structures on uncertainty analysis would be manifested in a fashion similar to that reported
by Brown (1990) when incorporated into a numerical water quality model applied to multiple reach
stream network This assessment is made using field data from representative case studies.

                                      QUAL2E-UNCAS

       QUAL2E (Brown and Barnwell, 1985) is a comprehensive steady state water quality model
that can simulate up to  15 different state variables, including: DO; carbonaceous biochemical oxygen
demand (CBOD); temperature; algae (as chlorophyll-a);  organic, ammonia, nitrite, and nitrate nitrogen;
phosphorus; coliform bacteria; and conservative constituents.  The model is applicable to dendritic
streams that are well mixed, and can accommodate multiple waste discharges,  withdrawals, tributary
flows, and incremental inflow and outflow.  QUAL2E-UNCAS is a recent enhancement to QUAL2E
(Brown and Barnwell, 1987)  which allows the modeler to perform uncertainty analysis on the steady
state water quality simulations. Three uncertainty options are available: sensitivity analysis, first order
error analysis, and monte carlo simulation.

       One of the limitations of the original UNCAS framework in QUAL2E is that the uncertainty
analysis methodologies  assumed that each input variable or parameter used in the model acts
independently of the others.  This assumption was appropriate for obtaining a first approximation to
the magnitude of the effects of input uncertainty on the  precision of water quality model predictions.
However, the hypothetical studies cited previously clearly demonstrate the necessity for accounting for
the potentially important effect of input correlation  in uncertainty analysis. Thus, the first project task
was to modify the QUAL2E-UNCAS code to accommodate correlation among input parameters and
variables. The second was to perform the case study analyses.

       The original strategy in developing QUAL2E-UNCAS was to keep user input coding
requirements to a minimum.   Supplying an input correlation matrix for the large number of QUAL2E
inputs (potentially over 100) is a daunting task, from both a data gathering and an input coding view-
point.  The work of Song and Brown (1990) showed that the important correlation contributions come
from just a small sub-set of the total set of input  pairs.  Based on these findings, it seemed
unnecessary and unreasonable to require the user to specify a large, complete,  rigidly formatted
correlation matrix to perform an UNCAS simulation.  It was decided that QUAL2E-UNCAS
simulations with correlated inputs would be done simply by having the user specify the correlation
coefficients only for the pairs of inputs for which correlations were desired. This input is formatted as
a simple  list identifying the input pair and its corresponding  correlation coefficient.  The list is '
supplied  in any order, and appears at the end of the data file containing the variances of the input
parameters and  variables used in the uncertainty analysis. All input pairs not appearing in this list are
assumed  independent by default.

                         UNCERTAINTY ANALYSIS TECHNIQUES

       Two of the uncertainty analysis methodologies in QUAL2E-UNCAS were compared in this
study: first order error analysis (FOEA) and monte  carlo simulation (MCS). Both techniques offer
unique insight into the manner and extent to which uncertainty in model  inputs propagates through a
model and is manifested as uncertainty in the output water quality variables. First order error analysis
provides  the user with a components of variance  analysis, while monte carlo simulations supply  the
frequency distributions  of the output water quality variables.
                                           197

-------
         In first order error analysis, uncertainties in model input parameters and variables are
 propagated through the model using the first order terms in a Taylor series expansion about the mean
 value of each input.  The mean and the variance of the output variable can be obtained from the
 linearized equation as follows (Benjamin and Cornell, 1970):
                                Ay \2
 where; ¥(¥,,) denotes the variance of the output variable Yk, n is the number of input variables and
 parameters, V(Xj) is the variance of the input variable Xi; and cov(Xi;Xj) represents the covariance
 between inputs Xj and Xj.  The derivative approximations, AY/AX represent the sensitivities of the
 model output to the model inputs.  They are computed in QUAL2E-UNCAS as the difference in the
 function values evaluated at the mean values of the input variables and at the mean values with X,
 perturbed by an amount AXj. The first term in Eq. 1 represents the contributions to output variance
 from the variance of each input variable acting independently. The second term in Eq. 1 denotes the
 contributions to the output variance from correlation among the different pairs of input variables.

        In monte carlo simulation, a probability distribution is identified or assumed for each uncertain
 model input parameter or variable. Repeated model simulations are performed for a predetermined
 number of times, with each simulation using a value for each uncertain input randomly selected from
 the assumed probability distribution.  The resulting distribution of simulated output values can be
 analyzed statistically to compute an output variance that represents the combined effects of all model
 input uncertainties. Several references describe methods  for determining an adequate number of
 simulations to provide information on model output frequency distributions. Song and Brown (1990)
 and Brown (1990) used 1000 simulations, and Burgess and Lettenmaier (1975) used 2000.
 Considering the reliability of the output results and the computational time requirements, 1000
 simulations were  chosen for the monte carlo analysis in this study.  The multivariate normal •
 probability distribution was used for all monte carlo simulations.

                                 EXPERIMENTAL DESIGN

       Two data sets were used to assess the effects of correlation among model inputs on the output
 variable uncertainty in the QUAL2E model. The first is from the Withlacoochee River in north
 Florida, a small stream with a single industrial discharge.  The second is from the Ouachita River in
 Louisiana, a large slowly moving navigation stream with multiple industrial and municipal inputs.
 Two representative correlation matrices were selected for study.  One is typical of correlations found
 while using the model in a calibration mode, the other is representative of a forecast mode.

       The correlation among water quality model input  variables and parameters has not been well
 studied. Attempts to locate appropriate field data on which to base quantitative estimates were met
 with limited success.  Even the extensive field and laboratory data from the two case studies used in
 this paper were not sufficient to extract reliable quantitative estimates of these correlations.  Thus,
 correlations were assigned (assumed) on a qualitative basis using two criteria; sign (positive and
negative) and strength (strong (0.7-1.0), moderate (0.4-0.6), weak (0.1-0.3), or negligible(O.O)). For
example, assigning a correlation coefficient of 0.6 between a CBOD rate coefficient and an ammonia
(nitrification) rate coefficient means that these two inputs have a moderately positive correlation.

                                           198

-------
                 Table 1.  Input Correlation Coefficients for Uncertainty Analysis
                                   Using QUAL2E-UNCAS
   Input Parameter/Variable Pair
CBOD rate coefficient
CBOD rate coefficient
CBOD rate coefficient
CBOD rate coefficient
CBOD rate coefficient
CBOD rate coefficient
NH3 rate coefficient
NH3 rate coefficient
NH3 rate coefficient
NH3 rate coefficient
NH3 rate coefficient
Point load CBOD
Point load CBOD
Point load CBOD
Point load CBOD
Point load CBOD
Point load NH3
Point load NH3
Point load NH3
Point load NH3
Algal max. growth rate
Algal max. growth rate
Algal respiration rate
SOD rate coefficient
Reaeration rate coef.
Reaeration rate coef.
Reaeration rate coef.
       Forecast   Calibration
                                                                  Mode
                                            Mode
 NH3 rate coefficient               0.4
 Point load CBOD                  0.6
 Point load NH3                   0.3
 Algal max. growth rate
 Algal respiration rate
 SOD rate coefficient               0.5
 Point load CBOD                  0.3
 Point load NH3                   0.6
 Algal max. growth rate             0.2
 Algal respiration rate              0.2
• SOD rate coefficient               0.2
• Point load NH3                   0.8
• Algal max. growth rate             0.3
• Algal respiration rate              0.3
• SOD rate coefficient               0,4
• Point load DO                    0.2
• Algal max. growth rate             0.5
• Algal respiration rate              0.5
- SOD rate coefficient               0.4
-Point load DO  .                  0.1
- Algal respiration rate              0.7
- SOD rate coefficient               0.2
- SOD rate coefficient               0.4
-Point load DO                    0.1
- Algal max. growth rate            -0.5
- Algal respiration rate              -0.4
- SOD rate coefficient              -0.4
-0.3
-0.5
 0.2
 0.3
 0.3

 0.2
-0.5
-0.1
-0.1

-0.6
-0.4
-0.4
 0.2
 0.2
 0.7
-0.3
-0.3
        In a manner consistent with prior work (Brown, 1990), two different correlation matrices were
 constructed for this study (see Table 1). The first set of values was designed to represent the input
 correlation structure for the case of applying the model in a forecast mode.  The second set was
 designed to represent a calibration mode.  The primary difference between these two sets is mat in the
 forecast mode, values of model inputs are generally assumed or fixed by regulation to represent some
 unknown future condition; while in the calibration mode, many of the model inputs are measured and
 the remaining ones are adjusted (within acceptable ranges) to match an observed set of state variables.
 The correlation relationships for these two cases are quite different.

        For the forecast mode correlation matrix, 25 pairs of inputs were assumed to have non-zero
 correlation coefficients. The values assigned represent the general tendency for the mean of one input
 parameter to increase or decrease as the mean of the second input increases.  For example, the
 correlations between CBOD rate coefficient and CBOD point load, and between ammonia
 (nitrification) rate and ammonia point load were modeled as moderately positive (+0.6).  Strong
 correlations were assumed between CBOD and ammonia point loads (+0.8), and between algal
 maximum growth and respiration rates (+0.7). The correlations between headwater conditions and all
                                           199

-------
 other inputs were assumed to be negligible (0.0), as were those between climatological factors and all
 other inputs.  The correlations between reaeration rate coefficients and the algal growth and
 respiration parameters was assumed moderate to strong and negative (-0.4 to -0.5).

         For the calibration mode correlation matrix,  17 pairs of inputs were assumed to be correlated.
 The major differences from the correlations in the forecast mode  are in assessing whether measurement
 of these particular inputs can be achieved independently of each other.  For  example, the value of the
 reaeration rate coefficient, if measured by a tracer technique, can be obtained independently of all
 other input measurements. On the other hand, the value of the CBOD load  cannot be measured
 independently from the value of the organic nitrogen or ammonia load, nor independently from the
 CBOD or nitrification rate coefficients, if a standard long term BOD test is performed to obtain values
 of these inputs. Furthermore, because many input values  are often obtained by taking differences
 (e.g.CBOD = Total BOD - NBOD) or by regression analyses (e.g. fitting first order kinetic expressions
 to obtain estimate of ultimate CBOD and CBOD rate coefficient), the sign of these correlation
 coefficients is usually negative.  Thus the calibration correlation matrix has many more negative
 entries than the forecast correlation matrix.

        The values of the correlation coefficients used in this study are similar to the set of input
 correlations was developed by Brown (1990) for the uncertainty analysis simulations with the Streeter-
 Phelps model. Because the kinetic formulations and model structure in QUAL2E are not exactly the
 same as those in the Streeter-Phelps model, these values were assigned to the input parameter or
 variable in QUAL2E that most closely matched the inputs in the analytical model (e.g. NBOD in the
 Streeter-Phelps model was matched to ammonia point load in QUAL2E). Every attempt was made to
 keep the correlation matrices as close  as possible to those of the analytical simulation in order to
 provide a common basis for comparing the  results of the two studies.

        Finally, it should be noted that these correlation coefficient assignments are subjective and are
 likely to be refined with further study.  However, in the absence of extensive, replicate field data, the
 two sets of values in this study are believed to represent a reasonable starting point for assessing the
 effects of input parameter correlation on the precision of model predictions.

                         CASE STUDY - WITHLACOOCHEE RIVER

        The first data set used in this study  of input parameter correlation  was obtained from a
 USEPA survey of the Withlacoochee River  during October 1984 (Koenig, 1986).  The river is
 subjected to both municipal and industrial waste loads and there is a significant accretion of flow from
 groundwater inputs (50%). The river has a  uniform low slope, but is characterized by alternating
 shoals and pools (occasionally in excess of 25 ft deep).  Average depths during the survey periods
 were from 5.2 to 14.8 feet, widths ranged from 90 to 140 feet, and flows varied from 150 cfs at the
 headwater to 660 cfs at the end of the  system.  Water quality is affected by algal activity resulting
 from municipal waste discharges  above the section of stream studied. The addition of industrial waste
 at river mile (RM) 24 dramatically reduces tight penetration to the extent that the algal population
 diminishes rapidly in the downstream direction.

        Ten state variables were simulated in this study; temperature, dissolved oxygen (DO),
 carbonaceous BOD (CBOD), four nitrogen forms  (organic, ammonia, nitrite, and nitrate), two
phosphorus forms (organic and dissolved), and algae as chlorophyll-a.  A summary of the calibrated
inputs and their relative standard deviations (standard deviation divided  by the mean) is shown in
Table 2. The calibrated values in general were obtained by adjusting field and/or laboratory
measurements of the model inputs.  The variance estimates were computed from replicate data  taken

                                            200

-------
during the survey period and by inference from other published data. (McCutcheon, 1985 and Tetra
Tech, 1985). Detailed results of the uncertainty analysis without correlation is presented elsewhere
(Brown, 1987).
               Table 2.  Summary of Input Data for QUAL2E-UNCAS Simulations
                              Withlacoochee River Survey 1984
       Input Parameter or
          Coefficient

Hydraulic Data (7)*
   Flows (cfs)
   Depths (ft)
   Velocities (fps)
   Others

Reaction Coefficients (8)*
   CBOD Decay (I/day)
   Reaeration (I/day)
   SOD (gm/tf-day)
   N, P, Algae

Algae, Nutrient, Light Coefficients (17)*
   Maximum Growth Rate (I/day)
   Respiration Rate (I/day)
   Others

Climatology, Temperature Variables (23)*
   Wet, Dry Bulb Air Temps (C)
   Temperature Coefficients
   Others

Headwater, Incremental, Point Loads (27)*
   DO, Temperature
   CBOD, N, P, Algae
Calibrated (Mean)
     Values
    150 - 660
    5.2 - 14.8
     .1 -.8
       a,b
    .04 - .10
    .08 - .80
    .04-.13
       a,b
       1.3
       .15
       a,b
   18.0 - 23.5
   1.00 -  1.083
       a,b
       a
       a
Relative Standard
 Deviation (%)
      3%
      8%
      10%
   10 - 20%
     15%
      13%
     20%
    15 - 25%
      10%
      10%
      10%
      2%
      3%
    1 - 15%
    1 -  3%
    8 - 25%
 ()* Value indicates number of inputs in each category.
 (a) Basin specific values from Koenig, 1986.
 (b) Typical values from Table III-3, Brown and Barnwell, 1985
                                          201

-------
        The QUAL2E-UNCAS model provides uncertainty output for all simulated variables.
However, only the results for dissolved oxygen will be discussed in this paper.  A plot of observed
and modeled dissolved oxygen concentrations is shown in Figure 1.  There is a pronounced minimum
in the DO profile (sag point) a few miles downstream from the industrial waste discharge.  The
dilution from the spring water at RM 12 doubles the flow and depresses the dissolved oxygen only
slightly because of its lower DO content. The results of the uncertainty analysis using the forecast and
calibration mode correlation matrices is presented in Table 3 for both first order error analysis (FOEA)
and monte carlo simulation (MCS).  The results are reported at three locations: RM 26, near the
beginning of the river section; RM 20 at the dissolved oxygen sag point below the industrial waste
discharge; and at RM 2 near the end of the section studied.
   Inputs

Independent
Correlated -
 Forecast Mode
Correlated -
 Calibration Mode

Ratios
Forecast / Indep
Calibration / Indep
                               Table 3.  Effect of Input Correlation
                      Simulated Dissolved Oxygen Standard Deviation (mg/L)
                            Withlacoochee River Basin - 1984 Survey
First Order Error Analysis

RM26  RM20    RM 2

.18     .27    .30

.18     .34    .32

.17     .26    .30
1.00
.94
1.26
.96
1.07
1.00
                     Monte Carlo Simulation

                     RM 26 RM 20 RM 2

                     .18    .28     .31

                     .18    .34     .34

                     .18    .27     .31
1.00
1.00
1.21
.96
1.10
1.00
        A number of important observations can be made from examining the results in Table 3.
First, the simulated DO standard deviation using the forecast mode correlation matrix is larger than
that from the base case analysis which assumes independence among the inputs. At RM 20, for
example, the increase is the largest, and is from .27 to .34 mg/L (26%) for FOEA and from .28 to .34
mg/L (21%) for MCS.  The increases are smaller in the upper reaches because the inputs with strong
correlation coefficients (e.g. the loads) do not have a strong impact on the simulated dissolved oxygen.
Likewise, the increases are diminished near the end of the system because the DO is recovering from
the effects of the upstream industrial waste input.

        The second observation is that the effect of input correlation using the calibration mode matrix
is to decrease the simulated DO standard deviation relative to the case of independent inputs.  The
magnitude of the decrease averages about 0.01 mg/L or  about 5%, and is  not as great as the increase
computed in the case of the forecast correlation matrix.  The positive and negative correlation
coefficients in the calibration scenario apparently tend to offset one another, resulting in only a small
net change (decrease) in the magnitude of the simulated DO standard deviation.
                                           202

-------
8.0 -i
Q 7.0 -
I ":
c 5-0-
O
2 4.0-
X
0
T>
2
1 2'° ~
o *
Qo -
3

T Withlacoochee River

B] ' '
1 RM 2
RM 20
jMox
Industrial Spring .„.,
Waste Dilution " Ave
i i _L
V V
0 25 20 15 10 5 0
River Location (mile)
Figure 1  - Observed and QUAL2E Simulated Dissolved Oxygen Concentrations

Withlacoochee River Survey 1984.
        o>
        E
        c
        >.
        X
       O
        (D
       _>

        O
        w
        VI
8.0  -


7.0  -_


6.0  ^


5.0  -_


4..Q  f


3.0  {


2.0  -_


1.0  -
            0.0
                                   Ouachlta  River
                                                  RM 120
               230  220  210  200  190  180  170 160 150 140 130 120 110

                         River Location  (mile)
Figure 2  - Observed and QUAL2E Simulated Dissolved Oxygen Concentrations

Ouachita River Survey 1980.
                                 203

-------
        Thirdly, the DO standard deviations from the monte carlo simulations are 3-5 % larger than
 those from the first order error analysis. This observation suggests that model non-linearities are not
 severe for these simulations, which is consistent with prior findings (Brown, 1985 and 1987).  It is
 interesting to note that this difference is about the same magnitude as the error in estimating the DO
 standard deviation from 1000 monte carlo simulations as reported by Brown (1987).

        These findings from the Withlacoochee case study are generally supportive of those by Brown
 (1990) using similar correlation matrices and the  hypothetical data sets of tiie Streeter-Phelps  model.
 The major difference is in the  magnitude of the effects. The increases using the forecast mode matrix
 and the decreases using the calibration mode matrix are smaller by about one-half in the QUAL2E-
 UNCAS simulations than in the Streeter-Phelps ones. This attenuation is likely attributable to the
 averaging effects of a large multi-reach simulation characteristic of QUAL2E applications.

                             CASE STUDY -  QUACfflTA RIVER

        The second data set is from an NCASI survey of the Ouachita river during  July 1980 (NCASI,
 1982).  The length of river studied is a navigational pool located between two Corps of Engineers lock
 and dams. The section is just  over 100 miles in length and receives waste discharge from a variety of
 industries and a number of municipalities.  The summer of 1980 was a particularly  dry period, with
 little or no precipitation occurring in the basin during the two months prior to the survey. Thus flows,
 while controlled by the lock and dams,  was low and relatively steady. Time of flow in the pool was
just over two weeks.

         All but one of the 11 discharges and tributaries are located in the upper two-thirds of the
 basin.  Average depths during  the survey period were from 15 to 45 feet, and widths varied from 380
 to 515 feet. How ranged from 2450 cfs at the  head of the section to 3170 cfs at the end. The river
 has a very low uniform slope,  with gradual meanders.  Velocities were low, averaging no more than
 0.25 fps.  Mixing action of large eddy roll cells and stream meanders keep the water column fairly
 well mixed. This was confirmed by periodic sampling traverses of the stream cross section as
indicated by the range of DO concentrations shown in Figure 2. The river water quality is affected by
 algal activity, but light penetration into the turbid water is poor, limiting  this activity to the upper foot
or so of the water column.

        Six state variables were simulated in this  study; DO, CBOD,  three nitrogen forms (ammonia,
nitrite, and nitrate),  and algae as chlorophyll-a.  A summary of the calibrated inputs and their  standard
deviations is given in Table 4.  The calibrated values and variance estimates were obtained in a
manner similar to that used in  the  Withlacoochee  study. Detailed results of the uncertainty analysis
without correlation has been discussed by Brown  (1985).

        As with the Withlacoochee example, only the dissolved oxygen uncertainty  results will be
evaluated.  A plot of the observed and modeled dissolved oxygen concentrations is shown in Figure 2.
Note that there is a steady decline in the DO concentration, with little evidence of a recovery from the
applied waste loads. The results of the uncertainty analysis using the forecast and calibration mode
correlation matrices is presented in Table 5 for  both first order error analysis and monte carlo
simulation. Again, the results are  reported at three locations in the basin; near the headwater at RM
220, in the middle of the modeled section at RM  172, and at the downstream end of the river  at RM
120.
                                            204

-------
                Table 4. Summary of Input Data for QUAL2E-UNCAS Simulations
                                 Ouachita River Survey 1980
       Input Parameter or
          Coefficient

Hydraulic Data (4)*
   Flows (cfs)
   Depths (ft)
   Velocities (fps)
   Others

Reaction Coefficients (10)*
   CBOD Decay (I/day)
   Reaeration (I/day)
   SOD (gm/tf-day)
   N, Algae

Algae, Nutrient, Light Coefficients (11)*
   Maximum Growth Rate (I/day)
   Respiration Rate (I/day)
   Others

Climatology, Temperature Variables (10)*
   Temperature Coefficients
   Others

Headwater, Point Loads (14)*
   DO, Temperature, CBOD
   N, Algae
Calibrated (Mean)
     Values
   2450 - 3170
     15-45
    .15 - .45
       a,b
    .01 - .02
    .02 - .20
    .07 - .14
       a,b
       2.5
       .13
       a,b
   1.00 - 1.047
       a,b
       a
       a
Relative Standard
 Deviation (%)
      6%
      5%
      10%
   10 - 20%
     20%
      12%
      20%
    10 - 20%
      10%
      10%
    5 -  10%
      3%
     2-5%
    2 - 10%
    10 - 20%
  ( )* Value indicates number of inputs in each category.
  (a) Basin specific values from NCASI, 1982.
  (b) Typical values from Table III-3, Brown and Barnwell, 1985
       Inspection of the results in Table 5 show a pattern remarkably similar to that obtained for the
Withlacoochee River. Using correlated inputs in a forecast mode increased the DO standard deviations
about the same amount (3 - 28%). There is also  a clear pattern of the correlation effect increasing as
the DO gets lower and closer to zero. The decrease in DO standard deviation from using the
calibration mode correlations is also repeated in this case study.   While smaller than the increase with
the forecast mode, it is almost twice as large as that found in the Withlacoochee example.
                                          205

-------
                               Table 5. Effect of Input Correlation
                      Simulated Dissolved Oxygen Standard Deviation (mg/L)
                               Ouachita River Basin - 1980 Survey
  Inputs
First Order Error Analysis

RM 220    RM 172   RM 120
Independent .32
Correlated -
Forecast Mode .32
Correlated -
Calibration Mode .32
Ratios
Forecast / Indep 1.00
Calibration / Indep 1.00
.38
.45
.37
1.18
.97
.71
.91
.67
1.28
.94
Monte Carlo Simulation

RM 220    RM 172   RM 120

.32       .41       .70

.33       .46       .81,

.33       .38       .65
                                                               1.03
                                                               1.03
                                                      1.12
                                                      .93
                    1.16
                    .93
calibration mode correlations is also repeated in this case study.  While smaller than the increase with
the forecast mode, it is almost twice as large as that found in the Withlacoochee example.

        The primary difference between the two case studies is the smaller DO standard deviations
predicted in the Ouachita monte carlo simulations at RMs 172  and 120 with the forecast mode matrix,
and at RM 120 with the calibration mode matrix.  This observation is contrary to most QUAL2E
applications which have shown that monte carlo simulation produces larger standard deviations than
first order analysis.  The reason for this  unexpected finding in this case study is not clear. In two
instances,  the differences are small (less than 3%), and thus can be attributed to the error in estimating
DO standard deviations from only 1000 monte carlo simulations. (This factor may also explain the
3% increase in simulated DO standard deviation at RM 220 when using the calibration mode
correlation matrix.)  In the remaining instance (RM 120 -forecast mode), the discrepancy occurs  at a
low DO (less than 2 mg/L) and after a long travel time (greater than 20 days).  The difference is over
10%,  and  may indicate an important, yet unidentified characteristic of uncertainty analysis under these
extreme conditions.

                                       CONCLUSIONS

        In general, the results from the QUAL2E-UNCAS simulations compare favorably with those
reported by Brown (1990) for the Streeter-Phelps simulations.  Input parameter correlation was found
to both  increase and decrease the dissolved oxygen standard deviation, depending on the form of the
parameter correlation matrix used in the uncertainty simulation. The impact of correlated inputs in a
forecast mode was to increase the simulated dissolved oxygen standard deviation on average about 15
percent  (range 0 to 30%).  Accounting for correlation among input parameters in a calibration mode
decreases  the simulated dissolved oxygen standard deviation on average about 5% percent (range -3 to
7%).  The effect of  input parameter correlation estimated in the QUAL2E-UNCAS case studies,  while
similar in  sign, is  about one-half that determined for a hypothetical, single reach Streeter-Phelps
example (Brown,  1990).  In all simulations,  except those with long travel times (greater than 20  days),
monte carlo analysis showed slightly larger (2-6%) output standard deviations for dissolved oxygen
than those from first order error analysis. Thus the effect of model non-linearity on the results of
dissolved oxygen  uncertainty analysis is not believed to be severe for these simulations.
                                           206

-------
                                SELECTED REFERENCES

 Benjamin, J.R., and C.A. Cornell, (1970), Probability, Statistics and Decision for Civil Engineers,
        McGraw-Hill, New York, NY.
 Brown, L.C. (1985), "Uncertainty Analysis Using QUAL2E", Proceedings, Conference on Stormwater
        and Water Quality Management, December 5-6, 1985, E.M. and W. James, ed., pp. 125-140.
 Brown, L.C. (1987), "Uncertainty Analysis in Water Quality Modeling Using QUAL2E", in Systems
        Analysis in Water Quality Management, M.B. Beck, ed., Advances in Water Pollution Control,
        Pergamon Press, London, UK., pp. 309-319.
 Brown, L.C. (1990), "Effect of  Correlated Inputs on Water Quality Models Uncertainty", a paper
        presented at the 2nd International Symposium on Fish Physiology, Toxicology and Water
        Quality Management, USEPA, Davis ,CA, Sept 15-17.
 Brown, L.C. and T.O. Barnwell, Jr. (1985), Computer Program Documentation for the Enhanced
        Stream Water Quality Model QUAL2E, EPA/600/3-85/065, USEPA, Environmental Research
        Laboratory, Athens, GA.
 Brown, L.C. and T.O. Barnwell, Jr. (1987), The Enhanced Stream Water Quality Model QUAL2E and
        QUAL2E-UNCAS: Documentation and User Manual, EPA/600/3-87/007, USEPA,
        Environmental Research Laboratory, Athens, GA.
-Burgess, S.J. and D.P.Lettenmaier, (1975), "Probabilistic Methods in Stream Quality Management",
        Water Resources Bulletin, AWRA, Vol. 11, No. 1, pp. 115-130.
 Chadderton, R.A., A.C. Miller and A.J. McDonnell, (1982), "Uncertainty Analysis of Dissolved
        Oxygen Model", Journal of the Environmental Engineering Division, ASCE,  Vol. 108, No.
        EE5, pp. 1003-1013.
 Koenig, M., (1986), Withlacoochee River 7 QUAL2E model calibration from Clyatville, GA to
        Ellaville, FL, USEPA, Region IV, Environmental Services Division, Athens,  GA.
 McCutcheon, S.C., (1985), Water Quality and Streamflow Data for the West Fork Trinity River in Fort
        Worth, TX, USGS, Water Resources Investigation Report 84-4330, NTSL, MS.
 NCASI, (1982), A Study of the  Selection, Calibration,and Verification of Mathematical Water Quality
        Models,  Technical Bulletin No.  367, National Council for Air and  Stream Improvement, New
        York, NY.
 NCASI, (1985), A Study of Uncertainty Analysis Techniques and their Applications to the
       .Mathematical Water Quality Model SNSIME, Technical Bulletin No. 463, National Council
        for Air and Stream Improvement, New York, NY.
 Song, Q. and L.C. Brown, (1990), "DO  Model uncertainty with Correlated inputs", Journal of
        Environmental Engineering, ASCE, Vol. 16, No.6, PP. 1164-1
 Tetra Tech, Inc.,  (1985), Rates,  Constants, and Kinetics Formulations in Surface Water Quality
        Modeling, 2nd ed., EPA/600/3-85/040, USEPA, Environmental Research Laboratory, Athens,
        GA.
                      • U.S. GOVERNMENT PRINTING OFFICE: 1994-5 50-00 1 / 00 184
                                          207

-------

-------