EPA/600/R-93/157
August 1993
FISH PHYSIOLOGY, TOXICOLOGY, AND
WATER QUALITY MANAGEMENT
Proceedings of an International Symposium
Sacramento, California, USA
September 18-20,1990
Edited by
Rosemarie C. Russo and Robert V. Thurston^
Environmental Research Laboratory
U.S. Environmental Protection Agency
Athens, Georgia 30605-2720
Fisheries Bioassay Laboratory
Montana State University
Bozeman, Montana 59717
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30605-2720
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 recommen-
dation for use by the U.S. Environmental Protection Agency.
u
-------
FOREWORD
In recognition of the importance of international cooperation in environmental research,
the United States and the People's Republic of China have agreed to a Protocol for
Environmental Protection. Under Annex 3, Item 4 of the USA-PRC Protocol, cooperative
research is jointly conducted by U.S. and PRC researchers in several important environmental
areas. These include the environmental processes and effects of pollution on freshwater
organisms, soils, and groundwater and on the application of transport and transformation models
for use in water quality management decisions. Activities under this component of the Protocol
include seminars, workshops, joint symposia, training programs, joint research, and information
exchange.
Symposia are an effective means of fostering cooperation among scientists from different
countries as environmental organizations seek to acquire the scientific knowledge necessary to
predict the effects of pollutants on ecosystems and apply the results on a global scale. The
symposia provide a forum through which scientists and engineers from laboratories and institutes
can exchange scientific expertise on environmental problems of concern to EPA and the
international environmental community.
This document is the proceedings of the Second International Symposium convened under
this portion of the Protocol and held on September 18-20, 1990, in Sacramento, California, USA
The Symposium was sponsored jointly by the U.S. Environmental Protection Agency, the Chinese
National Environmental Protection Agency, the University of California at Davis, the American
Fisheries Society, the Chinese National Science Foundation, and the Chinese Society of
Ichthyology. The Symposium provided a forum for presentation of research findings and
exchange of knowledge on fish physiology, toxicology, and water quality management.
Proceedings of the First International Symposium, held in Guangzhou, PRC, on
September 14-16, 1988, were published as EPA Report EPA/600-9-90/011.
Rosemarie C. Russo, Ph.D.
Director
Environmental Research Laboratory
Athens, Georgia, USA
in
-------
ABSTRACT
Scientists from five countries presented papers at the Second International Symposium on
Fish Physiology, Toxicology, and Water Quality Management, which was held in Sacramento,
California, on September 18-20, 1990. This proceedings includes 21 papers presented in sessions
on the physiological effects of pollutants on fishes, the uptake and depuration of toxicants by
fishes, and water quality management. Papers address reproduction and growth of fishes,
respiratory physiology, bioaccumulation of toxicants, microcosms, ecotoxicology, surface water -
quality including acid mine drainage, metal complexation and xenobiotics, and water quality
models and management strategies.
IV
-------
CONTENTS
Page
FOREWORD ..................................... ......................... jii
ABSTRACT .................... „ ..................... ... .................. jv
ACKNOWLEDGMENT ............. . ....................... ........ . . . ..... . . vii
SESSION I. PHYSIOLOGICAL EFFECTS OF POLLUTANTS
Lin, Hao-Ren and David J. Randall, Chairpersons
Effects of Some Pollutants on Reproduction of Fishes: A Review . . . . .............. 1
Lin, Hao-Ren
Use of Fish to Monitor Toxicity and Accumulative Capacity
of Instream Contaminants ..................... ........ ........ ..... 13
William H. Benson and Jeffrey A. Black
Use of Biomonitoring to Control Toxics in the United States ..... ............ ..... 21
Nelson A Thomas
Effects of Water Pollution on the Growth and Development
of Fishes in the Lower Yangtze River of China . .......... . ............... 31
Yuan, Chuan-Fuh
Ecotoxicology of the Synthetic Detergent LAS in the Aquatic Ecosystem ____ . . ....... 43
Zhang, Yong-Yuan and F. Korte
Effects of Temperature and Hypoxia on California Stream Fishes ..... ...... ..*..... 53
Joseph J. Cech, Jr., Stephen J. Mitchell,
Daniel T. Castleberry, and Maryann McEnroe
Effects of Varying Water pH on Gill Function ..... .......... . . ................ 59
Hong JJn and David J. Randall
Effects of Zinc on Respiratory Physiology of Fish (Tilapia sp.)
and Exploration of Detoxification ....................... ....... ....... 71
Chai, Min-Juan, Zhou, Xue-Cheng, and Huang, Yu-Ling
SESSION II. UPTAKE AND DEPURATION OF TOXICANT
Jin, Hong-Jun and Robert V. Thurston, Chairpersons
Bioaccumulation of Toxic Hydrophobic Organic Compounds
at the Primary Trophic Level ..................... ......... .......... 79
Deborah L. Swackhamer
Effects of a New Insecticide, CCU, on Carp:
Acute Toxicity and Metabolism In Vitro ........ ......................... 87
Xu, Li-Hong, Zhang, Yong-Yuan, Xu, Ying, and Chen, Zhu-An
-------
Respiratory Oxygen Requirements of Fishes: Description of OXYREF,
A Data File of Test Results Reported in the Published Literature 95
Robert V. Thurston and Peter C. Gehrke
Toxicant Uptake Across Fish Gills 109
David J. Randall and Colin J. Brauner
Toxicity and Distribution of Copper in an Aquatic Microcosm
under Different Conditions of Alkalinity and Hardness 117
Jin, Hong-Jun, Zhang, Yi-Min, and Yang, Rong
SESSION III. WATER QUALITY MANAGEMENT
Liu, Jennie Jing-Yi and Rosemarie C. Russo, Chairpersons
Environmental Implications of Metal Complexation 129
Liu, Jennie Jing-Yi
Prediction of Metal Contaminant Exposure in Natural Waters
using Geochemical Equilibrium Modeling 149
Nicholas T. Loux and David S. Brown
Impacts of Acid Mine Drainage on Water Quality of the Lo An River,
Poyang Lake Area, PRC 165
Lin, Yu-Huan
Water Environment Management in China 177
Zhu, Xing-Xiang
Water Quality Projections for Lake Bosten, Xinjiang, PRC 181
Zhang, Guo-An, Zhu, Dong-Wei, Steve C. McCutcheon,
Pei, Xin-Guo, and Zhong Xin-Cai
Effect of Correlated Inputs on Water Quality Model Uncertainty 197
Linfield C. Brown
Water Quality Modeling for a Tidal Harbor:
Application to Zhenjiang Harbor, PRC 213
Zhu, Dong-Wei, Steve C. McCutcheon, Zuo, Yu-Hui,
Yue, Zhen-Hua, and Tim A. Wool
Water Resource Management Strategies for Restoring and Maintaining
Aquatic Life Uses 227
Thomas Willingham and Allen Medine
List of Participants : 243
VI
-------
ACKNOWLEDGMENTS
Thanks are due to the sponsoring organizations for this Symposium: National
Environmental Protection Agency (China), National Science Foundation (China), Society of
Ichthyology (China), U.S. Environmental Protection Agency (USA), University of California at
Davis (USA), American Fisheries Society (USA). The research scientists/engineers and
environmental managers who participated in the Symposium are deserving of primary
recognition. Special appreciation is accorded to Dr. David J. Randall for arranging and
coordinating the international aspects and to Dr. Joseph J. Cech, Jr., for arranging and
coordinating the local activities. Thanks are expressed to Ms. Martha Mann and Mr. Robert
Ryans for their assistance in preparation of the Proceedings.
Vll
-------
Vlll
-------
EFFECTS OF SOME POLLUTANTS ON REPRODUCTION
OF FISHES: A REVIEW
by
Lin, Hao-Ren1
INTRODUCTION
The rapid growth of industrial activity and the modernization of agriculture has resulted
in widespread contamination of water, due to indiscriminate use of pesticides and fungicides in
agriculture and discharge of industrial wastes and factory effluents into rivers, lakes and ponds.
One of the serious consequences is that aquatic animals, including fishes, become the victims of
water pollution. An important reason for the large scale mortality of fishes and elimination of
fish populations in natural water is the increased concentrations of pollutants and their/
detrimental action on fish growth and reproduction (Saunders 1969).
In recent years the influences of environmental pollutants on the physiological control of
fish reproduction have become a subject of great interest. In fishes, like all other vertebrates,
the nervous and endocrine systems act in concert to coordinate reproduction. Major links in the
chain of events leading from the perception of environmental stimuli to the release of gametes
in fish are shown in Figure 1. The reception of such environmental stimuli is mediated by the
nervous system and involves the passage of information from sensory receptors to the brain.
This neural information, upon reaching the hypothalamus, regulates the gonadotropin (GtH)
secretion into the general circulation from the pituitary gland by a dual control, with
gonadotropin-releasing hormone (GnRH) stimulating GtH release and inhibition by dopamine,
which acts as a gonadotropin release-inhibitory factor (GRIP) on the action of GnRH as well as
spontaneous release of GtH (Peter et al. 1986). The target organ of GtH is the gonad. Its
effect is to stimulate the production of sex steroids in the gonads; these, then, are responsible
for the maturation and release of gametes. This is so-called hypothalamo-hypophyseal-gonadal
axis (Harvey and Hoar 1979).
Toxicity tests with various pollutants on reproduction of fishes have been performed by
various,laboratories around the world in the last 10-20 years. The present paper reviews the
works dealing with the effects of depressed pH, heavy metals, and pesticides on fish
reproduction, with special reference to their effects on the endocrine physiology of the
hypothalamo-hypophyseal-gonadal axis.
EFFECTS OF POLLUTANTS ON THE HYPOTHALAMO-HYPOPHYSIAL COMPLEX IN
RELATION TO REPRODUCTION
Recently, Ram and co-workers (1986, 1988) reported the effects of mercury on the
hypothalamo-hypophysial complex in relation to reproduction in Channa punctatus (Bloch).
Department of Biology, Zhongshan University, Guangzhou, PRC.
1
-------
Environmental Factors
Brain
Hypothalamus
GnRH
GRIP
(Dopamine)
Pituitary
GtH
Follicle Cell of
Oocyte
MIS
Oocyte Surface
| MPF
Germinal Vesicle
1
Oocyte Maturation
Ovulatton
Figure 1. Major links in the physiologic chain of events leading from the
reception of environmental stimuli to the release of mature oocytes.
MIS = maturation inducing steroids; MPF = maturation promoting factor;
GtH = gonadotropin; GnRH = gonadotropin releasing hormone;
GRIP = gonadotropin release-inhibitory factor.
Fish in groups I and n were exposed to "safe concentrations" of 0.01 ppm of inorganic mercuric
chloride (HgCl2) and 0.2 ppm of an organic mercurial fungicide, Emisan (methoxy ethyl
mercuric chloride; MeEHgCl). The untreated group m served as the control. After 6 months
treatment, in the control fish the nucleus preopticus (NPO) neurons of hypothalamus were large
and actively secreting, characterized by their prominent rounded nuclei and median nucleoli with
an adequate quantity of neurosecretory material (NSM) in the perikarya. In fish of both
-------
experimental groups, the neurons were smaller, containing a scanty quantity of NSM and
exhibiting various degrees of degeneration. In MeEHgCl-treated fish, the degenerative changes
were more pronounced, as most of the neurons were pyknotic and necrotic. Corresponding with
nuclear changes in the neurons, the brain monoamine oxidase (MAO) activity was also
significantly inhibited in groups treated with HgCl2 (1.782 ± 0.105, p < 0.05) and MeEHgCl .
(1.808 ± 0.106, p < 0.05) as compared with controls (2.051 ± 0.132). In the control fish, the
proximal pars distalis (PPD) of the pituitary was dominated by large, actively secreting,
hypertrophied, vacuolated gonadotrophs; their ovaries were fully ripe with majority of vitellog-
enic matured oocytes. In the testis, seminiferous tubules were filled with sperm masses, and
Leydig's cells were large with prominent rounded nucleus and nucleolus. In the pituitaries from
fish of both experimental groups, the gonadotrophs were small, inactive, involuted, and fewer in
number; gonadal maturation was significantly inhibited, the ovaries were in immature state I,
totally devoid of vitellogenic oocytes; in the testis, sperms were lacking and the Leydig's cells
were inactive and atrophied. These results suggest that mercurial-induced inhibition of the
gonadal maturation may be mediated through the impairment of the hypothalamo-hypophysial-
gonadal axis.
Katti and Sathyanesan (1986) also reported that exposure of Clarias batrachus (L.) to 5 ppm
of lead nitrate for 150 days caused marked accumulation of NSM in the anterior
neurohypophysis; the neurons of both nucleus preopticus and nucleus lateralis tuberis (NLT)
showed degenerative changes; the ovaries were in immature stage, totally devoid of vitellogenic
oocytes; in the testis, sperms were lacking and the Leydig's cells were involuted. These
observations clearly indicate that lead significantly influences the hypothalamo-hypophysial axis
by adversely affecting both NPO and NLT neurons, then, inducing dysfunction of the gonads;
the accumulation of NSM in the anterior neurohypophysis suggests the inactivity or blocking of
transport.
Singh and Singh (1981, 1982a, b) investigated the site and route of actions of four pesticides,
viz. Cythion (malathion), Paramar M50 (methylparathion), Hexadrin (endrin), and aldrin on
hypothalamo-hypophyseal-ovarian axis in the freshwater catfish, Heteropneustes fossilis (Bloch).
The observations revealed that: (1) Hypothalamic extracts drawn from the fishes exposed to
these four pesticides showed significantly low amounts of the GnRH-like factor; (2) Exposure of
fishes to these four pesticides decreased the gonadotrophic potency of the pituitary gland
significantly; (3) Exposure of sham-hypophysectomized fishes to these four pesticides decreased
the ovarian 32P uptake significantly; in hypophysectomized fishes, Cythion and Paramar M50 had
no effect on ovarian 32P uptake, while aldrin and Hexadrin were still very potent in inhibiting
ovarian 32P uptake. Thus it seems that Cythion and Paramar M50 act through the hypothalamus
where they inhibit the secretion of GnRH, which in turn decreases synthesis and release of
gonadotropin from the pituitary gland followed by reduced ovarian activity. Aldrin and
Hexadrin suppressed ovarian 32P incorporation even in hypophysectomized fish, and therefore
appear to have two effective routes of action for suppressing ovarian activity, one like that of
Cythion and Paramar M50 in the hypothalamo-hypophyseal-ovarian axis, and the other involving
direct action on the ovary. On the basis of these findings, it was concluded that these pesticides
probably interfere with gonadotropin secretion with resultant decrease in ovarian activity during
all phases of the annual reproductive cycle in H. fossilis.
EFFECTS OF POLLUTANTS ON STEROIDOGENESIS
The level of 3-p-HSD(hydroxy-steroid dehydrogenase) along with that of various other
steroid dehydrogenases is indicative of steroidogenesis in the gonadal tissue. Kaput et al. (1978)
-------
observed that treatment with the pesticide fenitrothion caused a great reduction in the activity of
3-p-HSD in the testicular and ovarian extracts of common carp (Cyprinus carpioi). The activity
of 3-P-HSD was reduced to about 80% and 71.1% in the case of males and about 89.6% and
83.8% in the case of females, as compared to the controls, when treated with safe and sublethal
dosages, respectively. Recent studies on fish have indicated that the level of 3-P-HSD may be
related to the level of gonadotropin in the circulating system (Nagahama et al. 1982); hence, the
decline in the level of 3-P-HSD as a result of the fenitrothion treatment may be caused by the
action of this pesticide on the pituitary by inhibiting the release or synthesis of the gonadotropin.
Truscott et al. (1983) found that the effect of acute exposure to crude petroleum on the
plasmatic concentration of the sexual steroids varied according to the state of gonad
development. The concentrations of total (free and conjugated) androgens (testosterone,
11-ketotestosterone) in plasma of sexually mature male salmon and flounder were significantly
lower in petroleum-exposed fish (30 ml petroleum/270 L/day); petroleum exposure had no
significant effect on levels of total plasmatic androgen or estradiol in male salmon and male or
female flounder during gonadal recrudescence.
The long-term ecological consequences of selenium pollution of freshwater environments
have not been elucidated. However, selenium accumulation can be associated with decreased
reproductive success of feral fish (Cumbie and Van Horn 1978). Recently, comparisons were
made of the accumulation of selenium and reproductive status of redear sunfish (Lepomis
microlophus) collected in July 1986 from Martin Lake (a contaminated site) and Lake Tyler
(a reference site) (Sorensen 1988). The ovaries of mature fish collected from Martin Lake (with
higher selenium accumulation) frequently had atretic follicles, abnormally shaped follicles,
asynchronous oocyte development, and an overall reduction in the number of developing
oocytes. These histopathological changes were not accompanied by alterations in sexual steroid
level in the blood. However, most of the males collected from the contaminated site were
immature and had lower circulation levels of steroid hormones than did reference males. These
results indicate that the reproductive status of selenium-contaminated fish collected from Martin
Lake is still seriously impaired.
Heavy metals such as cadmium inhibited'the biosynthesis of 11-ketotestosterone from the
radioactive precursor and the conversion of exogenous pregnenolone to steroid metabolites in
the testis of brook trout (Salvelinus fontinalis) (Sangalang and O'Halloran 1973). These results
give conclusive evidence that cadmium directly affects testicular steroidogenesis in vitro.
Freeman and Sangalang (1985) investigated the effects of an acidic river, caused by acidic
rain, on steroidogenesis and reproduction in the Atlantic salmon (Salmo salar). Wild male
Atlantic salmon captured in the low pH (4.7) Westfield River, Nova Scotia, Canada, had lesser
plasma androgen levels at sexual maturity compared to levels in wild male salmon sampled in
the nearby less acidic Medway River. In addition, Atlantic salmon held in cages and fed daily in
the Westfield River (pH range 5.1 to 5.3) during the last 3 months of their sexual maturation
cycle gained less weight, produced smaller eggs, and had abnormal steroid hormone metabolism
compared to similar fish held in the less acidic Medway River (pH range 5.4 to 6.1). These
observations indicated that the acidic river water can alter the biosynthesis of critical steroid
hormones involved in the reproductive process. Consequently, the steroidogenesis tests, applied
to fish to detect changes in steroid hormone synthesis, have great promise in providing an early
warning of pollution problems in water before they become acute, and their effects irreversible.
-------
EFFECTS OF POLLUTANTS ON GONADAL DEVELOPMENT AND MATURATION
Kumar and Pant (1988) investigated the effects of exposure of three pesticides
(methoxychlor, aldrin, and monocrotophos) on the ovarian development in Puntius conchonius.
The results showed that all three pesticides were severe poisons to the ovaries of P. conchonius
even in sublethal concentrations. There was a sharp and progressive increase in the population
of atretic oocytes with the 4-month treatments. These studies also indicated that the
organophosphate pesticides, which were generally considered to be milder than organochlorines,
were also hazardous to fish reproduction over long exposures. Kumar and Mukherjee (1988)
reported that when sexually maturing female common carp were exposed for 30 days to
sublethal concentrations of phenol or sulfide, the GSI (gonadosomatic index) was reduced
significantly, suggesting the inhibitory effects of these pollutants on the development and
maturation of the ovary.
It has been shown that sublethal concentrations of heavy metals that did not affect the
survival and growth of fish over a given period although they were capable of impairing
reproduction (Brungs 1969, Spehar 1976). Kumar and Pant (1984) have observed the most
common industrial aquatic pollutants in the freshwater teleost P. conchonius. After 3 or 4
months of treatments, all three metals (zinc, copper, lead) induced significant atresia in the
ovary; zinc damaged mainly the younger oocytes, whereas copper and lead were more effective
on relatively older oocytes. The female Lebistes reticulatus when exposed to sublethal
concentrations of zinc sulfate (278-300 mg/L) also displayed an increasing percentage of atretic
oocytes, reducing the percentage of mature oocytes, pyknotic nuclei, and numerous small
vacuoles in cytoplasm (Sehgal and Saxena 1986). Similarly, the male L. reticulatus treated with
sublethal concentration of zinc exhibited elaborate vacuolization in spermatocytes, significant
reduction of spermatids and mature sperms, and remarkable increase in the percentage of
atretic spermatophores (Sehgal and Saxena 1986). Recently, Kirubagaran and Joy (1988) have
investigated the toxic effects of mercury ovarian recrudescence in the catfish, Clarias batrachus.
After 90 and 180 days treatment with different mercurial compounds (HgCl2, CHsHgCl, and
methoxyethyl mercury chloride), ovarian recrudescence was completely arrested and the oocytes
were in a non-vitellogenic stage, suggesting that mercury impairs vitellogenesis regardless of its
chemical nature.
The effects of low pH on the oogenesis of fish are most clearly evident by the absence of
fully matured oocytes. The reproductive index calculated from the final developmental stage 6
oocyte of flagfish, Jordanella floridae, showed that the production of fully mature, oocytes capable
of being fertilized was reduced to 20.7 and 15.8% at pH 6.0 and 5.5, and 2.1 and 8.2% at pH 5.0
and 4.5 (Ruby et al. 1977). This reduction was primarily produced by the loss of ability of
oocytes to deposit secondary yolk within the cytoplasm. At pH 4.5, both primary and secondary
yolk depositions were severely affected. Histological ovarian indices (HOI), an index of
interrupted egg development, was developed based on the percentage by volume of preovulatory
corpora atretica (POCA) when compared to the total volume of all oocyte stages present in the
ovary; since the diagnosis of reproductive impairment is based on elevated atretic oocyte
volumes, HOI values of 20% or greater usually signal reproduction failure. More recently,
McCormick et al. (1989) investigated the effects of environmental acidification on oocyte atresia
and reproductive success in fathead minnow (Pimephales promelas,) and found that elevated
HOI value was related to pH and sampling date (the stage of reproductive cycle). In June, no
fish exhibited elevated HOI values at any pH. However, in July and August, the oogenic
process had been in progress long enough so that oocyte atresia was exhibited among maturing
oocytes; fish with elevated HOI values (greater than 20%) were found only in pH 6 and 5, and
those in pH 5 were higher than in pH 6. In September, several weeks after cessation of
-------
spawning, the HOI value of fish at any pH had returned to prespawning values. These results
indicated that the oocyte atresia and the final reproductive success were directly associated with
the environmental acidification.
Weiner et al. (1986) reported the effects of low pH on the gametogenesis of rainbow
trout (Oncorhynchus myJdss'). Adults (parents) were exposed to water of pH 4.5, 5.0 or 5.5
during the final 6 weeks of reproductive maturation; control acidities were pH 6.5-7.1. Progeny
of cross breeding between control males and acid-exposed females had reduced survival through
development, hatching and yolk-sac absorption, demonstrating that oogenesis is sensitive to
acidic conditions. Similar reductions in the survival of the progeny of acid-exposed males and
control females indicate the sensitivity of spermatogenesis to low ambient pH.
EFFECTS OF POLLUTANTS ON OVULATION AND SPAWNING
Effects on ovulation and spawning, or on the number of eggs produced, are also known
for pesticides in general. Carlson (1971) reported that after a long-term exposure (9 months) at
the 0.68 mg/L concentration of the carbamate insecticide carbaryl (Sevin), the mean number of
eggs produced per female of the fathead minnow, as well as the mean number of eggs per
spawning, were significantly less than in the control, and no hatching occurred. This study
demonstrated that a concentration of 0.68 mg/L carbaryl adversely affects spawning of fathead
minnows. Also, diazinon, an organophosphate insecticide, employed in continuous flow-through
tests with sheepshead minnow (Cyprinodon variegatus) caused a reduction of up to 55% in the
number of eggs produced per female at concentration higher than 0.47 p-g/L in seawater (Von
Westernhagen 1988). Similar effects were caused by polychlorobiphenyls (PCBs) when applied
orally (Phaxinus phoxinus; Bengtsson 1980) or in solution (P. promelas; Nebeker et al. 1974) at
low (1.8 £tg/L) doses. Among the hydrocarbons, benzene, a toxic component of petroleum, is
known to be very active in reducing the number of viable ovarian eggs of fish. When female
pacific herring (Clupea pallasi) were exposed to low levels (ppb) of benzene for 48 h just prior
to spawning, a significant reduction in survival of ovarian eggs was recorded in the range of
10-25% (Struhsaker 1977).
Exposure of mature fish to low levels of zinc, cadmium, copper, or mercury may lead to
significant reductions in eggs produced. Adult zebrafish (Brachydanio rerio), when held in water
containing a sublethal concentration (5 ppm) of zinc for a 9-day period in which the gametes
were maturing, showed a delay in spawning; when spawning did occur, the experimental pairs of
fish produced an average of 165 eggs of which only 21.1% were viable. In contrast, control pairs
of fish produced an average of 434 eggs of which 90.2% were viable (Speranza et al. 1977). A
reduction in the number of eggs spawned by the minnow P. phoxinus (up to 21%) after exposure
to zinc at 0.13 and 0.2 mg/L (Bengtsson 1974). In a chronic test (100 days exposure), spawning
of flagfish was reduced at zinc concentrations of 85 and 51 /ig/L, and was inhibited at cadmium
concentrations of 8.1 ju.g/L (Spehar et al. 1978). Cadmium and copper at low concentration
(Cd, 0.6-60 /ig/L; Cu, 3.7-31 ^.g/L) caused progressively decreasing spawning activity and egg
number spawned per female among fathead minnows (Eaton 1973). During a 22-month
exposure of bluegill (Lepomis macrochirus) to copper in soft water, its spawning was inhibited at
a concentration of 162 jj,gfL (Benoit 1975). Pickering et al. (1977) found that prespawning
exposure time to copper (3 or 6 months) had no significant effect on spawning of fathead
minnows; although the number of eggs produced per female decreased with increase in copper
concentration; egg production at copper concentrations of 37 /ig/L and higher was significantly
lower than in the controls.
-------
Impairment of ovulation and spawning in fish is also known to be caused by depressed pH
of the holding water. For example, fathead minnows when kept in water at pH levels between
4.5 and 6.0 produced significantly fewer eggs than in an environment at pH 6.8 (Mount 1973).
Craig and Baksi (1977) reported that after exposure to depressed pH (below 6.0) for 20 days,
egg production, egg fertility and fry growth of flagfish were impaired at all exposure levels; fry
survival was reduced at pH 5.5 and 5.0, and no fry survived at pH 4.5. These results coincided
with the investigations on the desert pupfish, Cyprinodon nevadensis .(Lee and Gerking 1980).
Egg production of this species was significantly reduced at every pH level tested below the
control (pH 6.5); egg-laying virtually ceased at pH 5.0, while egg viability was reduced to less
than 50% of the control value at pH 6.5. Tarn and Payson (1986) investigated the effects of
chronic exposure (from early February to early December) to sublethal pH on egg production
and ovulation in brook trout, and found that at the end of the experiment the body weights of
both male and female fish in pH 5.16 and 4.48 were only 70-77% of the control fish at pH 7.34.
Rapid oocyte development occurred simultaneously over all pH groups in June, suggesting that
the initiation of gametogenesis was not affected over the range of pH tested; however, the
number of eggs produqed was significantly correlated to body weight, consequently the number
of eggs produced by the smaller pH 5.16-4.48 females was reduced. Ovulation was also
significantly delayed in the acidic groups. Beamish et al. (1975) suggested that due to altered
pH, normal calcium metabolism required for successful ovarian maturation (Urist and Schjeide
1961) was altered, resulting in abnormally low calcium concentration in female serum, which in
turn affected the female reproductive physiology.
SUMMARY
The loss of fish populations in polluted natural waters has frequently and primarily been
attributed to reproductive failure and resultant lack of recruitment of new generations into the
population.
The reproductive physiological mechanisms affected by some pollutants (pesticides, heavy
metals, etc.) appear to be mediated through the impairment of the hypothalamo-hypophysial-
gonadal axis. Involvement of hypothalamus in the action of these pollutants has been
confirmed. Some pesticides (e.g., Cythion and Paramar M50) and heavy metals (e.g., mercury
and lead) act through the hypothalamus where they inhibit the secretion of GnRH which in turn
decreases the synthesis and release of gonadotropin from the pituitary gland followed by
reduced gonadal activity. Some other pesticides (e.g., aldrin and Hexadrin) appear to have two
effective routes of action for suppressing gonadal activity: one involving hypothalmo-
hypophysial-gonadal axis and another directly acting upon the gonad.
There is a significant reduction in the level of 3-P-HSD, a principal enzyme involved in
steroid biosynthesis, in the ovary of common carp followed treatment with fenitrothion. This
detrimental effect may be due to the action of this pesticide on the pituitary by inhibiting the
release or synthesis of the gonadotropin. Acute exposure to crude petroleum, or long-term
exposure in the selenium-contained lake and acidic river, can also alter the biosynthesis of
critical steroid hormones involved in the reproduction process. Heavy metals such as cadmium
directly inhibited testicular steroidogenesis in the brook trout in vitro.
Sublethal concentrations of some pesticides and other chemical pollutants are severe
poisons to the ovary of fish and cause a sharp and progressive increase in the population of
atretic oocytes.
-------
Gonads of fishes exposed to "safe" or sublethal concentrations of heavy, metals (zinc,
copper, mercury, lead, etc.) for a long-term remained in the resting condition of the
reproductive cycle and exhibited degenerative tendencies. Correlative histological changes were
also observed in the pituitary; gonadotrophs were inactive and fewer in number.
The effects of depressed pH on the gametogenesis of fish are most clearly evident by the
reduction or absence of fully matured gamates. In female fish, the physiological mechanisms
affected by low pH and restricting egg maturation and production appear primarily to be the
interference of the synthesis of yolk proteins (vitellogenin) and the deposition of yolk within the
oocytes.
Exposure of mature fish to low levels (/ig/L) of pesticides (e.g., carbaryl, diazinon,
polychlorobiphenyl) or heavy metals (e.g., zinc, cadmium, copper, mercury) may lead to
significant interference of ovulation and spawning and reduction of eggs produced. Impairment
of ovulation and spawning in fish is also caused by depressed pH (4.5-6.0) of the holding water.
Detrimental effects on reproduction of fish caused by low level of pollutants present in
the environment necessitate the establishment of water quality control systems and safe levels of
pollutants for fishes.
REFERENCES
Beamish, R.J., W.L. Lockart, J.C. van Loon, and H.H. Harvey. 1975. Long term acidification of
a lake and resulting effects of fishes. Ambio 4:98-102.
Bengtsson, B.R 1974. The effects of zinc on the mortality and reproduction of the minnow,
Phoxinus phoxinus. Arch. Environ. Contam. Toxicol. 2:342-355.
Bengtsson, B.E. 1980. Long term effects of PCB (Clophen A50) on growth, reproduction and
swimming performance in the minnow, Phoxinus phoxinus. Water Res. 14:681-687.
Benoit, D.A. 1975. Chronic effects of copper on survival, growth and reproduction of the
bluegill (Lepomis macrochirus). Trans. Amer. Fish. Soc. 2:353-358.
Brungs, W.A. 1969. Chronic toxicity of zinc to the fathead minnow, Pimephales promelas.
Trans. Amer. Fish. Soc. 98:272-279.
Carlson, A.R. 1971. Effects of long-term exposure of carbaryl (sevin) on survival, growth and
reproduction of the fathead minnow (Pimephales promelas). J. Fish. Res. Board Can.
29:583-587.
Craig, G.R., and W.F. Baksi. 1977. The effects of depressed pH on flagfish reproduction,
growth and survival. Water Res. 11:621-626.
Cumbie, P.M., and S.L. Van Horn. 1978. Selenium accumulation associated with fish mortality
and reproductive failure. Proc. Ann. Conf. SE Assoc. Fish and Wildlife Agencies
32:612-624.
Eaton, J.G. 1973. Chronic toxicity of a copper, cadmium and zinc mixture to the fathead
minnow (Pimephales promelas RaGnesque). Water Res. 7:1723-1736.
-------
Freeman, H.C., and G.B. Sangalang. 1985. The effects of an acidic river, caused by acidic rain,
on weight gain, steroidogenesis, and reproduction in the Atlantic salmon (Salmo salar).
Pages 339-349. In: Aquatic Toxicology and Hazard Assessment: Eighth Symposium,
ASTM STP 891. R.C. Bahner and DJ. Hansen (Eds.). American Society for Testing and
Materials, Philadelphia.
Harvey, B J., and W.S. Hoar. 1979. The theory and practice of induced breeding in fish.
IDRD-TS21e. Publications of International Development Research Centre of Canada
Ottawa, Ont. pp. 1-48.
Kapur, K., K. Kamaldeep, and H.S. Toor. 1978. The effect of fenitrothion on reproduction of a
teleost fish, Cyprinus carpio communis Linn: A biochemical study. Bull. Environ Contain
Toxicol. 20:438-442.
Katti, S.R., and AG. Sathyanesan. 1986. Changes in the hypothalamo-neurobypophysial
complex of lead treated fish, Clarias batrachus (L). Z. Mikrosk-Ant Forsch Leipzig
100:347-352. V B
Kirubagaran, R., and K.P. Joy. 1988. Toxic effects of mercuric chloride, methylmercuric
chloride, and Emisan 6 (an organic mercurial fungicide) on ovarian recrudescence in the
catfish, Clarias batrachus (L.). Bull. Environ. Contam. Toxicol. 41:902-909.
Kumar, V., and D. Mukherjee. 1988. Phenol and sulfide induced changes in the ovary and liver
of sexually maturing common carp, Cyprinus carpio. Aquatic Toxicology 13:53-60.
Kumar, S., and S.C. Pant. 1984. Comparative effects of the sublethal poisoning of zinc, copper,
and lead on the gonads of the teleost Puntius conchonhis Ham. Toxicology Letters
23:189-194.
Kumar, S., and S.C. Pant. 1988. Comparative sublethal ovarian pathology of some pesticides in
the teleost, Puntius conchonhis Hamilton. Bull. Environ. Contam. Toxicol. 41:227-232.
Lee, R.M., and S.D. Gerking. 1980. Survival and reproductive performance of the desert
pupfish, Cyprinodon n. nevadensis (Eigenmann and Eigenmann), in acid waters J Fish
Biol. 17:507-515.
McCormick, J.H., G.N. Stokes, and R.O. Hermanutz. 1989. Oocyte atresia and reproductive
success in fathead minnows (Pimephales promelas) exposed to acidified hardwater
environments. Arch. Environ. Contam. Toxicol. 18:207-214.
Mount, D.I. 1973. Chronic effect of low pH on fathead minnow survival, growth and
reproduction. Water Res. 7:87-993.
Nagahama, Y., H. Kagawa, and G. Young. 1980. Cellular sources of sex steroids in teleost
gonads. Can. J. Fish. Aquat. Sci. 39:56-64.
Nebeker, A.V., F.A Puglisi, and D.L. DeFoe. 1974. Effect of polychlorinated biphenyl
compounds on survival and reproduction of the fathead minnow and flagfish Trans
Amer. Fish. Soc. 103:562-568.
-------
Peter, R.K, J.P. Chang, C.S. Nahorniak, RJ. Omeljaniuk, M. Sokolowska. S.H. Shih, and R.
Billard. 1986. Interactions of catecholamines and GnRH in regulation of gonadotropin
secretion in teleost fish. (Invited paper at 1985 Laurentian Hormone Conference, Banff,
Alberta, 8-13 September 1985). Rec. Progr. Horm. Res. 42:513-548.
Pickering, Q., W. Brungs, and M. Gast. .1977. Effect of exposure time and copper
concentration on reproduction of the fathead minnow (Pimephales promelas). Water Res.
11:1079-1083.
Ram, R.N., and K.P. Joy. 1988. Mercurial induced changes in the hypothalamo-neurohypo-
physial complex in relation to reproduction in the teleostean fish, Channa punctatus
(Bloch). Bull. Environ. Contam. Toxicol., 41:329-336.
Ram, R.N., and A.G. Sathyanesan. 1986. Effect of mercurial fungicide on the gonadal
development of the teleostean fish, Channa punctatus (Bloch). Ecotoxicology and
Environmental Safety 11:352-360.
Ruby, S.M., J. Aczel, and G.R. Craig. 1977. The effects of depressed pH on oogenesis in
flagfish Jordanella floridae. Water Res. 11:757-762.
Sangalang, G.B., and MJ. O'Halloran. 1973. Adverse effects of cadmium on brook trout testis
and on in vitro testicular androgen synthesis. Biology of Reproduction 9:394-403.
Saunders, J.W. 1969. Mass mortalities and behaviour of brook trout and juvenile Atlantic
Salmon in a stream polluted by agriculture pesticides. J. Fish. Res. Board Can.
26:695-699.
Sehgal, R., and A.B. Saxena. 1986. Toxicity of zinc to a viviparous fish, Lebistes reticulatus
(Peters). Bull. Environ. Contam. Toxicol. 36:888-899.
Singh, H., and T.P. Singh. 1981. Effect of parathion and aldrin on survival, ovarian 32P uptake
and gonadotrophic potency in a freshwater catfish, Heteropneustes fossilis (Bloch).
Endokrinologic 77:173-178.
Singh, H., and T.P. Singh. 1982a. Effect of some pesticides on hypothalamo-hypophyseal-
ovarian axis in the freshwater catfish, Heteropneustes fossilis (Bloch). Environmental
Pollution (Series A) 27:383-388.
Singh, H., and T.P. Singh. 1982b. Effect of pesticides on fish reproduction. Ichthyologia
15:71-81.
Sorensen, E.M.B. 1988. Selenium accumulation, reproductive status, and histopathological
changes in environmentally exposed redear sunfish. Arch. Toxicol. 61:324-329.
Spehar, R.L 1976. Cadmium and zinc toxicity to flagfish, Jordanella floridae. J. Fish. Res.
Board Can. 33:1939-1945.
Spehar, R.L., E.N. Leonard, and D.L. DeFoe. 1978. Chronic effects of cadmium and zinc
mixtures on flagfish (Jordanella floridae). Trans. Amer. Fish. Soc. 107: 354-360.
10
-------
Speranza, A.W., RJ. Seeley, V.A. Seeley, and A. Perlmutter. 1977. The effect of sublethal
concentration of zinc on reproduction in the zebra fish, Brachydanio rerio Hamilton-
Buchanan. Environ. Pollut. 12:217-222.
Struhsaker, J.W. 1977. Effects of benzene (a toxic component of crude oil) on spawning Pacific
herring, Clupea harengas pallasl Fish. Bull. 75:43-49.
Tarn, W.H., and P.D. Payson. 1986. Effects of chronic exposure to sublethal pH on growth, egg
production and ovulation in brook trout^JSalvelmus fontinalis. Can. J. Fish. Aquat Sci
43:275-280.
Truscott, B., J.M. Walsh, M.P. Burton, J.F. Payne, and D.R. Idler. 1983. Effect of acute
exposure to crude petroleum on some reproductive hormone in salmon and flounder.
Comp. Biochem. Physiol. 75G121-130.
Urist, M.R., and AO. Schjeide. 1961. The partition of calcium and protein in the blood of
oviparius vertebrates during estrus. J. Gen. Physiol. 44:743-756.
Von Westernhagen, H. 1988. Sublethal effects of pollutants on fish eggs and larvae. Pages
253-331. In: Fish Physiology. W.S. Hoar and DJ. Randall (Eds.), Volume II, Part A.
Academic Press, New York.
Weiner, G.S., C.B. Schreck, and W.L. Hiram. 1986. Effects of low pH on reproduction of
rainbow trout. Trans. Amer. Fish. Soc. 115:75-82.
11
-------
-------
USE OF FISH TO MONITOR TOXICITY AND
ACCUMULATIVE CAPACITY OF INSTREAM
CONTAMINANTS
by
William H. Benson1 and Jeffrey A Black2
INTRODUCTION
Chemical monitoring has been a critical component in assessing the biological impact of
chemicals on aquatic ecosystems. However, because of the importance of accurately assessing
the biological impact of chemicals, more recently, attention has been given to the use of fish to
monitor toxicity and accumulative capacity of instream contaminants. In addition, analytical
chemical assessment may, at times, be of limited value given the number of environmental
parameters that must be considered in such an assessment. There is, therefore, a need to
develop a comprehensive risk assessment approach. Such an approach would, when possible,
include concurrent chemical and biological monitoring. An obvious advantage to conducting
both chemical and biological monitoring is that if toxicity is demonstrated the cause may be
discernible. However, cause-effect relationships are not always easily established. To select
between chemical and biological monitoring, certain aspects of each must be considered.
The conventional laboratory toxicity test, in which conditions may be controlled within
desired limits, is valuable for studies of the properties of specific toxicants. However, for many
practical purposes it is difficult to simulate the full range of combinations and variations of
toxicant concentrations which are likely to be encountered in the field. In such cases it may be
more informative to observe the responses of organisms exposed to actual environmental
conditions. This is possible through the use of on-site studies with mobile laboratories or with
the use of specially constructed instream exposure chambers to permit in situ evaluations. This
paper reviews laboratory on-site and in situ methods for evaluating toxicity and accumulative
capacity of instream contaminants. For purposes of the following discussion, on site refers to
procedures used for field testing of effluent or ambient samples in a mobile laboratory. In situ is
defined as actual instream exposure.
RATIONALE FOR BIOLOGICAL MONITORING
Table 1 presents several underlying considerations for employing a biological monitoring
approach. First, water quality criteria, valuable in a risk assessment process, have been
established for relatively few chemicals. Additionally, a chemical monitoring approach alone
likely will not detect all the chemicals in a contaminated ecosystem. Although chemical
Department of Pharmacology and Research Institute of Pharmaceutical Sciences, School of
Pharmacy, The University of Mississippi, University, Mississippi, USA
2EA Engineering Science and Technology, Inc., Hunt Valley Loveton Center, Sparks, Maryland,
USA
13
-------
Table 1. Rationale for using chemical and/or biological monitoring.
Availability of water quality criteria
Number of chemicals present
Toxicant interactions
Assessments of bioavailability
Temporal variability/persistence
Bioaccumulative chemicals
Implementation of toxicity reduction procedures
Expense relative to program objectives
monitoring is sufficient for quantifying a limited number of selected contaminants, it usually is
not sufficient when many chemicals are present, especially considering possible synergistic or
additive toxicant interactions. In these situations, an organism may be used as a biosensor to
detect those chemicals and chemical interactions occurring in an effluent or receiving water.
Bioavailability, temporal variability, and persistence of toxicity also can be evaluated by using a
biological monitoring approach. For determining the presence of bioaccumulative chemicals,
both chemical and biological monitoring are appropriate. For rapid assessments, chemical
monitoring is arguably more efficient for detecting carcinogens. Examination of indigenous
aquatic organisms for tumor formation and other anatomical and cytological endpoints also may
serve as effective biomarkers of carcinogen exposure and effect. Both chemical and biological
assessments are reliable for design of toxicity reduction systems (e.g., chlorine reduction). In
selecting methods, expense is important; consideration must be given to specific program
objectives and availability of funds.
The purposes of biological testing vary; different procedures are used to accomplish
different objectives. Biological testing may be used for regulatory purposes. An industry's
discharge permit may be based on characteristics of the mixing zone of the stream. In this case,
biological monitoring could be used to determine compliance in the mixing zone. In another
application, biological monitoring may be used to develop a scientific data base in which findings
of toxicity tests can be compared with ecological endpoints (i.e., field validation studies).
Biological monitoring also may be used to determine instream sources of impact. For industries
with only one or two discharges, the sources may be obvious. For nonpoint sources such as
agricultural runoff, biological monitoring can be useful in identifying important toxic inputs.
Additionally, the effects of water quality characteristics (e.g., salinity, hardness, pH) on toxicity
may be evaluated with biological endpoints.
METHODS FOR ON-SITE AND IN SITU EVALUATIONS
SELECTION OF FIELD STATIONS
The selection of field stations includes a number of important considerations. First,
choosing a station using map sitings alone is usually unsatisfactory, as many limitations of access
may not be discernible unless on-site reconnaissance is performed. The proximity of stations to
impact sources also is very important in attempting to determine if there is any spatial variability
in effect The establishment of a control or reference site can be difficult. For instance, an
14
-------
upstream site may be contaminated to the extent that it is unsuitable. Thus, one may be
restricted to using laboratory (e.g., reconstituted) water. The sites should be evaluated as to
their importance as spawning areas or recreational areas, and for their potential for producing
human health effects. For field validation studies described above, which involve evaluation of
benthic invertebrates, fish populations, periphyton, etc., sites for toxicity testing should be
selected near the ecological sampling sites.
TOXICITY TESTING
Toxicity tests used in a monitoring program should incorporate appropriate biological
endpoints. For example, in many field applications acute tests are not adequately sensitive, and
more sensitive chronic endpoints such as effects on growth and/or reproduction must be
examined. Whether on-site or in situ evaluations are conducted, the sampling procedures,
selection of test organisms, cost and personnel needs also are important considerations.
Variations in toxicity between laboratory and on-site evaluations of effluents often have
been observed. Generally, laboratory tests on stored effluents are less sensitive, in some
instances by nearly two orders of magnitude (Birge and Black 1990). In such cases, on-site
testing with a mobile facility has advantages over traditional laboratory tests.
A comparative ecological and toxicological investigation was conducted by Birge et al.
(1989) on a secondary treatment plant point-source discharge and the receiving system. The
principle objectives were to asses downstream persistence of aquatic contaminants, to quantify
their effects on structure and function of aquatic communities and to evaluate the fathead
minnow embryo-larval test for measuring instream toxicity and estimating chronic effects on
biota. A good predictive correlation was found between embryo-larval survival and independent
ecological parameters, especially species richness of macroinvertebrates.
There are a number of standardized toxicity test procedures available for on-site
biological monitoring (Table 2). Acute tests have been well documented in "Methods for
Measuring the Acute Toxicity of Effluents to Freshwater and Marine Organisms" (Weber 1991).
Established methods for estimating the chronic toxicity of receiving waters to freshwater
organisms (Weber et al. 1989) as well as marine and estuarine organisms (Weber et al. 1988)
also are available.
Unlike on-site testing, there are no standardized test procedures available for in situ
evaluations. However, there are a variety of test procedures which have been used successfully
for some time (Table 3). For example, as early as 1958, Allan et al. used floating cages in
evaluations of the survival of fish exposed to a sewage effluent. More recently, Jones and Sloan
(1989) used an in situ exposure vessel to evaluate accumulative capacity of caged organisms in
the Hudson River in New York. Vessel construction permitted exposure of a large number of
juvenile fish which could be maintained for time-series sampling.
Concurrent mobile laboratory on-site and in situ evaluations have been conducted by Hall
et al. (1988) in the Chop tank River and Upper Chesapeake Bay using prolarval and yearling
striped bass. Prolarval survival data from both the on-site and in situ evaluations were similar
for each spawning area examined. A thorough description of the design, as well as a discussion
of the capabilities and significance in biological monitoring of the in situ apparatus used by these
investigators has been published (Ziegenfuss et al. 1990).
15
-------
Table 2. On-site toxicity test procedures.
Acute - freshwater and marine*
• Daphnids (Daphnia magna, D. pulex)
• Mysids (Mysidopsis bahia)
• Fathead minnow (Pimephales promelas)
• Silverside (Menidia spp.)
Chronic - freshwater1"
• Fathead minnow larval survival and growth
• Fathead minnow embryo-larval survival and teratogenicity
• Ceriodaphnia survival and reproduction
• Algal (Selenastrum) growth
Chronic - marine0
• Sheepshead minnow larval survival and growth
• Sheepshead minnow embryo-larval survival and teratogenicity
• Silverside larval survival and growth
• Mysid survival, growth and fecundity
• Algal (Champia) reproduction
•Weber (1991)
bWeber et al. (1989)
"Weber et al. (1988)
Table 3. Endpoints for in situ evaluations.
Survival - larval, juvenile, adult
Growth
Production
Biochemical/Physiological
Accumulative capacity
DISCUSSION
As new tests are developed for on-site and in situ evaluations, reliable and reproducible
test endpoints must be established. Several novel in situ methods for evaluating the deleterious
effects of instream contaminants are currently under development. For example, chemically-
induced alterations in the biochemistry and physiology of aquatic organisms may be used as
potential diagnostic tools in biological monitoring. Such an example is the use of detoxification
processes, such as mixed-function oxygenase enzymes in the case of selected organic
contaminants (Payne 1984, Andersson et al. 1988, Lindstrom-Seppa and Oikari 1990, Mater-
Mihaich and DiGiulio 1990). Additional detoxification processes include metallothionein
16
-------
induction and subcellular distribution of heavy metals (Olsson and Haux 1986, Benson et al.
1990). Fish bile metabolites have been used successfully to assess contamination by
chorophenolics and resin acids (Oikari and Holmbom 1986), as well as chlorine-bleached kraft
pulpmill effluents (Oikari and Kunnamo-Ojala 1987). Freeman and Sangalang (1985) examined
the effects of an acidic river on weight gain, steroidogenesis, and reproduction in the Atlantic
salmon. Hematological and blood chemistry parameters also are attractive as indicators of
environmental health because they offer a rapid and sensitive means of monitoring the impact of
instream contaminants on aquatic organisms (Benson et al. 1988, Watson et al. 1989).
SUMMARY
When implementing biological monitoring in the field, investigators must be aware of
problems inherent in such testing programs. Several examples follow. First, it can be difficult to
demonstrate cause-effect relationships for ambient samples when toxicity is minimal or is
observed only occasionally. Secondly, while statistical procedures are established for evaluating
results of standardized toxicity tests that involve a graded concentration series and laboratory-
quality controls, these analyses may not be appropriate to determine the significance of field-
related responses. Finally, it is important to understand that the creation of manipulated
exposure'conditions (e.g., fish cages), albeit in the field, potentially produce non-toxicant related
stresses that may obscure observations of sublethal effects.
Despite its limitations, on-site or in situ biological monitoring provides a unique
opportunity for evaluating the effects on environmental conditions on indigenous organisms, or
their surrogates. It is expected further research in this area will result in more refined
methodologies that, in time, will assist in the development of improved ecological risk assessment
strategies.
ACKNOWLEDGMENTS
The authors thank Lenwood W. Hall and Richard T. DiGiulio for their assistance in
preparation of this manuscript.
REFERENCES
Allan, I.R.H., S.W.M. Herbert, and J.S. Alabaster. 1958. A field and laboratory investigation of
fish in a sewage effluent. Fishery Invest. London (I) 6:1-76.
Andersson, T., L. Forlin, J. Hardig, and A Larsson. 1988. Physiological disturbances in fish
living in coastal water polluted with bleached kraft pulp mill effluent. Can. J. Fish. Aquat.
Sci. 45:1525-1536.
Benson, W.H., C.F. Watson, K.N. Baer, and R.A. Stackhouse. 1988. Response of hematological
and biochemical parameters to heavy metal exposure: implications in environmental
monitoring. Mar. Environ. Res. 24:219-222.
17
-------
Benson, W.H., KN. Baer, and C.F. Watson. 1990. Metallothionein as a biomarker of
environmental metal contamination: Species-dependent effects. In: Biological Markers of
Environmental Contamination. J.F. McCarthy and L.R. Shugart (eds.). Lewis Publishers,
Boca Raton, EL pp. 255-265.
Birge, W.J., J.A. Black, T.M. Short, and A.G. Westerman. 1989. A'comparative ecological and
lexicological investigation of a secondary wastewater treatment plant effluent and its
receiving stream. Environ. Toxicol. Chem. 8:437-450.
Birge, W.J., and J.A. Black. 1990. In situ toxicological monitoring: Use in quantifying ecological
effects of toxic wastes. In: In Situ Evaluations of Biological Hazards of Environmental
Pollutants. S.S. Sandhu et aL (eds.). Plenum Press, NY. pp. 215-231.
Freeman, H.G, and G.B. Sangalang. 1985. The effects of an acidic river, caused by acidic rain*
on weight gain, steroid oogenesis, and reproduction in the Atlantic Salmon (Salmp salar).
In: Aquatic Toxicology and Hazard Assessment: Eighth Symposium, ASTM STP 891.
R.C Banner and DJ. Hansen (eds.). American Society for Testing and Materials,
Philadelphia, PA. pp. 339-349.
Hall, L.W., Jr., S.J. Bushong, M.C. Ziegenfuss, W.S. Hall, and R.L. Herman. 1988. Concurrent
mobile on-site and in situ stripped bass contaminant and water quality studies in the
Choptank River and upper Chesapeake Bay. Environ. Toxicol. Chem. 7:815-830.
Jones, P.A., and R J. Sloan. 1989. An in situ river exposure vessel for bioaccumulation studies
with juvenile fish. Environ. Toxicol. Chem. 8:151-155.
Lmdstrom-Seppa, P., and A. Oikari. 1990. Biotransformation and other toxicological and
physiological responses in rainbow trout (Salmo gairdneri Richardson) caged in a lake
receiving effluent of pulp and paper industry. Aq. Toxicol. 16:187-204.
Mater-Mihaich, E., and R.T. DiGiulio. 1990. Oxidant, mixed-function oxidase, and peroxisomal
responses in channel catfish exposed to a bleached kraft mill effluent. MS submitted.
Oikari, A, and T. Kunnamo-Ojala. 1987. Tracing of xenobiotic contamination in water with the
aid of fish bile metabolites: A field study with caged rainbow trout (Salmo gairdneri). Aq.
Toxicol. 9:327-341.
Oikari, A., and B. Holmbom. 1986. Assessment of water contamination by chlorophenolic and
resin acids with the aid of fish bile metabolites. In: Aquatic Toxicology and
Environmental fate: Ninth Volume, ASTM STP 921. T.M. Poston and R. Purdy (eds.).
American Society for Testing and Materials, Philadelphia, PA. pp. 252-267.
Olsson, P-E., and C. Haux. 1986. Increased hepatic metallothionein content correlates to
cadmium accumulation in environmentally exposed perch (Perca fluviatilis). Aq. Toxicol.
9:231-242.
Payne, J.F. 1984. Mixed-function oxygenases in biological monitoring programs: Review of
potential usage in different phyla of aquatic animals. In: Ecotoxicology Testing for the
Marine Environment. G. Persoone, E. Jaspers, and C. Claus (eds.). State Univ. Ghent
Inst. Mar. Sci. Res., Bredene, Belgium, pp. 625-655.
18
-------
Watson, C.F., K.N. Baer, and W.H. Benson. 1989. Dorsal gill incision: A simple method for
obtaining blood samples in small fish. Environ. Toxicol. Chem. 8:457-461.
Weber, C.I. (ed.) 1991. Methods for measuring the acute toxicity of effluents and receiving
waters to freshwater and marine organisms, 4th edition. EPA-600/4-90-027. U.S.
Environmental Protection Agency. Cincinnati, OH.
Weber, C.I., W.B. Horning, H, D.J. Klemm, T.W. Neiheisel, P.A Lewis, E.L. Robinson, J.
Menkedick, and F. Kessler. 1988. Short-term methods for estimating the chronic toxicity
of effluent and receiving waters to marine and estuarine organisms. EPA-600/4-87/028.
U.S. Environmental Protection Agency, Cincinnati, OH.
Weber, C.I., et al. 1989. Short-term methods for estimating the chronic toxicity of effluent and
receiving waters to freshwater organisms, 2nd edition. EPA-600/4-89/001. U.S.
Environmental Protection Agency. Cincinnati, OH.
Ziegenfuss, M.C., L.W. Hall, Jr., S.J. Bushong, J.A Sullivan, and M.A Unger. 1990. A remote
in situ apparatus for ambient toxicity testing of larval and yearling fish in river or estuarine
streams. Environ. Toxicol. Chem. 9:1311-1315.
19
-------
-------
USE OF BIOMONITORING TO CONTROL
TOXICS IN THE UNITED STATES
Nelson A. Thomas1
ABSTRACT
A biomonitoring program has been developed in support of the National Policy for the
Development of Water Quality-Based Permit Limitations for Toxic Pollutants. The program
focuses on the use of laboratory toxicity tests with aquatic plants and animals to predict
ecosystem impacts caused by toxic pollutants. Both acute and chronic toxicity tests were
developed to test effluents and ambient waters. Laboratory and biological field studies were
conducted at nine sites. Single species laboratory toxicity tests were found to be good predictors
of impacts on the ecosystem when two or more species were used. Biomonitoring can be
undertaken either on effluents and/or on the receiving waters. When toxicity upstream from a
discharge is suspected to originate from inputs such as seeps, leachates, or storm sewers, it is
beneficial to conduct both effluent and ambient biomonitoring. This will allow one to determine
more accurately the impact of a discharge on a receiving water body.
INTRODUCTION
In the United States, as control of conventional pollutants is achieved, increased emphasis
is being placed on reduction of toxic chemicals. The U.S. Environmental Protection Agency
(EPA) has developed a Water Quality Based Approach (See Glossary) to achieve desired water
quality where Treatment Control Based Discharge Limits have proven to be insufficient. To
control discharges by water quality, it is necessary to demonstrate quantitatively the connection
of biological effects to effluent limits. Single chemical criteria or toxicity limits on effluents can
be developed to do this. The high cost of data generation makes single chemical criteria
development a slow process. The cost to develop a single criteria document is estimated to be
$100,000. Application of water quality criteria to the Waste Load Allocation Process is
complicated by the inability to predict chemical action in natural waters and to predict the
effects of .effluent mixtures. Biomonitoring is often more cost-effective and incorporates the
effects of receiving water body chemistry on toxicity.
FIELD STUDIES
Greatest emphasis has been placed on validating the use of a battery of toxicity tests to
predict the impact of toxics on an effluent and on the biological community of a receiving water
body. As toxicity limits are incorporated into permits, it becomes necessary to develop
methodologies for biomonitoring permit compliance. Biological field evaluations of discharges
on receiving waters have been made for the past 50 years. Generally, these studies have
. Environmental Protection Agency, Environmental Research Laboratory, Duluth,
Minnesota, USA.
21
-------
focused on losses of important components of the^ biological community. Population studies on
fish and benthic invertebrates have included the principal taxonomic groups. Other groups for
which measures have been developed include algae, protozoans, and bacteria. The results of
most pollution field studies have been used to assess the ecological health of water bodies. It is
difficult to establish quantitative correlations between measures of impact on a community and
the cause of that impact. In order to establish the necessary treatment required to achieve a
desired reduction of impact on a biological community, a quantifiable cause and effect
relationship must be developed. The measures of community response to toxic inputs are often
not sensitive, because of population variability and other reasons. The impact must often be
severe before a change can be measured. Biological field assessments are difficult to
incorporate into the National Pollutant Discharge Elimination System (NPDES) Permit process.
Permits require frequent sampling to incorporate effluent variability.
Field studies are useful in evaluating the overall effect of all disturbances on ecosystems.
Usually, effects of non-point sources, multiple discharges and habitat alterations make it
impossible to evaluate the impact of individual effluents.
TISSUE RESIDUE
Measurements of tissue residues of both free and captive organisms have been used to
monitor the effects of many compounds that bioaccumulate, especially pesticides.
Programs such as "Mussel Watch" have provided quick and accurate assessments of the
occurrence of residue-forming compounds. Often the residues are associated with chemicals
originating from non-point sources. These require action levels which are generally associated
with human food consumption. To date, in the United States, very few dischargers are required
to conduct tissue residue monitoring tests. Most monitoring programs focus on general base-line
activities to establish and maintain fish consumption warnings. Research is underway to develop
a procedure to identify and quantify the presence of chemicals in effluents that bioaccumulate to
form residues, which can then be related to a fish consumption action level. This method will
provide for effluent-specific evaluations, in contrast to residue determinations, of free-roaming
fish from receiving water bodies to which effluent limits are often not assignable.
TOXICITY
The most influential impetus to increase biomonitoring in the United states was the
issuance by EPA, in March 1984 (Anon. 1984), of the National Policy for the Development of
Water Quality-Based Permit Limitations for Toxic Pollutants. This policy recommended the use
of toxicity testing of effluents to achieve state narrative standards that prohibited the discharge
of toxic materials in toxic amounts. For this policy to be effective, proof that toxicity tests are
good predictors of impacts on aquatic communities had to be developed. Single species toxicity
tests have been found to be a good predictor of ecosystem impact resulting from toxic input.
Seven field studies were conducted at sites where effluents and ambient toxicity tests were being
conducted (Mount et al. 1984,1985a, 1985b, 1985c, 1985d, 1986; Norberg-King et al. 1986). A
comparison of tests at 83 sampling locations was made between receiving water toxicity and
aquatic community impact as measured by species lost (Figure 1). The correlation of toxicity to
22
-------
species loss was 0.54. The comparison of the most sensitive toxicity test to the most severe
ecosystem response, as measured by species loss, correlated better than, the various measures of
ecosystem impact.
100
80
111 60
CO
CO
UJ
40
20
*•*" * •
m m» • *
3^5-^
20 40 60 . 80
LABORATORY TOXICITY (%)
100
Figure 1. Laboratory toxicity versus field response at 83 sites.
Currently, more than 6,000 discharge permits have toxicity limits to protect against
chronic toxicity. However, some states use acute toxicity limits to protect against chronic
toxicity. In a majority of the cases, biomonitoring with toxicity testing is required for the effluent
but not for the receiving water body. A few permits require biological surveys, but these are of
the traditional field population studies. The use of toxicity of ambient waters is the next logical
step in implementation of the Water Quality-Based Approach. The advantages of the ambient
toxicity tests include: (1) the establishment of a quantifiable relationship between the discharges
and the water quality of the receiving water body; (2) lower costs than general biological surveys
or chemical scan; and (3) knowledge that a toxicity test could elicit a response from a chemical
that might go undetected in a normal chemical scan.
Ambient toxicity testing can be used as a screen to determine whether additional effluent
testing is required (Figure 2). Of the 20 sites in Table 1 where ambient toxicity was tested,
toxicity was measured at eight of the sites. Six were found to be lethal and two had effects on
fathead minnow growth. All of the effluents, where toxicity was observed downstream, had an
acceptable effluent concentration of 17.3% or less. Ambient toxicity testing was conducted at 20
stations, five of which were toxic. This suggests that if an impact of an effluent is to be
evaluated, the toxicity of the upstream site should be investigated along with the effluent. It
may be necessary to include effluent dilutions with water from other sources known to be free of
toxicity. In the case of the second paper plant on Owl Creek, the effluent had no effect on
23
-------
30
LU
i
£20
O
10
50
40
30
20
10
KILOMETERS
Figure 2. Ambient toxicity as measured by Ceriodaphnia young
production, Ottawa River, Ohio.
Ceriodaphnia, and only the 100% concentration was lethal. The water upstream from this plant
was acutely toxic to Ceriodaphnia, most likely the result of the upstream paper plant discharge.
The water upstream from both of the paper plants was not toxic; therefore it was used as the
dilution water for both effluents. The effluent of the plant located upstream on Owl Creek was
lethal at the 100% concentration and could therefore have caused toxicity immediately upstream
from the second plant.
Most of the water bodies of the United States discharge into receiving waters where the
dilution ratio is less than 10 at low flow and greater than 100 at annual mean flow (Anon. 1985).
If the toxicity of the discharges in Table 1 is typical of discharges suspected of containing toxics,
one would expect to observe ambient toxicity during low flow in 75% of the cases and during
average flow in only 5% of the cases. One must compare the lethal concentrations to the
dilutions at both the low and annual mean flows.
In developing a biomonitoring program, factors such as species sensitivity, effluent
variability, and stream flow must be considered. Aquatic organisms, both in the laboratory and
in the field, exhibit different sensitivities to different toxicants. Data collected in our studies
indicate the greater sensitivity of Daphnia as opposed to fish is a common observation (Tables 2
and 3). In a study on industrial discharges the two daphnids were the most sensitive, followed
by fathead minnows, green sunfish and rainbow trout. The importance of species sensitivity has
been recognized in the development of Water Quality Criteria. Minimum data sets have been
established that require representatives from the major taxonomic groups (Stephan et al. 1980).
24
-------
Table 1. Effluent toxicity and ambient toxicity data from 14 rivers and creeks.
Effluent Type
Truck Mfg.
Leachate
Seeps
As Arco1
Oil
Paper
Paper
Chemical
Chemical
POTW3*! (4 samples)
Textile (5 samples)
POTW #2
POTW#3
POTW #4
POTW #5
POTW #6
POTW #7
POTW #8
POTW #9
POTW #10
Upstream
Toxicity
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Present
Present
Absent
Present
Present
Absent
Absent
Absent
Absent
Acceptable
Effluent
Concentration
(«)
17
1
0.4
100
1.7
17
55
55
5.5
< 1
6
5.5
17.3
17.3
54.8
100
54.8
5.5
17.3
1.7
Downstream
Toxicity
Absent
Absent
Present
NS2
Present
Present
Present
Absent
Absent
Absent
Absent
Present
Present
NS
Absent
Absent
Absent
Absent
Present
Present
River
or
Creek
Maumee
Maumee
Coeur
d'Alene
Coeur
d'Alene
Otter
Owl
Owl
Taylor
Taylor
Wantaga
Wantaga
Arkansas
Arkansas
Arkansas
Arkansas
Chat
Cedar
Mill
Wilson
So. Dry Sac
lAs Arco = Industrial/Petroleum Effluent
2NS = Not Sampled
3POTW = Publicly Owned Treatment Works
25
-------
Table 2. LC50 values for species sensitivity test conducted on effluent from Sherwood
Medical on 20 November 1985.
Species
Rainbow Trout
Daphnia magna
Fathead minnow (juvenile)
Green sunfish
Fathead minnow (larval)
Ceriodaphnia dubia
Diluent
Lake Superior
Lake Superior
Lake Superior
Lake Superior
Lake Superior
Lester River
LC50 (95% confidence interval),
% effluent (v/v)
17.3
3.9 (2.8-5.4)
7.5 (5.9-9.7)
13.4 (9.5-18.8)
8.7 (6.4-11.7)
5.5
Table 3. LC50 values for species sensitivity test conducted on effluent from Anchor Fastener
on 4 December 1985.
Species
Rainbow Trout
Daphnia magna
Fathead minnow (juvenile)
Green sunfish
Fathead minnow (larval)
Ceriodaphnia dubia
Diluent
Lake Superior
Lake Superior
Lake Superior
Lake Superior
Lake Superior
Lake Superior
LC50 (95% confidence interval),
% effluent (v/v)
10.9 (7.7-15.6)
6.9 (5.2-9.2)
27.5 (20.7-36.3)
43.5 (35.1-54.0)
23.1 (17.8-30.0)
17.3
In developing toxicity biomonitoring techniques or protocols for effluents, EPA recommends
testing with a fish, an invertebrate and a plant to reduce the uncertainty associated with a
sensitive species (Anon. 1985). It is important that the test species have a sensitivity similar to
the ecological groups that are to be protected. Effluent variability is a problem in any
monitoring program. Often 24-hour composite samples are used in place of grab samples. The
magnitude of the problem is shown in Figure 3. If one sample only, collected at 1600, 1800, or
2000 on 5/5, had been used, the toxicity might not have been detected. One might assume that
the toxicity is flow-dependent and that point source samples should be obtained at low flow.
26
-------
100
TIME
Figure 3. Time variable toxicity as measured by Ceriodaphnia survival
and mean young production, Trinity River.
In the case of Trinity River, toxicity was observed on the receding part of a high flow
event. In designing an ambient biomonitoring program it is important to understand the flow
dependence of the sources of the toxics. For most point sources; ambient toxicity is best
conducted during the time of the year when the receiving water body is at low flow. However, if
one suspects the toxicity to be the result of sewer overflow, seeps, leachates, or runoff, sampling
must be conducted under several flows.
Another consideration is to decide when to use acute and/or chronic tests. The level of
protection that is generally provided to an ecosystem requires the elimination of chronic toxicity.
Many agencies are using acute toxicity as an effluent limit to protect against chronic effects. In
early studies of acute to chronic ratios (ACR) of effluents, the ACR was generally found to be
about 10 (Anon. 1985).
APPLICATION
The final step in a biomonitoring program is to use the information obtained. By using
field studies which rely on diversity and abundance, one can judge the health of systems. In
cases where impacts are not observed, interpretation is quite easy. However, when impacts are
noted, connection to the cause is often difficult. If ambient and/or effluent toxiciiy has been the
focus of the biomonitoring program, both the amount of toxicity reduction and the identification
of the toxicant can be determined. To protect an aquatic ecosystem, the Instream Waste
Concentration (IWC) must be less than the No Observable Effect Level (NOEL) or Acceptable
Effluent Concentration (AEC). This relationship is most protective when design flow conditions
are specified. The allowable effluent toxicity can easily be calculated based on available
27
-------
dilution. This can be done in the absence of toxicity tests. The Allowable Effluent Toxicity can
then be compared to the Measured Acceptable Effluent Concentrations to determine whether
further treatment is required. If the effluent and receiving water body flows are known at the
time of sampling, one can compare the predicted impact based on effluent toxicity to the
observed ambient toxicity. The data presented in Table 1 indicate the frequent presence of
toxicity downstream from toxic effluent where no toxicity had been observed upstream.
TOXICITY IDENTIFICATION
When toxicity is found in an ambient water body or in an effluent in sufficient
concentration that it would still be toxic after dilution, the first question is, "What is causing this
toxicity and how do I remove it?".
Usually when an effluent is having an impact, identification of the toxicant is often
required before an action plan can be developed. Previous attempts have been made to identify
toxicity through the use of chemical screens. These have almost always failed. Most recently,
research has been undertaken to use toxicity in the fractionation of effluents. Chemical-physical
characteristics of the effluent are determined. These include filterability, complexity, pH
dependence, solubility, volatility, etc. Knowing these characteristics aids in developing a
treatment strategy and/or direction in further chemical identification. If toxicity is found in a
fraction containing many compounds; i.e., hydrophobic organic, further fractionation is
undertaken to ease the final chemical identification. Once an identification is made, spiking the
sample can verify that the suspected compound is causing the problem.
Comparisons of the Coefficient of Variation of single species toxicity tests and chemical
tests indicate that the precision of the toxicity test is usually less than 50%. For a variety of
organic and inorganic measurements, the Coefficient of Variation is in the same range. The
Coefficient of Variation of these same analytical measurements near the detection limit is much
greater than 50%. Since toxicity test endpoints are near the "detection limit" (no-effect
concentration or kills 50% of the organisms) the most appropriate comparison with analytical
measurements is at the detection limit. Viewed in this way, the precision of toxicity tests is
better than many analytical measurements.
SUMMARY
Biomonitoring has accelerated in the United States in the last five years as a result of
the recognition that laboratory aquatic tests can be used to predict aquatic community impact.
Effluent toxicity has been shown to be quantifiably associated with ambient toxicity, thus
providing a numerical cause and effect relationship. This quantification provides regulators with
the tools to develop discharge limits. Biomonitoring of ambient waters is becoming an
important component in evaluating the water quality status of the nation's waters.
ACKNOWLEDGMENT
Toxicity test data in Table 3, provided by Joseph Amato and Teresa Norberg-King, is
gratefully appreciated.
28
-------
GLOSSARY
Water Quality Based Approach - Use of whole effluent toxicity limits, chemical specific criteria
and biological criteria to control the discharge of toxics into the nation's waterways.
Treatment Control Based Discharge Limits - The control of toxics (discharge permit limit) is
based on waste treatment technologies.
National Pollutant Discharge Elimination System (NPDES) - As part of the Clean Water Act
numerical limits were to establish a permit for each discharge.
Instream Waste Concentration: The concentration of a chemical contained in an effluent after
dilution by the receiving water body.
No Observed Effect Concentration (NOEC) - The highest concentration of an effluent or a
toxicant at which no adverse effects are observed on the aquatic test organisms.
Acceptable Effluent Concentration - The concentration of chemical in effluent after dilution
does not exceed specific chemical toxicity limits. This can either be calculated or
measured.
REFERENCES
Anon. 1984. Development of Water Quality-Based Permit Limitations for Toxic Pollutants:
National Policy. U.S. Environmental Protection Agency, OW-FRL-2533-1.
Anon. 1985. Technical Support Document for Water Quality-Based Toxics Control. U.S.
Environmental Protection Agency, EPA-440/4-85-032.
Mount, D., N. Thomas, M. Barbour, T. Norberg, T. Roush, and W. Brandes. 1984. Effluent
and ambient toxicity testing and instream community response on the Ottawa River, Lima,
Ohio. U.S. Environmental Protection Agency, EPA-600/3-84-080.
Mount, D.I., and T. J. Norberg-King. (Eds.). 1985a. Validity of effluent and ambient toxicity
tests for predicting biological impact, Scippo Creek, Circleville, Ohio. U.S.Environmental
Protection Agency, EPA-600/3-85-044.
Mount, D.I., A.E. Steen, and T. J. Norberg-King. 1985b. Validity of effluent and ambient
toxicity for predicting biological impact, Five Mile Creek, Birmingham, Alabama. U.S.
Environmental Protection Agency, EPA-600/8-85-015.
Mount, D.I., A.E. Steen, and TJ. Norberg-King. 1985c. Validity of ambient toxicity tests for
predicting biological impact, Ohio River,,near Wheeling, West Virginia. U.S.
Environmental Protection Agency, EPA-600/3-85-071.
Mount, D.I., and TJ. Norberg-King. 1985d. Validity of effluent and ambient toxicity test for
predicting biological impact, Knawha River, Charleston, West Virginia. U.S.
Environmental Protection Agency, EPA-600/3-86-006.
29
-------
Mount, D.I., T. J. Norberg-King, and A.E. Steen. 1986. Validity of ambient toxicity tests for
predicting biological impact, Naugatuck River, Waterbury, Connecticut. U.S.
Environmental Protection Agency, EPA-600/8-86-005.
Norberg-King, T.J., and D.I. Mount. 1986. Validity of effluent and ambient toxicity tests for
predicting biological impact, Skeleton Creek, Enid, Oklahoma. U.S. Environmental
Protection Agency, EPA-600/8-86-002.
Stephan, C.E., D.I. Mount, DJ. Hansen, J.H. Gentile, G.A. Chapman, and W.A. Brungs. 1980.
Guidelines for deriving numerical national water quality criteria for the protection of
aquatic organisms and their uses. U.S. Environmental Protection Agency, NTIS PB 85-
227049.
30
-------
EFFECTS OF WATER POLLUTION ON THE GROWTH AND
DEVELOPMENT OF FISHES IN THE LOWER YANGTZE
RIVER OF CHINA
by
Yuan, Chuan-Fuh1
ABSTRACT
Water pollution is an important problem in China. For example, its effects on the
growth and development of fishes in the lower Yangtze River in recent years are as follows: (1)
many bodies and sexual glands of fishes are congenitally malformed; (2) heavy metals such as
Hg, Cr, Cd, As, Pb accumulate in the bodies of fishes; (3) the species of fishes are very few in
numbers, and (4) the population numbers of fishes have been .reduced.
INTRODUCTION AND GENERAL ACCOUNT OF OUR INVESTIGATION
In recent years, many of China's rivers, lakes, and other bodies of water have been
seriously polluted as a consequence of the rapid development of industry and farming Water
pollution gravely affects the growth and development of fishes living in various waters To
investigate the influence of water pollution on fishes in China, I have collected a large variety of
fish specimens and a lot of data which reveal effects of water pollution on the growth and
development of fishes. In this paper I take, for example, fishes in the lower Yangtze River and'
its waters to illustrate the grave effects of water pollution on the "growth and development of
fishes in China.
Time: 1973 to the present, about 18 years in total.
Places: Lower Yangtze River and its attached lakes, canals, and other bodies of water.
Number of malformed fishes: total number of malformed fishes collected as specimens is about
1000, the total number of visibly malformed fishes in 34 species being about 300.
EFFECTS OF WATER POLLUTION ON THE GROWTH OF FISHES __
Two points are worthy of note: many species of fishes cannot survive in severely polluted
waters; a large variety of fishes, although still living and growing, have poisonous substances in
their bodies, and will become malformed in moderately polluted waters. In the lower Yangtze
River, the malformed fishes make up 24% of the total of 140 species (Table 1).
Department of Zoology, Nanjing University, Nanjing, PRC.
31
-------
Table 1. Species names of fishes found to be malformed.
CLUPEIFQRMES
Clupeidae
Macnura reevesi (Richardson)
Engralidae
Coilia ectenes (Jordan et Seale)
Coilia ectenes taihuensis (Yuan et Lin)
Coilia brachygrathus (Kreyenberg et Pappenheim)
Coilia mystys (Linnaeus) ' . ,
Salangidae
Neasalanx tangkakkeii taihuensis (Chen)
ANGUILUFORMES
Anguillidae
Anguilla japotuca Jemminck et Schlegel
CYPRINIFORMES
Cyprinidae
Cyprinus carpio (Linnaeus)
Carassius auratus (Linnaeus)
Pseudorasbora parrva (Temminck et Schlegel)
Sarcocheittchthys nigripinnis (Gunther)
Abbottina rivularis (Basilewsky)
Hemibarbus maculatus (Sleeker)
Qpsariichthys unciroostris bidens (Gunther)
Mylopharyngodon piceus (Richardson)
Clenopharyngpdon idellus (Cuvier et Valenciennes)
Megalobrama ambfycephaia Yih
Megalobrama terminalis (Richardson)
Hemiculler leucisculus (Basilewsky)
Erythroculter iUshaeformis (Bleeker)
Acheilognathus gracilis Nichols
Acanihorhodeus tankinensis Vaillant
Acantharhodeus taenianaUs Gunther
Pseudoperilampus light Wu
Arisdchthys nobilis (Richardson)
Hypothalmichthys moUtrix Cuvier et Valenciennes
Cobitidae
Misgumus anguillicaudatus (Cantor)
SILURIFORMES
Bagridae
Pseudobagrns fulvidraco (Richardson)
SYMBRANCHIFORMES
Symbranchidae
Monopterus abus (Zuiew)
PERCIFORMES
Serranidae
Siniperca chuaisi (Basilewsky)
Gobiidae
Rhinogobius giurinus (Rutter)
Anabantidae
Macropodus chinensis (Bloch)
Gchlidae
Sarotherodon mossatnbica Peter
Sarotherodon nilotica (Linnaeus)
32
-------
PROPORTION OF MALFORMED FISHES IN RECENT YEARS
An analysis of malformed fish specimens has shown that the quantity and number of species
of malformed fishes have increased gradually over time. At present, malformed fishes are
distributed in various kinds of waters, which has attracted the attention of scientists. The
process of malformation increase is roughly divided into three periods.
(1) 1973 - 1979: During this period, there was only a small quantity of malformed fishes in
a small number of species. We could not find many malformed fishes. The proportion of
malformed fishes to the total number of specimens was very low, approximately 1-3% of the
collected specimens.
(2) 1980 -1984: The quantity and the number of species of malformed fishes increased,
such as Cyprinus carpio, Carassius auratus, Hypothalmichthys molitrix, Aristichthys nobilis,
Clenopharyngodon idellus, Monopterus abus, Misgumus anguUlicaudatus, and Cottia ectenes. The
distribution area of malformed fishes was enlarged. In terms of the number of separate
malformed fishes collected, we found that the proportion of malformed fishes to the total
number of specimens was about 5-10%.
(3) 1985 - 1990: The number and species of malformed fishes has increased obviously.
Almost all species of fishes listed in Table 1 were collected during this period, and the distribu-
tion area of them has gradually enlarged. In terms of the number of separate malformed fishes
collected as specimens, we found that the proportion of malformed fishes to the total number of
specimens is about 10-20%. In some waters, the proportion even rose to about 50%. In a few
particular regions it was 100%.
EXTERNAL FEATURES OF MALFORMED FISHES
Generally, from the collected malformed specimens, we found variations in their external
features, some variations being very distinctive. Now, let me talk briefly about the variations.
(1) Hemorrhage of skin: It is a very evident sign and can be observed easily. The
hemorrhage of skin may appear in dots or in stripes (Figure 1).
(2) Malformations of the body: Compared with normal fishes, the bodies of malformed
fishes are abnormal. Some bodily forms become longer than the normal body. Other bodily
forms are shorter from head to tail, yet higher from top to bottom. Some fishes have other
obvious variations in their bodily forms, for instance, the body becomes crooked, or some parts
of the body becomes projected, or atrophied, or the head has the right side unsymmetrical to
the left, or the tail grows abnormally and becomes very crooked (Figure 2).
(3) Scales arranged irregularly or deformed: In malformed scaled fishes, the scales are
irregularly arranged and have various kinds of deformation. Observing the scales through the
microscope, we can find that their annual rings and ring veins appear irregular (Figure 3).
(4) Abnormal fins: In malformed fishes, the position and shape of fins also have variations.
For example, some caudal fins grow abnormally; some dorsal fins and anal fins have abnormal
shapes, and their positions are greatly irregular. The structure of some pectoral fins and pelvic
fins often show the right side unsymmetrical to the left (Figure 4).
33
-------
Figure 1. Hemorrhage of the skin (Carassius auratus).
Figure 2. Malformations of the fish bodies
34
-------
Figure 3. Malformations of scales (Cyprinus carpio)
Figure 4. Malformations of the caudal fins (Carassius auratus).
35
-------
(5) Abnormal heads: Very often, heads of the malformed fishes have the right side
unsymmetrical to the left, sometimes, their mouths are crooked and their mandibular portion
are projecting, and their operculae may also become abnormal and appear incomplete. Also
the gill rakers and gill filaments appear abnormal.
(6) Abnormal eyes and nostrils: Some malformed fishes are marked by their deformed
eyes and nostrils. The right eye or nostril is often unsymmetrical to the left eye or nostril
(Figure 5). Their eyeballs are evidently projecting. The crystalline lens and retina have partly
disappeared or are full of air bubbles so they can no longer see (Figure 6).
Figure 5. Abnormal nostrils: (a) Ctenopharyngodon idellus; (b) Carassius auratus.
Figure 6. Abnormal eyes (exophthalmus) (Sarotherodon nilotica).
36
-------
INTERNAL STRUCTURE OF MALFORMED FISHES
I have selected from many specimens some models of malformed fishes, and have dis-
sected them and observed their internal structures in order to study important effects of water
pollution on fishes. Now, some serious malformations of internal structures have been found.
(1) Abnormal brains: In some fishes, the brain's right side is unsymmetrical to the left
For example: the right mesencephalon is in front of the left mesencephalon; the cerebellum,
the medula oblongata facial lobe, and the vagal lobe have become deformed, are misplaced, or
are unsymmetrical; some of the medula oblongata facila lobes have become smaller; some vagal
lobes have become flat and spread laterally; the right diencephalon are unsymmetrical to left;
the hypophysis have become deformed (Figure 7).
(a)
Figure 7. Malformations of the brains: (a) Carassius auratus; (b) Hemiculter
lenciculus.
(2) Abnormal skeletons: Some fishes have obvious zigzag vertebrates, and other fishes
have abnormal ribs that look like claws (Figure 8).
(3) Abnormal otolith: The otolith is an important part inside the ear. It has a fixed shape
and does not change its shape easily. Deformed otolithes were also found when I dissected
abnormal fishes. I found that some right and left otolithes were not of the same size. Some
otolithes were extended into abnormal shapes; others were atrophying into vestiges (Figure 9).
(4) Abnormal air-bladders: Some fishes had a big front bladder and a small back bladder.
Some air-bladders were atrophying or becoming a tube. Some air-bladders were marked by
hyperplasia. The tubes of some air-bladders were extended and curved or were misplaced
(Figure 10).
EFFECTS OF POLLUTION ON GLANDS OF FISHES
If the sexual glands are abnormal or stop developing, the growth of the coming generation
will be surely affected. Water pollution affects not only individual fish but also the survival of all
37
-------
Figure 8. Malformations of vertebrae and ribs (Cyprinus carpio).
(Normal)
Figure 9. Malformations of the otoliths (Asteriscus) (Cyprinus carpio).
38
-------
(b)
(c)
Figure 10. Malformations of air bladders: (a) Carassius auratus; (b) C. auratus;
(c) Hypothalmkhthys molitrix.
fishes. This is a very important problem. The results of specimen dissection lead us to believe
that deformed shapes and abnormal sexual glands have been found in fishes not only from lakes,
rivers, and pools, but also from the main stream of the Yangtze River. For example, in the case'
of Coilia ectenes living in the Yangtze River and offshore, abnormal, atrophied, and unsymmetri-
cal sexual glands can be easily found during their breeding season (Figure 11).
Figure 11. Malformations of the sex glands (Coilia ectenes).
39
-------
EFFECTS OF POLLUTION ON EGGS AND LARVAE OF FISHES
I would like to point out that eggs in abnormal ovaries are mostly marked by dysplasia. For
example, their shape and size are changed and are out of order (Figure 12). The eggs grow
abnormally and also develop unusual forms. Some sex glands seem to be normal, but eggs
produced by such sex glands cannot be hatched normally, or the larvae have a high rate of
death, or their larvae grow abnormally and become abnormal fish. Phenomena in this connec-
tion have been found in my collected specimens, for example, a Sarotherodon nilotica I collected
was shaped like Siamese twins with a small body grown incompletely and attached to one side of
its abdomen.
(Normal)
Figure 12. Malformations of the eggs (Cottia ectenes).
CONCLUSIONS AND ANALYSIS OF CAUSES
From the various abnormalities of fishes, we can conclude that a lot of waters in China have
been seriously polluted by now and some serious results have occurred. If we determine the
level of poison in malformed fishes, the causes of pollution can be obviously discovered.
According to identified data, some fishes look normal, but the amount and varieties of
poison in their bodies have increased. The varieties of poison include: Hg, Pb, As, Cd, Ra, Cl,
Zn, Cr, Cu, Ni, ammonia, phenol, fluoride, and cyanide. Some of these are evidently above
normal level, for example, Coilia ectenes in the Yangtze River have a poison level of 0.43 mg/kg
Hg in their bodies, which is eight times as high as the standard level stipulated by the World
Health Organization. Thus, although some fishes look normal, their flesh tastes bad, their
nutritive value has decreased, and they may have bad effects on people's health if eaten.
There are many causes which lead to fish abnormalities, but the most important cause is
that the water has been polluted. According to the data which I have investigated for many
years, we find the poison in water comes from many sources. The main sources of poison
40
-------
include wastewater from industries, wastewater from daily living, chemical fertilizers and
pesticides, and wastewater from medicine-production. In other words, no matter what waters
are polluted we can say it is man who causes pollution. Many factories and small towns were
planned from a long-term point of view. Some people think of only the present interest and
neglect the future interests. They allow untreated wastewater to flow into rivers and lakes.
Lacking knowledge of science, some people throw poisonous materials into water and let it
spread everywhere, and they even believe it is the best method to treat poisonous matter.
BIBLIOGRAPHY
Anonymous. 1983. Pollution investigation of the Yellow River. Joint Report of the Long River
Fishery Institute in Shar City and Yellow River Water Conservation Research Institute.
Ariens, E.J. 1976. Introduction to general toxicology. Academic Press, New York.
Bartik, M., and Piskac, A. 1981. Veterinary toxicology. Elsevier, Amsterdam.
Bond, C. E. 1979. Biology of fishes. W.B. Saunders Company. Philadelphia, London, Toronto.
Brent, R.L. 1977. Radiation and other physical agents. Handbook of Teratology. l(5):153-223.
Cahen, P.L. 1964. Evaluation of the teratogenicity of drugs. Clin. Pharmacology and Thera.,
5(4):480-514.
Chang, Zuei Tao. 1980. Study of fish abnormality by fluorides. Fresh Water Fishery 2:11-14.
Chang, Zuei Tao. 1982. Investigation of the effect of fluoride toxicity on fish abnormality.
Environmental Science (Part III) 4:1-15.
Chang, Zuei Tao. 1983. Preliminary investigation of the effect of water pollution on fishery.
Animal Report 4:23-26.
Chin, Ann Ling, and Chuan Mee Yuan. 1988. Effect of water pollution on fishery resources in
Great Canal (Chang-Jou City). Nanjing University Report (Natural Science edition)
24(1):97-107.
Chinese Academy of Sciences, Aquatic Biology Research Institute. 1980. Environmental
Pollution and Ecology Report. Jiang-Su Science and Technology Publishers.
Chinese Academy of Sciences, Geology Research Institute. 1976. Water pollution and water
conservation. Science Publishers.
Clarke, E.G.C., and M.L. Clarke. 1978. Veterinary toxicology. 2nd ed., Bailliere Tindall,
London.
Gutherie, F.E., and Perry, JJ. 1980. Introduction to environmental toxicology, Elsevier, North
Holland, Inc., New York.
Huang, Jur. 1986. The growth of abnormality. Biology Report 9:1-3.
41
-------
Ju, Pei Lei. 1989. Animal toxicology. Shanghai Science and Technology Publishers.
Kurent, I.E. 1977. Infectious diseases. Handbook of Teratology l(6):225-259.
Lin, Chang Shan, and Yee Ming Wu. 1986. Environmental biology. Chinese Environmental
Science Publishers.
Murph, M.L. 1965. Factors influencing teratogenic response to drugs. Teratology: Principles
and Techniques 7:145-184.
Radeleff, R.D. 1970. Veterinary toxicology. 2nd ed., Lea S. Febiger, Philadelphia.
Smithells, R.W. 1976. Environmental teratogens of man. British Medical Bulletin 32(l):29-33.
Wang, Fan Yuan, and Ying Chung. 1978. Shir-Chuan University Academic Report (Natural
Science edition) 4:81-90.
Wang, Fan Yuan, and Ying Chung. 1981. Preliminary study of the effect of thiadiazole on
abnormality. Pharmacy Academy Report 16(9):654-660.
Wilson, J.G. 1973. Principles of teratology. Environment and Birth Defects 2:11-34.
Wilson, J.G. 1973. Causes of developmental abnormality. Environment and Birth Defects 3:35-
82.
Wilson, J.G. 1977. Status of teratology: General principles and mechanisms derived from
animal studies. Handbook of Teratology l(2):49-62.
Wilson, J.G. 1977. Environmental chemicals. Handbook of Teratology l(2):357-385.
Yu, Yee Fu, and Jia Juing Mao. 1985. Environmental pollution and human health. Fu Tan
University Press.
Yuan, Chuan Mee, and Jen Hwa Liu. 1977. Fish abnormality and water pollution. Nanjing
University Press (Natural Science edition) 1:99-112.
Yuan, Chuan Mee. 1989. Effects of pollution in Jiang-Su rivers and lakes on fish growth and
resources. Pages 259-265 In Jiang-Su Resources and Environment. Jiang-Su Education
Publishers.
Yung, Hwei Yee, and Yei Sin Yu. 1983. Effects of microwave radiation on abnormality of
mice. Shir-Chuan University Report (Natural Science edition) 1:101-106.
42
-------
ECOTOXICOLOGY OF THE SYNTHETIC DETERGENT LAS
IN THE AQUATIC ECOSYSTEM
by
Zhang, Yong-Yuan1 and F. Korte2
INTRODUCTION
Linear alkylbenzene sulfonate (LAS) is one of the anionic surfactants, and is an active
component in commercial detergents. In recent years, the synthetic detergents industry devel-
oped rapidly in China with an annual production of 1.15 million tons in 1987, of which LAS is
the main product. Owing to the fact that satisfactory facilities for wastewater treatment have
not been established for most cities in China, large amounts of LAS have been sprayed into
rivers, lakes, and other water bodies. Most inland waters in China are being used for fishery
utilization. Therefore, great attention is paid to the effect of LAS on the aquatic environment.
Many reports have been dealing with the biodegradation (Swisher 1972, Willets and Cain 1972*
Gledhill 1975, Divo 1980) and toxicology of LAS (Kikuchi 1978; Kimerle et al. 1981,
Wakabayashi et al. 1981). The purpose of this paper is to study the degradation rates of MC-
labeled LAS under aerobic and anaerobic conditions and the distribution of 14C-labeled LAS in
various ecological compartments in the model laboratory aquatic ecosystem, as well as the
toxicity of LAS to various aquatic organisms, so as to provide a basis of water pollution control
for this pollutant.
METHODS, RESULTS, AND DISCUSSION
THE AEROBIC DEGRADATION OF "C-LAS IN THE AQUATIC ENVIRONMENT
Water samples were collected from Lake Donghu, Wuhan, a typical eutrophic lake. The
variation of LAS degradation rates with renewed dose in lake water was observed. Each dose
was approximately 2 mg/L, Sterilized lake water, containing the same concentration of LAS as
the test group, was used for control. The results of three successive doses are given in Figure 1.
It can be seen that the degradation rates of LAS increase with the increase of number of the
dose in the natural lake water. The time course of LAS concentration for each dose can be
described by the following equations:
P! = iOO[l-0.021t * exp(0.601t)]
P2 = 100[l-0.115t * exp(0.340t)]
P3 = 100[l-0.275t * exp(0.290t)]
Institute of Hydrobiology, Academia Sinica, Wuhan, Hubei Province, PRC.
2Institut fur Okologische Chemie, Gesellschaft fur Strahlen und Umweltforschung, Munchen,
Germany.
43
-------
or as the general expression
P =
exp(k2t)].
The half-life of LAS, calculated from Figure 1 for the first, second and third doses of LAS were
3.2,2.1 and 1.2 days but the concentration of LAS in the control group remained unchanged. It
is obvious that the ability of bacteria to degrade LAS is enhanced when the bacteria are
acclimated by multiple introduction of the substrate into the medium. On the other hand
Yediler et al (1989) reported that the degradation rate of LAS was positively correlated with
the density of bacteria in lake water. Therefore, it might also be connected with the increase of
bacterial density due to renewal of carbon sources by repetitive dose.
TO
3
Half life
of LAS l.a
\
Exposure Time (day)
Figure 1. LAS degradation kinetics hi Dong Hu lake water with renewed doses.
When "C-LAS (approximately 2 mg/L) was dosed in the test apparatus, the radioactivities
of intact "C-LAS, intermediate metabolites of "C-LAS, and MC-carbon dioxide were measured
at different times. During the test period, when the temperature was kept at 25-27° C, most of
C-LAS was degraded after 72 hours and the content of 14C-LAS intermediates rose steeply
(Figure 2). The results from isotope thin layer chromatography assay indicated that on the third
day, two kinds of more polar metabolites were formed with Rf values of 0.58 and 0 27 The Rf
value of LAS was 0.77 (Figure 3).
44
-------
100
ci
o
£ 50
cd
C!
(1)
O
(J
o
o
OD
LAS(1.5-3.5°C)
LAS(25-27°C)
AS
Metabolites(25-27 C)
C02(25-27°G)
Metabolites
-3.5°c)
2 ^ 6 8 10 12 14 16 18
Time (day)
Figure 2. Ultimate biodegradability of 14C-LAS in Dong Hu lake water
at 25-27° C and 1.5-3.5° C.
LAS
Rf 0.77 0.58
0.27
-
a: standard IH"G-LAS
b: tested water sample
Figure 3. Radioisotope TLC chromatogram of 14C-LAS metabolites in water.
45
-------
With the disappearance of the parent LAS in water and an increase in the content of MC-
LAS metabolites, 14C was detected both in the carbonates and carbon dioxide, and the radioac-
tivity of CO2 was gradually increased with the extension of experimental time. Willets and Cain
(1972) suggested that the attack of microorganisms to LAS involved the carboxylation of the
terminal methyl group of the alkyl side chain, and subsequent cleavage of the carbon chain at
every two carbon atoms via p-oxidation pathway, to form p-hydroxybenzoate or p-hydroxy-
phenylacetate. As the ring labelled LAS was used for this test, the results mentioned above
indicated that benzene ring had been cleaved and oxidized to form the ultimate oxidation
product—carbon dioxide. From the results shown in Figure 2, it can also be seen that under the
condition of low temperature of 1.5-3.5° C, the degradation of LAS followed the same rule as at
higher temperatures, but the rate was approximately 20 times lower than that at the tempera-
ture of 25-27° C
THE ANAEROBIC DEGRADATION OF LAS
Because of the lipophilic molecular structures of LAS, it is possible that LAS would be
readily adsorbed onto the remains of microorganisms, which could partly settle down to the
bottom of a water body and reach anaerobic areas. It is necessary to study the anaerobic
degradation of LAS for the overall evaluation of the behavior of LAS in the aquatic ecosystem.
The results indicated that in contrast with aerobic conditions, LAS degraded slowly under the
anaerobic environment; the half life was 57 days (Figure 4). Moreover, it was not found that
the degradation rate was enhanced by acclimating the bacteria with repeated dosing of LAS into
the anaerobic digestion apparatus.
Half life of LAS
20 30 40 50
Time (day)
60 70
Figure 4. Anaerobic degradation kinetics of LAS.
In addition, the variation of LAS concentration was also observed in the wastewater
stabilization ponds of Huang Gang City, Hubei Province, for treating municipal wastewater. The
pond system consists of four units operated in series. The first unit is an anaerobic pond; the
46
-------
others are aerobic ones. Analytical results indicated that the content of LAS did not show
noticeable change in the first pond (Table 1). When the sewage passed through the subsequent
aerobic ponds, there was a marked drop in the concentration of LAS. This result is quite
similar to that obtained from the laboratory model test.
Table 1. Variation of LAS concentration (mg/L) in wastewater stabilization ponds.
Variable
Influent
Effluent of pond
No. 1
anaerobic
No. 2
aerobic
No. 3
aerobic
No. 4
aerobic
Dissolved oxygen 1.5 0.0 4.5 6.1 5.2
LAS 0.26 0.25 0.02 0.01 < 0.01
TRANSFER AND DISTRIBUTION OF 14C-LAS IN A LABORATORY MODEL
AQUATIC ECOSYSTEM
In order to understand the distribution and bioaccumulation behavior of 14C-LAS in the
presence of multiorganisms, we carried out an experiment with a laboratory model aquatic
ecosystem (LMAE) in which zebrafish (Brachydanio redo), daphnid (Daphnia magna), snail
(Bellamya aeruginosa), and an aquatic plant (Vallisneria spiralis) were introduced along with one
dose of 14C-IAS. The bioaccumulation of 14C-LAS and its metabolites in each tested species, as
well as in water, were observed. The results show (Figure 5A) that in the LMAE system, the
time course of LAS and its metabolites concentration in water are
Clw = 2.7exp(-0.159t) (1)
Cmw = 0.363 exp[-0.0161(t-14)2]+0.109t * exp(-0.0714t) (2)
where Clw = concentration of LAS in water (/ig/ml), Cmw = concentration of metabolites
(calculated as /ig/ml LAS), and t = time (day).
It can be seen from Figure 5A and Equation (3), that the LAS concentration in water of
LMAE follows the pattern of an exponential decline and the metabolites reach maximum value
on the 14th day. The results in Figure 5B indicate that the bioaccumulation of 14C-LAS in each
test species in LMAE can apparently be divided into two stages. The first stage is accumulation
of intact LAS in organism, in which the highest concentration is observed on the second day of
exposure. The change of LAS concentration in zebrafish and snail at first stage can be
expressed by the following equations:
Of = 50t * exp(-0.54t) (3)
Ck = 115t * exp(-0.45t) (4)
where Of = concentration of LAS in fish (/tg/g), and Ck = concentration of LAS in snail (/tg/g).
47
-------
•
223
200
173
f-I 7J
° 30
.s
l
B
X -Fish
+ -Daphnia
A -Snail
o -Aquatic plant
0 s « » an as . M- aa 4a 13 so
Exposure Tine (da/)
Figure 5. The relation between concentration of LAS and its metabolites in
organisms and water in laboratory model aquatic ecosystem.
The second accumulation peaks were found on the 20th day in both fish and snail, when
more than 95% of LAS had been degraded. The second peak in organisms lagged behind 6
days, compared with the peak of metabolites in water (Figure 5A). Therefore, it may be
considered that the second accumulation peak is mainly caused by the metabolites of MC-LAS.
The time course of metabolites concentration in fish and snail are
C^r = 110 exp[-0.013(t-20)2]+t * exp(-0.03t)
Cm = 121 exp[-0.02(t-20)2]+t * exp(-0.03t)
where Cnf, Cm = concentration Oug/g) of metabolites in fish and snail, respectively.
(5)
(6)
48
-------
The results indicate that the time course of LAS and its metabolites both in fish and snail
have a similar rule, which can be described by
(7)
(8)
" C| = k,t * exp(-k2t)
Cn = k», expf-k^t-ti^+t *
where C, = concentration of LAS in organisms (jug/g), Q = concentration of metabolites in
organisms fcg/g), t = time (day), tt = time of peak for the first function (day), and k,,
= coefficient '
The bioaccumulation factors of LAS in fish and snail on the second day are 16.7 and 47 7-
BCFs of metabolites in fish and snail on the 20th day are 166.6 and 181.8. '
Accumulation of LAS by aquatic plant in the first stage was lower than other organisms
However, "C-radioactivity increased exponentially since the llth day exposure and the
maximum value was observed on the 30th day; then the radioactivity decreased gradually This
phenomenon may be due to the fact that the metabolites of LAS were further mineralized and
became CO2 which was assimilated by plant via photosynthesis to form cell constituents It was
pointed out in the literature that the toxicity of LAS markedly decreased after the oxidation of
the terminal methyl group of the alkyl chain. Consequently, the harmful effect of the second
accumulation in the aquatic organisms will be lower than that of the first one.
TOXICITY OF LAS TO AQUATIC ORGANISMS AND ACCUMULATION AND
DEPURATION OF LAS BY FISH
The toxicity tests were carried out for the various aquatic organisms, including fish
zooplankton, protozoa, and algae. The results of acute toxicity tests showed that the 96-hr LC50
for grass carp larvae, 48-hr EC50 for D. pulex and Tetrahymena americanis, as well as EC50 for
the activity of fish gill ATPase are 5.9, 7.0, 7.8 and 3.8 mg/L, respectively. But the tolerance of
bcenedesmus obliquus was far more than that of other organisms. No obviously acute effect was
observed when the concentration of LAS was as high as 100 mg/L (Table 2).
Table 2. Acute toxicity test of LAS for aquatic organisms.
Organism
Fish
Grass carp (larvae)
ATPase (gill)
Zooplankton
Daphnia pulex
Protozoa
Tetrahymena americanis
Algae
Scenedesmus obliquus
End Point
LC50
EC50
EC50
LC50
EC50
LAS mg/L
5.9
3.8
7.0
7.8
100.0
49
-------
The accumulation and distribution of "C-LAS were investigated for various tissues and
organs of the fish (carp). The results (Table 3) showed that equilibrium was reached in almost
all of the tissues and organs after 24-hour exposure in 14C-LAS. The accumulation in gall
bladder was an exception, in which the highest concentration of 14C-LAS was found after 72
hours of exposure, and the bioconcentration factor (BCF) was 2485. The BCFs in muscle and
other organs were relatively low. The BCF was 19 in intestines, 12 in liver, and less than 10 in
all of the others. Moreover, LAS accumulated in fish can be depurated rapidly in clear water.
The rate of depuration reached 98.8% after 72 hours in gill and in muscle (Figure 6). Kimerle
(1980) reported that the BCF in bluegill was calculated to be 5000 in gall bladder and 36 in
muscle. This might be due to the difference in the test species.
Table 3. Accumulation and distribution of LAS in carp.
Organs and
tissues
Scale
Skin
Gill
Intestines
Liver
Kidney
Heart
Air bladder
Eye
Gallbladder
Muscle
Backbone
Remain body
24-hr Exposure
Concentration
(mg/kg)
6.76
12.66
2531
51.96
33.19
23.65
24.70
8.13
5.22
2413.43
133
3.56
BCF
4.69
250
937
19.24
12.29
8.76
9.15
3.01
1.86
893.70
0.49
1.27
Distribution
rate (%)
9.29
Z65
2233
133
531
1.49
.0.44
0.28
055
1.71
5434
72-hr Exposure
Concentration
(mg/kg)
7.92
13.48
30.75
91.48
37.23
34.17
22.24
10.33
7.28
6710.90
1.74
2.78
BCF
4.96
2.93
11.38
33.73
13.79
12.66
8.24
3.83
1.93
2485.20
0.64
132
Distribution
rate (%)
2.86
1.47
6.61
6.82
3.86
0.46
0.29
0.19
39.65
37.69
One of the main purposes for investigating the bioaccumulation of environmental
chemicals in fish is to assess the degree of human safety based on ingestion via consumption of
fish exposed to the chemicals. Since LAS can be degraded quickly under aerobic conditions,
generally the concentration of LAS in aquatic environments is low. For example, it remains at
16 jig/L to 300 fig/L in Donghu Lake, Wuhan, which has been polluted by municipal waste-
water (Tan et al. 1985). In this case, the content of LAS in the edible parts of fish will be lower
than 1 mg/kg. It is assumed that 1.0 mg LAS per kg fish muscle and 150 g fish muscle per day
for human consumption, 0.15 mg or 0.0025 mg/kg/day of LAS would be ingested daily (60 kg
calculated for human body weight). The chronic no-effect concentration for rat was 250
mg/kg/day (Buchler et al. 1971). Thus it can be considered that the safety coefficient of LAS to
humans is rather high.
50
-------
10000 .
1000-
'0
24
uptake _
phase
72 ' 24
Time (hour)
72
depuration.
phase\
-I
Figure 6. Accumulation and depuration of LAS in various
organs and tissues in hybrid carp.
CONCLUSIONS
1. Under aerobic conditions, LAS can be metabolized and mineralized to form CO,
rapidly. Under anaerobic conditions, it degrades rather slowly.
2. In a laboratory model aquatic ecosystem, the bioconcentration of LAS took place in
two stages. The first stage was accumulation of intact LAS, while the second one was the
accumulation of the metabolites of LAS.
™™, 3' «?t bioconcentrati°n factor of LAS in muscle of carp was found to be extremely low
(?rL= ^ in Sallbladder il c0"1* reach as high as 2485. Because the bioconcentration
of LAS in the edible parts of fish was low, the safety factor for humans ingesting fish exposed to
LAb in the aquatic environment may be considered relatively high.
51
-------
, ACKNOWLEDGEMENT
This project was a cooperation between the Institute of Hydrobiology, Academia Sinica,
Wuhan, PRC and the Institut fur Okologische Chemie, Gesellschaft fur Stralen und Umweltfors-
chung (GSF), Germany, and was supported by the National Science Foundation, PRC, and GSF,
Germany.
REFERENCES
Buchler, F.V., E.A. Newmann, and W.R. King. 1971. Two-year feeding and reproduction study
in rats with linear alkylbenzene sulfonate (LAS). Toxic. AppL Pharmac. 18:83-91.
Divo, C. 1980. Primary and total biodegradation of linear alkylbenzene sulfonates. Tenside
Deterg. 17(l):30-36.
Gledhill, W.E. 1975. Screening test for assessment of ultimate biodegradability linear alkylben-
zene sulfonates. Appl. Microbiol. 30(6):922-929.
Kikuchi, M., M. Wakabayashi, H. Kojima, and T. Yoshida. 1978. Uptake, distribution and
elimination of sodium linear alkybenzene sulfonate and sodium alkylsulfate in carp.
Ecotox. and Environ. Safety, 2:115-127.
Kimerle, R.A., and R.D. Swisher. 1977. Reduction of aquatic toxicity of linear alkylbenzene
sulfonate (LAS) by biodegradation. Water Res. ll(l):31-37.
Kimerle, R.A., KJ. Macek, B.H. Sleight, and M.E. Burrows. 1981. Bioconcentration of linear
alkylbenzene sulfonate (LAS) in bluegill (Lepomis macrochirus). Water Res. 15(2):251-
256.
Steber, J., and P. Wierich. 1987. The anaerobic degradation of detergent range fatty alcohol
ethoxylates. Studies with MC-labelled model surfactants. Water Res. 21(6):661-667.
Swisher, R.D. 1972. Linear alkylbenzene sulfonate benzene rings biodegradation and acclima-
tion studies. Yukagaku 21:130-140.
Tan, Y.Y., and Y.Y. Zhang. 1985. Determination of LAS homologous series in water by
HPLC. Environ. Chem., 5(5):50-55.
Wakabayashi, M., M. Kiccuchi, A. Sato, and T. Yoshida. 1981. The relationship between
exposure, concentration and bioaccumulation of surfactants. Bull. Japan. Soc. Scient.
Fisheries 47(10):1383-1387.
Willets, A.J., and R.B. Cain. 1972. Microbial metabolism of alkylbenzene sulphonates.
129:389-402.
Yediler, A, Y.Y. Zhang, and J.P. Cai. 1989. Effect of the microbial population size on the
degradation of linear alkylbenzene sulfonate in lake water (Dong Hu = East Lake,
Wuhan, Hubei, P.R. China). Chemosphere 18(7-8):1589-159
52
-------
EFFECTS OF TEMPERATURE AND HYPOXIA ON
CALIFORNIA STREAM FISHES
by
Joseph J. Cech, Jr.1, Stephen J. Mitchell2,
Daniel T. Castleberry3, and Maryann McEnroe4
INTRODUCTION
California is a large state, encompassing 411,000 km2, and has diverse aquatic
environments. With several mountain ranges and a large central valley, many of these aquatic
environments are flowing streams and rivers which support an interesting array of fishes.
Ichthyologists in California (and elsewhere) have had a scientific interest in why particular
species of fish live where they do and which factors affect their distribution. Also, as water
development proceeds in California, resource managers and regulators need a greater
understanding of the environmental requirements of the fish and the other inhabitants of our
state's streams and rivers. For example, low-head hydroelectric dam development can
dramatically change flow, depth, and related variables in small streams. Metabolic responses of
resident fishes to dominant abiotic factors provide a quantitative assessment of their
environmental requirements and predictive value for distributional changes.
The objectives of the present study were to measure respiratory metabolic (oxygen
consumption) rates of seven species of native California fishes in response to environmental
temperature and dissolved oxygen concentration, and to predict longitudinal distributions and
species associations in California streams. We also offer predictions concerning changed fish
species distributions with alteration in stream flows.
MATERIALS AND METHODS
Seven species of native California stream fishes were used in the study: rainbow trout,
Oncorhynchus myfdss (Salmonidae); tule perch, Hysterocarpus trasld (Embiotocidae); Sacramento
sucker, Catostomus occidentals (Catostomidae); riffle sculpin, Cottus gulosus, (Cottidae);
Sacramento squawfish, Ptychocheilus grandis; hardhead, Mylopharodon conocephalus; and
California roach, Hesperoleucas symmetricus (Cyprinidae). Adult fish were captured from several
central California locations and quickly transported to the University of California, Davis,
Department of Wildlife and Fisheries Biology, University of California, Davis, California, USA
2Aquatic Systems, Inc., San Diego, California, USA
3U.S. Fish and Wildlife Service, National Fisheries Contaminant Research Center, Field
Research Station, Dixon, California, USA
4Division of Natural Sciences, State University of New York, Purchase, New York, USA
' 53
-------
laboratory. In the laboratory, they were held from 3 d to 6 wk and acclimated to, usually, three
experimental temperatures: 10, 20, and 30 °C.
Respiratory metabolic (oxygen consumption) rates were measured on fish in acrylic, flow-
through respirometers (Cech et al. 1979). Riffle sculpin, however, was an exception. Because
sculpin would not settle down in the cylindrical respirometers, they were acclimated to 2.9-L
flasks (Cech et al. 1990). After fish adjusted to the respirometer surroundings, simultaneous
measurements were made of dissolved oxygen and water flow through the respirometer.
Dissolved oxygen was measured by either titrimetric (Winkler) or electrometric (Radiometer
PHM71, E5046) methods which gave comparable results. Water flow was measured by the
timed, volumetric collection of water (L/min) multiplied by 60 (min/h). Oxygen consumption
(mg (Vhr) was calculated by multiplying the oxygen concentration difference across the
respirometer (inflowing mg OJL, - outflowing mg O/L) by the water flow rate (L/h).
Oxygen consumption rates were determined at four different environmental conditions for
each fish. After measurements under normoxic conditions (>145 torr PO2, 1 torr = 133.3 Pa) at
the acclimation temperature, water conditions in the respirometer were changed to hypoxia (40
torr PO2) over a 3-h period and oxygen consumption was measured again. Hypoxia was induced
by counterflows of inflowing water and nitrogen in a water equilibration column located
upstream of the respirometers (Cech et al. 1979). After the hypoxia measurements, nitrogen in
the column was replaced with air and normoxic conditions returned. Then, water temperature
was increased 5 °C over a 3-4 h period. After 12 h at the elevated temperature, oxygen
consumption rates were again measured at both normoxic and hypoxic conditions.
Oxygen consumption rate data were analyzed statistically to compare species and
responses to environmental temperature and dissolved oxygen conditions. We used either
paired t-tests with Bonferroni's adjustment for multiple comparisons (Neter et al. 1985) on mg/h
data or either unpaired t-tests (2 data sets) or one-way ANOVA and Duncan's multiple range
test (> 2 data sets) on logio transformations of mass-independent metabolic rates (mg Oa-kg^h'1)
as suggested by Heusner (1984). Further details concerning the methods are given in Cech et
al. (1990).
RESULTS
Under normoxia, all species generally increased their respiratory metabolic rates with an
increase in acclimation temperature (Figure 1). A notable exception to this pattern occurred in
Sacramento squawfish for which data were collected at acclimation temperatures of 15 and
25 °C. Squawfish metabolism showed no significant difference between 15 and 20 °C or between
20 and 25 °C (Figure 1). All species generally increased metabolism after the 5 °C temperature
increase.
Four species showed a similar response pattern to hypoxia at various temperatures: no
effect at a low temperature, metabolic depression at a threshold temperature, and death at the
adjacent higher temperature (Figure 1). Threshold temperatures were 15 °C (after a 5°C
increase) for rainbow trout, 20 (after 5 ° increase) to 30 °C for Sacramento squawfish, and 25 °C
for hardhead and Sacramento sucker. Two species, tule perch and riffle sculpin, showed
54
-------
250-
200-
150-
r* 100-
1
x
50-
i
O
* I50-,
IN
° 100-
~ 50-
LU
h-
5
10 i
4;
:
^
;
!
^ .
!
;
i
$ ! 30° C ACCLIMATION
5|
$1 ?
W fl ^
*i II \im
^mm RIFFLE CALIFORNlA
SOUAWFISH SCULPIN ROACH
T«-
J
4
;
^
^,
5
.,13 T
iSi 5K?
1 » r1 ^ nri 25° C ACCLIMATION
[M $^yy
^a x Bm
^Xx $; (CQ
DC SACRAMENTO TULE PERCH '. '
-, SQUAWFISH
1 CONSUMPTIOh
01 o oi
O O O
,.1.1.1
^
8 150 -i
0 100-
'
50-
ISO-i
100-
-
50-
-
i
i 13^
u
2 19
Y 8 20° C ACCLIMATION
^j2 IsL iJ?, 10 1,
^« I^^&lSl i^-xdwl 1.ISI2T 10
ija r^m \§m f^ffl l^1°7 is121212
"?imi?W ^OM^|KTU° SASJS|NT° HARDHEAD TULE PERCH RIFFLE CALIFORNIA
TROUT SOUAWFISH SUCKER SCULPIN ROACH
I
10 ^
CS;
I
1 0? 1 5° C ACCLIMATION
i V
ll
SACRAMENTO
9 UAWFISH
I 10° C ACCLIMATION
Rj
T ^ 9 11 2 2
I ^^ 11fS11l. 1114*1 3lS2l!&
^'S3 r'^M r^J™ nilli jJi6.li 12 12
TROUT SOUAWFiSH SUCKER
CALIFORNIA
ROACH
Figure 1. Mean (± 2 SE with numbers of fish), mass-independent oxygen
consumption rates for seven species of native California stream fishes
at acclimation temperature and normoxic (blank bar) and hypoxic
(diagonally hatched bar) conditions, and after a 5 °C temperature
increase and normoxic (horizontally hatched bar) and hypoxic
(crosshatched bar) conditions.
55
-------
metabolic depressions at every temperature tested, whereas California roach showed no
metabolic depression at any temperature, including 35 °C!
Rainbow trout had the highest mass-independent metabolic rates while riffle sculpin and
California roach had the lowest rates (Figure 1). Metabolic rates of the other four species were
typically within these extremes and were typically indistinguishable statistically (Figure 1).
Further details concerning the results are given in Cech et al. (1990).
DISCUSSION
Moyle and Nichols (1973) and Moyle (1976) described native fish associations in central
California streams by elevation-related zones and corresponding habitats. Rainbow trout
dominated the "rainbow trout zone" consisting of high elevation, cold (< 21 °C) permanent
reaches.
Currently, several of the rainbow trout zone streams in the Sierra Nevada mountains have
been occluded with low-head hydroelectric dams and water diversion (e.g., for agricultural
irrigation) structures. These structures can dramatically change stream flow, temperature, and
dissolved oxygen characteristics. For example, summer reductions in flow can increase exposure
time to solar radiation and to warm air temperatures, thereby wanning the stream. How
interruption from low-head hydropower development on Haypress Creek has dewatered large
sections of the former streambed (N. Erman, University California, Berkeley, unpublished data).
In addition, stream sedimentation from the Haypress Creek project construction has been
associated with thick diatom blooms in the low-water stream (N. Erman, personal observations).
Large blooms of plant material are known to result in dramatic, nocturnal decreases in dissolved
oxygen concentration (Boyd 1979). Thus, reduced stream flows and increased nutrient inputs
may result in eutrophic waterways, altering both temperature and dissolved oxygen. Because
rainbow trout cannot exist for long in even moderately warm (20 °C) hypoxic water (Figure 1),
they will be susceptible to these sorts of diversions.
According to Moyle and Nichols (1973), Sacramento squawfish, Sacramento sucker, and
hardhead dominated the "squawfish-sucker-hardhead zone" in larger, low elevation streams.
These species showed the same metabolic pattern as rainbow trout regarding their response to
hypoxia, but at warmer temperatures. These species showed an increased sensitivity to low
dissolved oxygen after an abrupt 5 °C increase at temperatures between 10 and 25 °C, suggesting
that they would be more commonly found in thermally stable, large pool habitats. Available
field studies support this prediction (Smith 1977, Moyle and Baltz 1985, Baltz et al. 1987).
Finally, Moyle and Nichols (1973) stated that California roach dominated the "California
roach zone" characterized by warm (possibly > 30 °C), intermittent tributaries of larger streams.
The California roach's small size (Moyle 1976), low oxygen consumption rates, and hardy
tolerance to high temperature and low dissolved oxygen extremes (Figure 1) should make it a
survivor in these harsh environments. Field data of Smith (1977) Moyle et al. (1982) show this
fish thrives in streams too warm (> 35 °C) for other species.
Tule perch and riffle sculpin also live in California streams but do not have zones
designated for them. Both of these species showed metabolic depressions at every temperature
56
-------
tested. We predict that both species should be found in well-oxygenated environments. Moyle's
review (1976) supports this prediction, but shows that they seldom occur together. The tule
perch is laterally compressed and lives in low velocity areas of deep streams, whereas the sculpin
is dorso-ventrally compressed and lives on the bottom of higher velocity areas (Moyle 1976).
We suspect that their widely divergent body shapes help explain their different distributional
patterns (Cech et al. 1990).
In conclusion, metabolic responses of adult, California stream fishes to temperature and
hypoxia followed patterns which help explain species' distribution and abundance. Correlations
show that species' distribution in streams is greatly influenced by stream temperature and
dissolved oxygen concentrations (Cech et al. 1990). In many species of fish, the larval stages are
more sensitive to hypoxia than are the adults (reviewed by Rombough 1988). If this is also true
for California stream fishes, then the environmental requirements for successful reproduction in
these species would be more stringent than those reported here. Changes in California stream
environments, such as low-head hydroelectric power development, which disrupt or drastically
reduce water flow will be expected to alter stream temperature and dissolved oxygen levels.
Such changes should make streams less hospitable for species with narrower environmental
requirements (e.g., rainbow trout) and more hospitable for species with wider environmental
requirements (e.g., California roach).
REFERENCES
Baltz, D.M., B. Vondracek, L.R. Brown, and P.B. Moyle. 1987. Influence of temperature on
microhabitat choice by fishes in a California stream. Trans. Amer. Fish. Soc. 116:12-20.
Boyd, C.E. 1979. Water quality in warmwater fish ponds. Agricultural Experiment Station.
Auburn University. Auburn, Alabama.
Cech, J.J., Jr., C.G. Campagna, and S.J. Mitchell, 1979. Respiratory responses of largemouth
bass (Micropterus salmoides) to environmental changes in temperature and dissolved
oxygen. Trans. Amer. Fish. Soc. 108:166-177.
Cech, J.J., Jr., SJ. Mitchell, D.T. Castleberry, and M. McEnroe. 1990. Distribution of
California stream fishes: influence of environmental temperature and hypoxia. Env. Biol.
Fish. 29:95-105.
Heusner, A.A. 1984. Biological similitude: statistical and functional relationships in comparative
physiology. Amer. J. Physiol. 246:(Reg. Integ. Comp. Physiol. 15):R839-845.
Moyle, P.B. 1976. Inland fishes of California, Univ. California Press. Berkeley, California.
Moyle, P.B., and D.M. Baltz. 1985. Microhabitat use by an assemblage of California stream
fishes: developing criteria for instream flow determinations. Trans. Amer. Fish. Soc.
114:695-704.
Moyle, P.B., and R.D. Nichols. 1973. Ecology of some native and introduced fishes of the -Sierra
Nevada foothills in Central California. Trans. Amer. Fish. Soc. 114:695-704.
57
-------
Moyle, P.B., JJ. Smith, R.A. Daniels, and D.M. Baltz. 1982. Distribution and ecology of stream
fishes of the Sacramento-San Joaquin drainage system, California~a review. Univ.
California Publ. Zoology 115:227-256.
Neter, J., W. Wasserman, and M.H. Kutner. 1985. Applied linear statistical models: regression
analysis of variance and experimental designs. Second edition. Irwin Press. Homewood,
Illinois.
Rombough, P J. 1988. Respiratory gas exchange, aerobic metabolism, and effects of hypoxia
during early life. In: Fish Physiology, VoL 11A, W.S. Hoar and DJ. Randall (Eds.),
Academic Press, New York, New York.
Smith, JJ. 1977. Distribution, movements and ecology of the fishes of the Pajaro River system,
California. Ph.D. Diss. Univ. California, Davis. 230 p.
58
-------
EFFECTS OF VARYING WATER pH ON GILL FUNCTION
by
Hong Lin1 and David J. Randall1
INTRODUCTION
The gills of fish are the primary site of gas exchange and acid-base and ion regulation.
Carbon dioxide, ammonia and acid are excreted into the water passing over the gills. Some of
the ion transport pathways, through which these excretory products are eliminated, are sensitive
to external water pH. The reaction of excreted carbon dioxide and ammonia with water is pH
dependent. Therefore, varying water pH might affect the excretory processes. This review
discusses some recent evidence for the presence of a pH sensitive electrogenic proton pump in
gill epithelium and the effects of water pH on acid, carbon dioxide and ammonia excretion.
GILL STRUCTURE AND ION TRANSFER MECHANISM
Two rows of filaments arise from each gill arch. Each filament has an upper and a lower
row of lamellae, which represent the respiratory portion of the gill. Water flows in thin sheets
between the lamellae. Blood perfusing the gills flows along the afferent and efferent vessels
inside the filament and the lumen inside the lamellae.
The lamellar epithelium separating the blood from the external water consists of a pillar
cell flange, basement membrane and two layers of epithelial cells plus a few chloride cells
(Figure 1A) (Randall 1990). In freshwater fish the epithelial and chloride cells are joined by
tight junctions, which act as a minimal barrier for diffusion of gases, but have a high resistance
to the transfer of ions and water. The basolateral side facing trie blood is much more
permeable to ions and water than the apical side facing the water.
Since the lamellar epithelium is permeable to oxygen, CO2 and NH3, carbon dioxide and
ammonia are excreted mainly in the form of CO2 and NH3 by passive diffusion (Figure IB). Ion
transfer across the gill epithelium, however, may be mediated by electroneutral exchange
processes or active transport processes (Figure IB). These have been postulated to include:
(1) Cl'/HCCv exchange occurring in the apical membrane, through which chloride is absorbed
and bicarbonate excreted. Although no more than 10% of the total carbon dioxide is excreted
by this means, it is a primary pathway for chloride uptake. SITS, an anion transport inhibitor,
blocks this exchange pathway (Perry and Randall 1981). (2) Na+/(H+)NH4+ exchange occurring
in the apical membrane, which was believed to be the major pathway for sodium uptake and
proton excretion and an optional excretory pathway for ammonia. Amiloride, a very potent and
Department of Zoology, University of British Columbia, Vancouver, British Columbia, Canada.
59
-------
specific inhibitor of sodium transport in a wide variety of cellular and epithelial systems, inhibits
the Na+/(H+)NH4+ exchange process (Perry and Randall 1981, Wright and Wood 1985).
(3) Na-VK+'ATPase in the basolateral membrane. K+ can be replaced by H+ or NH4+ (Evans et
al. 1989). This active transport pathway can be inhibited by amiloride when applied in mucosal
site in high concentration (Knauf et al. 1976, Kleyman and Cragoe 1988). (4) H+- translocating-
ATPase, or electrogenic proton pump and a sodium conductive channel in apical membrane.
The proton pump will remove protons from the cell and generate a negative potential in the
inner side of the apical membrane. Sodium influx is driven by the negative potential via the
sodium channel which is amiloride sensitive. The H+-translocating-ATPase can be inhibited by
vanadate (Ehrenfeld et al. 1985, Arruda et al. 1981).
WATER (BRANCHIAL EPITHELIUM I PLASMA
Na+
^r -ir
K+ (H+, NH4+)
HCO3-
(A)
Figure 1. (A) Cross section through the gill lamellae blood-water barrier of a fish.
(B) Simplified cross section of gill epithelium of freshwater fish. Gas and
ion transport mechanism is indicated. (See text for details.)
It is well documented that proton transport in mammalian kidney, amphibian urinary
bladder and frog skin is mediated by an electrogenic proton pump (Steinmetz 1985, Al-Awqati
1978, Ehrenfeld et al. 1985). Passive sodium uptake from the mucosal side is indirectly coupled
to this active electrogenic proton transport system. Little was known about the electrogenic
proton pump in the gill epithelium until recently; when we were studying the mechanism of the
expired water acidification in rainbow trout, we gathered some evidence for the presence of an
electrogenic proton pump on the gill epithelium (Lin and Randall 1990). It turned out that this
proton pump is the major source of acid added to the water passing over the gills, and Na+/H+
exchange played a very insignificant role. As a result we modified our model (Figure IB),
developing a new model of ion transfer in the gill epithelium (Figure 2).
60
-------
BLOOD
Figure 2. Hypothetical model of freshwater rainbow trout gill epithelium.
(See text for details.)
Since the active transport processes are energy consuming, they are more likely to happen
in the chloride cell, where mitochondria is abundant (Avella et al. 1987). Ammonia ion enters
the cell through the Na+-K+-ATPase pathway and converts to NH3 will diffuse out of the cell
and the proton will be excreted via the proton pump. The environmental water pH has a key
effect on the operation of this active pump.
EVIDENCE FOR PROTON PUMP MODEL AND EFFECTS OF pH ON THE PUMP
Evidence of the presence of an electrogenic proton pump on the gill lamellae of rainbow
trout, Oncorhynchus mykiss, was provided by our studies of expired water acidification. In the
experimental setup the fish was resting in a recirculating system (Figure 3). An opercular
cannula was fixed near to the opercular opening to sample expired water. Total CO^ total
ammonia, pH and buffer capacity of the inspired and expired water were measured. CO2,
ammonia and hydrogen ion excretion rates and ion concentrations in inspired and expired water
were calculated from these data. Amiloride, SITS and vanadate were added to the external
water to estimate the contribution of different ion transfer pathways to the excretory process.
Water pH was changed to examine its effect on excretion.
61
-------
pH •lactroctes
mixing
reservoir
chamber
with stirring bap
I* CO>»x
Anun«x
COiin
Ammin
cooling coil
Figure 3. The recirculating system with a black, chamber. Fish were prepared with
an opercular cannula. Inspired and expired water samples were
withdrawn from the outlets of the glass chambers.
Experimental results showed that when water pH was below 6, net H+ excretion was
around zero (Figure 4). Net proton excretion was inhibited by low pH. In neutral pH, however,
there was a net proton excretion across the gill epithelium which was unaffected by SITS
treatment, indicating Q-/HCO3- exchange plays in a minor role in this process. This net proton
excretion was not significantly inhibited by 0.1 mM amiloride. Increasing amiloride
concentration to 0.5 mM and 1 mM in the external media caused a reduction in proton
excretion (Figure 5). But more than 50% of the net proton excretion was still sustained even in
1 mM amiloride treatment. It was known that the Na+/H+ electroneutral exchange process is
blocked by low concentrations (1 juM-100 ^M) of amiloride (Benos 1982).
Perry and Randall (1981) demonstrated that 84% of the Na+ uptake by the gill of intact
rainbow trout was inhibited by 0.1 mM amiloride in external media. In our studies, Na+ flux was
not monitored, but net proton excretion was not inhibited by 0.1 mM amiloride. This indicates
that Na+/H+ exchange was not the mechanism of proton excretion. Similar to proton excretion,
ammonia excretion was not inhibited by 0.1 mM amiloride (Figure 6), indicating that NaVNH4+
exchange was not an excretory pathway for ammonia.
However, both ammonia excretion and net proton excretion were inhibited by 0.5 and
1 mM amiloride (Figures 5, 6). Although much less sensitive to amiloride than the Na+ channel,
Na+-K+-ATPase in the basolateral membrane can be inhibited by amiloride when applied to the
mucosal side in concentrations higher than 0.1 mM (Knauf et al. 1976, Kleyman and Cragoe
1988). Thus, the reduced ammonia excretion in 0.5 mM and 1 mM amiloride treatments can be
accounted for by the inhibitory effect of amiloride on Na+-K*(NH4+)-ATPase in the basolateral
62
-------
150
-50
CDControl BB 0.000 1 M Amlloride
.000 1M SITS
Figure 4. Net proton excretion across the gill epithelium of rainbow trout exposed
to different external water pH under control, amiloride, and SITS
treatments.
150
0)
0.0001 M
Amiloride
0.0005M
Amiloride
0.001 M
Amiloride
CHI Control
Treatment
Figure 5. Net proton excretion across the gill epithelium of rainbow trout exposed
to different concentrations of amiloride. ""Indicates a significant
difference at a = 0.05 between the control and treatment values.
63
-------
400
0.0001 M
Amiloride
0.0005M
Amfloride
0.001 M
Amiloride
0.0001M
Vanadate
CZl Control
Treatment
Figure 6. Ammonia excretion rates of rainbow trout under control, amiloride, and
vanadate treatments. *Indicates a significant difference at a = 0.05
between control and treatment values.
membrane. The deprotonation of NH4+ could supply the proton pump, and inhibition of this
NH4+ transport into the cell could cause a decrease of proton excretion, as well as ammonia
excretion. If we compare the ammonia excretion rate with the net proton excretion rate under
amiloride treatments (Table 1), we can see that the reduction of ammonia excretion was
equivalent to that of proton excretion, indicating the NH3 and protons could both originate from
NH<+ transported into the epithelium via Na+-K+(NH4+)-ATPase in the basolateral membrane.
Table 1.
Comparison between reduction of ammonia excretion and reduction of net proton
excretion under amiloride treatment.
Amiloride Concentration
(M)
0.0001
0.0005
0.001
Reduction of Ammonia
Excretion
GuM/kg-h)
47.968±20.568
130.542±9.346
280.560±35.561
Reduction of Net
Proton Excretion
(/iM/kg.h)
60.806±36.178
143.169±86.911
298.105±86.911
Net proton excretion was converted to ju,M/kg.h by assuming ventilation rate = 100 ml/min
(Lin and Randall 1990).
64
-------
Treatment by 0.1 mM vanadate resulted in a reduction of net proton excretion across the
gill by more than 50% (Figure 7). The similar inhibitory effect of vanadate on proton transport
mediated by an electrogenic proton pump was demonstrated in frog skin and turtle urinary
bladder (Ehrenfeld et al. 1985, Arruda et al. 1981). Since applying vanadate on the outer
surface did not affect the Na+-K+-ATPase in the basolateral membrane of frog skin (De Sousa
and Grosso 1979), the vanadate had no effect on proton backleak or bicarbonate excretion in
turtle bladder (Arruda et al. 1981), we conclude that the reduction of proton excretion was
induced by the inhibitory effect of vanadate of H+-translocating-ATPase, and the electrogenic
proton pump is the major pathway of proton transport in fish gills. The fact that vanadate had
no effect on ammonia excretion (Figure 6) indicates ammonia and proton excretion occur
through different pathways.
500
I i
rt^
|
400-
300-
200-
100-
pH7-8
CU Control
pH8-9
I Vanadate Treated
Figure 7. Net proton excretion across the gill epithelium of rainbow trout exposed
to neutral or alkaline water containing 0.1 mM vanadate. ""Indicates a
significant difference at a = 0.05 between the control and treatment
values.
When we plotted the net proton excretion against the inspired water pH, we obtained a
regression line fitting Michaelis-Menten kinetics (Figure 8). In inspired water pH below 5.5,
proton excretion was completely inhibited, due to the large hydrogen ion electrochemical gradient
against the proton pump. Over the pH range of 5.5 to 8, proton excretion increased linearly with
pH. Evidence from studies of turtle urinary bladder showed that the instantaneous
changes in the proton transport rate induced by varying pH involved changes of intrinsic
properties of the proton pump but not the number of pumps (Steinmetz 1986). Proton excretion
was saturated when inspired water pH was above 8, when all the active pumps are operating in
their maximal speed. The substantial proton excretion across the gill epithelium in neutral and
alkaline environment will cause acidification of expired water (Lin and Randall 1990).
65
-------
350
6 7
Inspired water pH
8
Figure 8. The relationship between the net proton excretion through the gills of
rainbow trout and the inspired water pH was expressed by a five-degree
regression curve (R value = 0.853).
McWilliaras and Potts (1978) found that exposing fish to a low pH environment was
associated with a reversal of transepithelial potential, which can easily be explained by our proton
pump model. In neutral pH, the proton pump generates a negative potential on the inner side of
the apical membrane, so transepithelial potential appeared negative. As external water pH
decreased, proton pump activity decreased as the electrochemical gradient against the proton
pump was increased. As a result, less and less negative potential was generated and TEP
switched from serosal negative to' serosal positive. All these factors indicate that proton excretion
across the gills was carrier-mediated and the electrochemical gradient for protons across the
membrane is a fundamental regulator of proton transport.
EFFECTS OF pH ON CARBON DIOXIDE AND AMMONIA EXCRETION
AND EXPIRED WATER ACIDIFICATION
Besides proton excretion, fish also excrete molecular CO2 and NH3 into the water passing
over the gills. The excretion rates are an order of magnitude greater for CO2 than NH3.
Theoretically, the excreted CO2 can be hydrated and form bicarbonate and carbonate in water.
These reactions are catalyzed by carbonic anhydrase in the gill mucus and water boundary layer,
and as a result water passing over the gill will be acidified (Wright et al. 1986). Since CO2
hydration/dehydration reactions are pH dependent, the CO2/HCO3- ratio changes according to
pH. In neutral and alkaline water most of the excreted CO2 will be converted to bicarbonate, but
under acid conditions only a small fraction will be hydrated. Thus, the extent of acidification of
expired water caused by CO2 hydration should decrease with water pH. Since the conversion of
CO2 to HCO3" will reduce the POCQ in the boundary water layer, thereby enhancing CO2 diffusion,
increased environmental water pH might facilitate CO2 excretion.
66
-------
Excreted NH3, on the other hand, can bind a proton to form ammonium ions in aqueous
solution and the rate of this reaction is extremely rapid. This reaction consumes protons in the
boundary water layer, so alkalization of expired water might result. The reaction is also pH
dependent and the NHj/NIV ratio increases with increasing pH. Under acid and neutral
conditions almost all the NH3 is converted to NH4+, but under alkaline conditions only a fraction
of the excreted NH3 forms NH4+ in gill water. Decreased external water pH might reduce P^ in
gill water and enhance NH3 excretion.
Experiments were designed to test these hypotheses. In these experiments we exposed fish
to pH 10 water and pH 4 water and examined the effect of varying pH on the expired water pH
and CO2 and ammonia excretion rates (Lin and Randall 1990).
Expired water pH was plotted against inspired water pH (Figure 9). In the high pH
treatment, expired water was acidified to a greater extent in alkaline water than in neutral water.
More protons were added to the water in alkaline conditions than in neutral conditions. In the
low pH treatment, on the other hand, water was alkalized instead of acidified as it passed over
the gills. Protons were consumed instead of generated in expired water.
Inspired water pH
o pHex (Mean ± SE)
Figure 9. The relationship between pH of exhalant water (o) and inhalant water of
rainbow trout. The exhalant water line is the regression line of the mean
pH values collected from both high and low pH treatments (r2 = 0.%).
67
-------
When we analyzed the ionic composition of inspired and expired water, we found that the
bicarbonate concentration in the inspired water was always higher than that of expired water in
neutral environment (Figure 10), which means there was actually no CO2 converted to HCO3- as
water passed over the gills. Instead, bicarbonate was dehydrated due to the net proton
excretion via the electrogenic proton pump. In the high pH environment, the proton pump was
operating at its maximal speed, therefore expired water was acidified to a greater extent
compared to that in a neutral environment. In the low pH environment, proton excretion ceased
because of the inhibitory effect of low pH on the proton pump; little CO2 would be hydrated at
this pH and all the excreted NH3 would form NH4+, so protons were consumed and pH of the
water was increased as it flowed over the gills.
500
.0
£
•+->
400--
o o 300
Q) C
b
c
o
I
o
m
200--
100
I—I Inspired water
•• Expired water
1
6.37 6.65 6.80
Inspired water pH
7.25
Figure 10.
Bicarbonate concentration of inspired and expired water of rainbow
trout in control experiments. Data are calculated from the mean
values of appropriate pH and total CO2 content.
Total carbon dioxide excretion was unaffected by environmental pH (Table 2), although
external water pH might affect the net proton excretion, and therefore affect the Pcca in
boundary layer. The C1-/HCO3- exchange across the red cell membrane is the rate-limiting step in
carbon dioxide excretion (Perry et al. 1982), and this appears to be unaffected by changes in
water pH.
Ammonia was accumulated within the body of the fish during both high and low pH
treatments (Table 2), indicating ammonia excretion was reduced by both alkaline and acid
environments. In the high pH treatment, NH3 protonation in the boundary water layer was
constrained due to the shortage of hydrogen ion in water. In the low pH treatment, proton
excretion was inhibited, which will reduce epithelial pH and trap ammonia as NH4+ in the cell
and decrease ammonia excretion.
68
-------
Table 2. Carbon dioxide excretion and plasma ammonia levels during high and low pH
treatment in rainbow trout.
High-pH
Low-pH
N
7
6
Control
3.18±0.89
2.03 ±0.14
Experiment 1
2.91 ±0.74
2.39±0.46
Experiment 2
2.27±0.45
2.05 ±0.44
Recovery
2.46+0.28
1.85+0.35
Ammonia in Plasma (/iM/L)
High-pH
Low-pH
6
6
48.9±8.2
57.3+14.9
85.0±15.4*
173 ±47.0*
81.8+17.6*
213±49.2*
54.4+11.1
95.4+15.3
Figure 9 showed that water was acidified in neutral and alkaline environments and
alkalized in acidic environments as it passed over the gill. At environmental pH around 5.6, there
was no change of pH as it flowed over the gills. This is to be expected if the ratio of NHs/CO2
excretion is 0.1 and only the undissociated species are excreted across the gill into the water.
Inspired water pH varied from 3.88 to 9.91, but expired water pH varied only from 4.33 to 7.10.
The overall result of proton, CO2 and NH3 excretion is to ameliorate the magnitude of the change
in water pH next to the gills in the face of changes in pH of the environmental water. Proton
excretion, which is mediated by the electrogenic proton pump and regulated by external water
pH, helps to maintain a relatively stable pH in the gill micro-environment.
CONCLUSION
The acidification of expired water in freshwater rainbow trout is caused by proton excretion
mediated by an electrogenic proton pump in the gill epithelium. This active proton pump is pH
sensitive. Its activity is inhibited in water pH lower than 5.5, saturated in pH higher than 8, and
increased linearly with pH in the water pH range from 5.5 to 8. Water pH had no effect on
carbon dioxide excretion, but ammonia excretion is reduced by both high and low pH. The
overall results of proton, carbon dioxide and ammonia excretion helped to maintain a relatively
stable pH in the gill micro-environment.
REFERENCES
Al-Awqati, Q. 1978. H+ transport in urinary epithelia. Am. J. Physiol. 235 (Renal Fluid
Electrolyte Physiol. 4): F77-F88.
Arruda, J.A.L., S. Sabatini, and C. Westenfelder. 1981. Vanadate inhibits urinary acidification by
the turtle bladder. Kidney International 20:772-779.
69
-------
Avella, M., A. Masoni, M. Bornancin, and N. Mayer-Gostan. 1987. Gill morphology and sodium
influx in the rainbow trout (Salmo gairdneri) acclimated to artificial freshwater
environments. J. Exp. Zool. 241:150-169.
Benos, D J. 1982. Amiloride: a molecular probe of sodium transport in tissues and cells. Am. J.
Physiol. 242 (Cell Physiol. 11):C131-C145.
De Sousa, R.C., and A. Grosso. 1979. Vanadate blocks cyclic AMP-induced stimulation of
sodium and water transport in amphibian epithelia. Nature 279:803-804.
Ehrenfeld, J., F. Garcia-Romeu, and B.J. Harvey. 1985. Electrogenic active proton pump in
Rana esculenta skin and its role in sodium ion transport J. Physiol. 359:331-355;
Evans, D.H., KJ. Moore, and S.L. Robbins. 1989. Modes of ammonia transport across the gill
epithelium of the marine teleost fish Opsanus beta. J. Exp. Biol. 144:339-356.
Kleyman, T.R., and E.J. Cragoe, Jr. 1988. Amiloride and its analogs as tools in the study of ion
transport. J. Membrane Biol. 105:1-21.
Knauf, H., B. Simon, and U. Wais. 1976. Non-specific inhibition of membrane-ATPase by
amiloride. Naunyn-Schmiedeberg's Arch. Pharmacol. 292-189-192.
Lin, H., and D J. RandalL 1990. The effect of varying water pH on the acidification of expired
water in rainbow trout. J. exp. Biol. 149-149-160.
McWilliams, P.G., and W.T.W. Potts. 1978. The effects of pH and calcium concentrations on
gill potentials in the brown trout, Salmo trutta. J. Comp Physiol. 126-277-286.
Perry, S.F., P.S. Davie, C. Daxboeck, and D.J. Randall. 1982. A comparison of CO2 excretion in
a spontaneously ventilating blood-perfused trout preparation and saline-perfused gill
preparation: contribution of the branchial epithelium and red blood cell. J. Exp. Biol.
101:47-60.
Perry, S.F., and DJ. Randall. 1981. Effects of amiloride and SITS on branchial ion fluxes in
rainbow trout (Salmo gairdneri). J. Exp. Zool. 215:225-228.
Randall, D J. 1990. Control and co-ordination of gas exchange in water breathers. Pages 253-
278 In: Advances in comparative and environmental physiology, Vol. 6, R.G. Boutilier
(Ed.). Springer-Verlag, Berlin, Heidelberg.
Steinmetz, P.R. 1985. Epithelial hydrogen ion transport. Pages 1441-1458 In: The Kidney:
Physiology and Pathophysiology, D.W. Seldin and G. Giebiach (Eds.). Raven, New York.
Steinmetz, P.R. 1986. Cellular organization of urinary acidification. Am. J. Physiol. 251 (Renal
Fluid Electrolyte Physiol. 20): F173-F187.
Wright, P.A., T.A. Heming, and D.J. Randall. 1986. Downstream pH changes in water flowing
over the gills of rainbow trout. J. Exp. Biol. 126:499-512.
Wright, P.A., and CW. Wood. 1985. An analysis of branchial ammonia excretion in the
freshwater rainbow trout: effects of environmental pH change and sodium uptake
blockage. J. Exp. Biol. 126:329-353.
70
-------
EFFECTS OF ZINC ON RESPIRATORY PHYSIOLOGY
OF FISH (TILAPIA SP.) AND
EXPLORATION OF DETOXIFICATION
by
Chai, Min-Juan1, Zhou, Xue-Cheng1, and Huang, Yu-Ling1
ABSTRACT
This paper, by exposing unanesthetized fish to solutions containing 1, 4, 8 and 12 ppm
ZnSO4 in either soft or hard water, and recording the opercular movement of naturally
swimming fish, deals with the effects of different concentrations of Zn2+ on respiratory and
cough frequency, and the detoxification function of EDTA! The results are as follows:
(1) Respiratory frequency and cough responses are obviously affected by Zn2* concentrations;
the higher the Zn2* concentration, the further the curve of respiratory frequency deviated from
the standard, and the higher the frequency of cough. These responses may be used as physio-
logical monitoring indexes of toxic symptoms and detoxification in fish. (2) Zn2+ poison to fish
in hard water is much less serious than that in soft water. (3) EDTA can relieve or lighten toxic
symptoms of fish, although its detoxicating effect shows no obvious differences between hard and
soft water. The suggestion is that hard water may be used as a considerably cheap substance of
detoxification.
INTRODUCTION
As a trace element, zinc has something to do with the activity of enzyme and the synthesis
of protein in living organisms. In natural environments, zinc is a pollutant which cannot be
degraded, but is easily accumulated by aquatic animals. Zinc, therefore, threatens the lives of
aquatic animals. For a long time people have studied the effects of Zn2+ on fish lives. Lloyd
(1960) studied the toxicity of Zn2+ to rainbow trout Crespo (1981), Satchell (1982), and Sellere
(1975) showed changes in ventilation, oxygen tension, and pH in the dorsal aortic blood of fish
due to Zn2+ and Cu2+, and Skidmare (1970) showed distribution in excretory organs when fish
were exposed to zinc solutions. Huang et al. (1988) have recently observed the effects of some
heavy metal ions on the respiration of fish. But so far there are few reports about continual
measure of respiratory changes while the fish are exposed to various sublethal Zn2+ concentra-
tions for long periods, and about the changes used as physiological monitoring index to explore
the function of toxification and detoxification of fish. In this paper, we recorded the changes in
respiratory frequency and cough responses of Tilapia sp. exposed to four different sublethal
concentrations of Zn2+ in either soft or hard water during 15 days. We observed detoxification
effects of EDTA complex on fish. We hope to find a way to relieve or lighten the toxic
symptoms of fish poisoned by zinc and hope to offer some suggestions for fish cultivation.
Department of Oceanography, Xiamen University, Xiamen, PRC.
71
-------
METHODS AND MATERIALS : ,
The fish (Tilapia sp.) were obtained from Furong Lake in Xiamen University. Fifty-two
specimens (mean weight 100±6 g, either male or female) were used. These fish were tempo-
rarily fed in round pond in the laboratory from 10 to 15 days before experiments, and were not
given food during the experiments in order to avoid pollution.
The experiment consisted of 12 groups of fish. One test period was 15 days. Each group
was treated by the chemicals shown in Table 1. The Zn2+ doses were respectively 10, 40, 80 and
120 times as much as Chinese Fishery Standard dose (0.1 ppm). The first six groups were kept
in soft water, the quality of which was the same as the feeding water. For the next six groups
the water was hard and was saturated with a solution of CaCO3.
Table 1. Constitution of the experimental water.
Number/
Chemical
ZnSO* (ppm)
CaCO3 (S.S.)*
EDTA (ppm)
1
0
0
0
2
1
0
0
3
4
0
0
4
8
0
0
5
12
0
0
6
12
0
60
7
0
S.S.
0
8
1
S.S.
0
9
4
S.S.
0
10
8
S.S.
0
11
12
S.S.
0
-
12
12
S.S.
60
*Saturated solution of CaCO,
EDTA was added to the 6th and 12th groups on the second day after the toxic symptoms
of fish appeared. The water temperature was 23 ± 1° C and pH was 5.7-6.3. The water was
continually aerated by a small pump during the experiment to maintain sufficient dissolved
oxygen.
The mechanical movement of operculum of unanesthetized fish, which swam unrestrained
in a 15-L aquarium, were transferred into electric signal by semiconductor strain gauge under
water. These electric signals were transferred to digital quantity through analog-digital conver-
sion circuit and then sent to computer to draw waves of opercular movement. When the data of
respiratory movements and cough responses were sent to computer, a smooth curve was drawn
according to Legendre orthogonal polynomial:
Y=f(x) =
p=0
The curves will accurately reflect the respiratory movement changes of fish poisoned by Zn2+
(Huang 1987).
72
-------
RESULTS
EFFECTS OF Zn2* ON RESPIRATORY MOVEMENT OF TILAPIA
The normal respiratory frequency of TUapia was about 77 times mur1, which was the average
value of the normal frequency of 228 measurements from 52 fish. This frequency was consid-
ered as the standard (100%), and the comparison of the changes of respiratory movement
measured for each group during 15 days are shown in Figure la. The respiratory frequency of
fish in all test groups decreased except in the control group, in which frequency fluctuated up
and down the average value. The decreasing tendency was inversely proportional to the
increase of concentration of zinc (see Figure la). This phenomenon is similar to the toxic
symptoms of fish by copper ions in sublethal concentration (Chai et al. 1990).
110
100
90
80
70
60
3.0
E 2. 0
1.0
itroU
10
15
5 - 10
Exposure time (day)
15
Figure 1. Respiratory changes of toxic fish in soft water during 15 days.
(a) Curves of respiratory frequency; (b) Curves of cough frequency.
Coughing response is a major index for Zn2+ toxic fish. It is a protective reflex of fish under
an unfavorable environment. It often occurs after a series of normal opercular movements.
The appearance of coughing and the change of frequency are closely connected with Zn2+
concentration. Figure Ib shows that cough frequency increased and the peaks of its curves
shifted forward as the Zn2+ concentration increased.
73
-------
r
DETOXinCATION EXPLORATION
(1) Exploration of detoxification
After the test fish adapted themselves to hard water for 1 or 2 days, the same four different
doses of ZnSO4 that were added to the soft water were added to the hard water. Although toxic
symptoms of fish become more serious with exposed time and the increase of Zn2+ concentration,
the phenomenon was significantly different from that in soft water. The respiratory frequency
reduced after fish contacted pollutant. But it rose again after the third day (see Figure 2). This
showed that the respiration of fish was recovering gradually. There was no such phenomenon in
soft water groups except for the 1 ppm group, the curves of which rose again at the sixth day (see
Figure la). The result suggested that hard water might help fish resist Zn2+ toxicity.
110
$100
§ 90
f
80
5 10
Exposure time (day)
IS
Figure 2. Respiratory changes of toxic fish in hard water during 15 days.
3.0 '
,«s
I 2.0
S«y
&
I
1.0
5 10
Exposure time (day)
15
Figure 3. Comparison of cough frequency of toxic fish between soft and hard
water during 15 days. 12 ppm Zn2+, 4 ppm Zn2+.
74
-------
Under the equal dose of Zn2* condition, cough frequency of fish in hard water had fewer
changes than that of fish in soft water (Figure 3). This proves strongly that hard water may raise
the resisting ability of freshwater fish to Zn2+ toxicity.
(2) Effects of EDTA
The respiratory frequency of fish in the 6th and the 12th groups began to increase the next
day after EDTA was added to the water (Figure 4). The concentration of EDTA was five times
as large as that of Zn2+. It was observed that the rising tendency of the curve for the fish in hard
water was more stable, and that the frequency was close to normal value (Figure 4a). Also, the
respiratory frequency for the fish in the two groups containing EDTA was much higher than that
for the fish living in water containing zinc but no EDTA, and was even higher than that in 8 ppm
Zn2+. The curve of respiratory frequency in these two groups came closer to the normal. This
implied that EDTA might remove or relieve Zn2* toxicity. Changes of cough frequency in groups
6 and 12 are shown in Figure 4b. The cough frequency of those living in water containing EDTA
was apparently lower than that in the group containing only the same dose of zinc ions. This
agrees with the result mentioned above. It is enough to prove that EDTA is an effective
detoxicant to remove Zn2+toxicity in fish.
12ppm+Ca*++EDTA '
5 10
Exposure time (day)
15
Figure 4. The detoxicated effects of EDTA. (a) Changes of respiratory frequency;
(b) Changes of cough frequency in 12 ppm groups.
75
-------
DISCUSSION
Respiratory movement is a synthetic index offish metabolism. When the water environ-
mental factors change, the receptors, which are widely distributed along the pathway traversed by
respiratory medium, accept this stimulation (Rehwoldt et al. 1971), and thus respiratory frequency
changes swiftly so that the amount of Zn2+ absorbed by fish body is lessened. With the accumula-
tion of Zn2+ on the surface of gill tissues, Zn2+ and mucus secreted by epithelial cells constitute
an undissolved protein compound depositing on the gill surface and stimulating J-receptor in the
gill (Satchell 1976). It reflexively causes the continuous cough response to clear away the foreign
body on the gill. This is a protection measure for fish to survive when they are in an unfavorable
environment. Since the respiratory and cough frequency change rapidly with the concentration of
Zn2+ and the extent of fish poisoning, we regard them as an index or physiological monitor of
poisoning and detoxifying of fish.
The same dose of Zn2+ has less influence upon the respiratory movement of fish in hard
water than in soft water. It shows that the toxicity of Zn2+ in hard water to fish is not serious. It
may be related to the acclimation of fish in hard water for 1 or 2 days before it is polluted.
Because the fish are placed in saturated solution of CaCO3, Ca2+ concentration outside the fish
body is far higher than that in tissues and cells of fish. The fish body is permeated with Ca2+
because of the ion concentration gradient, so the tissues and cells of fish contain a lot of Ca2+ in
advance. The existence of Ca2+ reduces the permeability of surface membrane to Zn2* or resists
the entrance of Zn2+ (Skidmare 1964). Because of the antagonism, which takes place mainly on
surface membrane, the speed of Zn2+ entering the fish slows down, and the chance of the
poisoning of fish would decrease. Wu et al. (1986) pointed out that the reciprocal effect of many
kinds of metals in compounds depends on bioaffinity (f) of metal ions, and the ion with larger
value/can replace those with smaller value. Because/is 14 for Ca2+ and 5.7 for Zn2*, Ca2* can
replace Zn2+ and make Zn2+ accumulation in fish decrease with the increase of activity of Ca2+ in
medium; the toxicity of Zn2+ to fish is decreased. Moreover, zinc sulfate was used in the
experiment, and the production of ZnCO3 sediment was not eliminated. The authors noticed that
the sediments at the bottoms of tank in hard water groups were generally much more than that in
soft water groups. This also causes the reduction of concentration of Zn2* in hard water to some
extent. According to this, it is supposed that the amount of Zn2+ in the body of fish in hard water
groups may be less.
The dithizone spectrophotometry was used to analyze the amount of Zn2+ in the gill, muscle
and liver of fish in each experimental group. It was detected that the amount of Zn2* in the
samples in hard water is obviously smaller than that in soft water of the same dose (Chai et al.,
Unpublished manuscript, 1989). It is demonstrated that hard water has antagonism to Zn2*,
which agrees with the conclusion of this paper. According to the suggestion proposed by
Skidmare (1964) that Ca2+ could antagonize the toxicity of most heavy metal ions, and to the
result of this paper, the authors consider that hard water may be of great advantage for prevent-
ing, curing and lightening the harmful effects on freshwater fish by heavy metal ions. Hard water
may be an ideal and cheap detoxifying agent.
As an effective complexing agent, EDTA is capable of complexing with almost all heavy metal
ions. The reaction is rapid in forming steady complex compounds. Because of the steady
complex compound combined by EDTA and Zn2+, lethal Zn2+ has become nonpoisonous metal
76
-------
chelate; therefore, the toxicity of Zn2+ to fish is relieved. "The fish resume their normal respira-
tion gradually. It is probably due to the detoxification effect of both EDTA and Ca2*; the
detoxifying effect of EDTA is more obvious under the condition of hard water. It proves that
EDTA is an effective detoxifying agent. As for whether EDTA is also complexed with Ca2+ at
the same time and weakens its detoxification effect, the authors think that because the steady
coefficient log K for the complex by Zn2+ and EDTA is larger than that of Ca2+, the experimental
pH conditions change from 5.7-6.5, and the conditional formation constant of Zn2+ is greatly
larger than that of Ca2+ (Laitineu and Harris 1975), EDTA could more easily be combined with
Zn2+ to form metal chelate to produce the effect of detoxification, but is not impossible to form a
small amount of Ca-EDTA This may be why the effect of detoxification of EDTA in hard water
is not so effective as expected. .
ACKNOWLEDGMENT
Funding support for this study was provided by Fujin Natural Scientific Foundation.
REFERENCES
Chai Minjuan, Huang Yuling, and Jiang Xinpo. 1990. Effects of sublethal concentration of
copper ions on respiratory physiology of Tilapia sp. J Fisheries of China, 14(l):50-54.
Crespo, S., E. Soriano, and C. Sampera. 1981. Zinc and copper distribution in excretory organs
of the dogfish Scyliorhinus canicula and chloride cell response following treatment with zinc
sulphate. Marine Biol. 65:117-123.
Huang Yuling. 1987. The Legendre orthogonal polynomial and the studies on fish physiology. In:
The symposium of bio-math. The Chinese Institute of Biomath. Press.
Huang Yiming, Mo Linyun, and Ma Tichun. 1988. Effects of heavy metal ions on cough
response of mud carp. Acta Scientiae Circumstautiae. 8(2):218-222.
Laitineu, H. A, and W.F. Harris. 1975. Pages 191-197 In: Chemical analysis. Complexation
equilibria: theory and applications. McGraw-Hill Press.
Lloyd, R. 1960. The toxicity of zinc sulphate to rainbow trout. Ann. Appl. Biol. 48:84-94.
Rehwoldt, R., B. Gerald, and B. Nerrie. 1971. Acute toxicity of copper, nickel and zinc ions to
some Hudson River fish species. Bull. Envir. Contain. Toxicol. 5:445-448.
Satchell, G.H. 1976. Type J receptors in the gills of fish. Pages 131-142. In: Studies in neuro-
physiology. R. Porler, (Ed.) Cambridge University Press, Cambridge.
Satchell, G.H. 1982. Respiratory toxicology of fishes. Pages 19-23. In: Aquatic Toxicology. LJ.
Weber (Ed.), Raven Press, New York.
77
-------
Sellers, CN., A.G. Heath, and M.L. Bass. 1975. The effect of sublethal concentration of copper
and zinc on ventilatory activity, blood oxygen and pH in rainbow trout (Salmo gairdneri).
Water Res. 9:401-408.
Skidmare, J.F. 1964. Toxicity of zinc compounds to aquatic animals, with special reference to
fish. The Quarterly Review of Biol. 39(3):227-245.
Skidmare, J.F. 1970. Respiratory and osmoregulation in rainbow trout with gills damaged by zinc
sulfate. J. Exp. Biol. 52:481-494.
Wu Yuduan, Chen Cimei, and Wang Lonhfa. 1986. Relationship between heavy metal pollution
and water productivity in Xiamen estuarine harbor. Oceanologia et limnologia Sinica.
17(3):173-186.
78
-------
BIOACCUMULATION OF TOXIC HYDROPHOBIC ORGANIC
COMPOUNDS AT THE PRIMARY TROPHIC LEVEL
by
Deborah L. Swackhamer1
INTRODUCTION
The fate of hydrophobic organic compounds (HOCs) in the water column of aquatic
systems is largely controlled by their association with particulate matter. Particulate matter is
very heterogenous in size and character, and results from both allochthonous and autochthonous
sources. A large portion of this particulate matter can be biogenic in nature, and dominated by
primary production. These lipid-rich algae are thus an optimal matrix for HOCs to sorb to. The
fate of HOCs associated with phytoplankton is controlled by their uptake of HOCs and by their
fate pathways, which includes sedimentation and grazing by zooplankton. A substantial' portion
of the carbon fixed in primary production is transferred to higher trophic levels, such that
phytoplankton serves as a major entry point of HOCs into food webs where they accumulate in
higher order consumers.
The bioaccumulation of HOCs in commercial and sport fish has resulted in the issuing of
health consumption advisories and the closing of many commercial fisheries in the Great Lakes,
the world's largest freshwater resource. Our ability to understand and model HOC
concentrations in fish is dependent on our ability to model HOC concentrations at the base of
the food web, phytoplankton. The currently held theory, or "conventional wisdom", is that HOGs
rapidly reach an equilibrium partitioning between the lipids of th^e organism and the water in
which they reside. Thus the lipid-normalized bioaccumulation factor (BAF) should be equal to
the compound's octanol-water partition coefficient (K^), as they both describe the equilibrium
distribution between water and a pure organic phase.
The chemicals most affected by these processes are those with log Kow> 4. One class of
chemicals of acute interest in the Great Lakes and elsewhere is the polychlorinated biphenyls
(PCBs). There are approximately 100-120 PCB components, or congeners, that can be found in
the environment. In addition to being a dominant contaminant, they have a wide range of
physical-chemical properties, with log K^s ranging from 4-8. Thus they provide a homologous
series of compounds that are a useful surrogate for HOCs.
Our objectives were to test the prevailing hypothesis, which states that the uptake~of
HOCs by phytoplankton is controlled by the compound's lipophilicity (K^). The approach taken
was to determine the factors that controlled the uptake of PCBs by phytoplankton under
controlled laboratory conditions, and to develop a model that would describe bioaccumulation of
PCBs in phytoplankton.
^Environmental and Occupational Health, School of Public Health, University of Minnesota,
Minneapolis, Minnesota, USA
79
-------
METHODS
Experiments were conducted by exposing single phytoplankton species in batch cultures to
a mixture of 40 PCB congeners. This suite of compounds was chosen to represent HOCs in
general and spanned a wide range of physical-chemical properties. Initial phytoplankton mass
was 10 mg/L. A spike of PCB was added by acetone carrier resulting in a total concentration of
1.6 ng/mL. Cultures were maintained at a specified temperature with a lighting regime of 16
hours light and 8 hours dark. Aliquots were taken at specified time intervals and separated into
algal and media phases by centrifugation. Determinations of lipid, mass, and PCB concentrations
were made of each aliquot
PCBs were batch-extracted from the media with hexane and isolated by Soxhlet extraction
of the biomass with 1:1 hexane-acetone. Surrogate standards were added prior to extraction to
monitor extraction efficiencies in all samples and included congeners #14, 65, and 166 (3,5-di-,
2,3,5,6-tetra-, and 2,3,4,4',5,6-hexachlorobiphenyl, respectively). All extracts were cleaned by
alumina and silica gel column chromatography. Columns contained 2-3 g anhydrous sodium
sulfate over 7 g of 10% deactivated neutral alumina (80-200 mesh, activated at 450 °C for 4 hr)
over 1 g 6% deactivated silica gel (100-200 mesh, activated at 300 °C for 4 hr) and were eluted
with 3 X 15 mL hexane. Cleaned extracts were reduced in volume, received internal standards,
and analyzed by capillary column gas chromatography (GC) with electron capture detection. The
internal standards were congener #30 (2,4,6-trichlorobiphenyl) and #204 (2,2',3,4,4',5,6,6'-
octachlorobiphenyl). A Hewlett-Packard 5890 GC with 30 m x 0.25 mm capillary column, ^Ni
electron capture detector, autoinjector, and Waters-Millipore Maxima data system was used with
the following conditions: injection port, 225 °C; carrier gas, H2, 2 mlVmin; oven programmed
from 100 to 280 °C at 1 deg/min; makeup gas, 95%/5% argon-in-methane, 20 mL/min; detector,
350 °C All quantitation was done by the internal standard method.
Bioaccumulation factors (BAFs) were calculated at every time point as the ratio of the
PCB concentration in phytoplankton biomass (ng/kg dry wt) to the PCB concentration in the
media (ng/L), expressed in equivalent units (L/kg).
RESULTS AND DISCUSSION
Most earlier reports have concluded that equilibrium between HOCs and phytoplankton
was achieved within hours (Hansen 1979, Baughman and Paris 1968, Sodergren 1968, Reinert
1979, Rice and Sikka 1973, Geyer et al. 1984, Mailhot 1987). There are no published uptake
experiments lasting more than three days. However, Neudorf and Khan (1975) and Wang et al.
(1982) noted that their BAFs were still increasing slowly with time at the end of their
experiments (3 and 28 hours, respectively), indicating a lack of equilibrium.
Our experiments were carried out for up to 20 days under conditions where algal growth
was minimized by temperature (0.03 day1) and under active growth conditions (0.13 day1)
(Swackhamer and Skoglund 1991). Under low growth conditions, the BAFs for congeners with
log K™ < 7 were related to log K<,w (Figure 1). However, most congeners had not reached
equilibrium even after 20 days of exposure, with the BAF values increasing continually over the
timeframe of the experiment. The rate of uptake was rapid initially, with 40-90% of the uptake
occurring in the first 24 hours. Uptake then appeared to level off, then gradually began to show
a small but steady increase that continued throughout the experiment. This is highly consistent
80
-------
5.0
6.0
Log Kow
7.0
8.0
Figure 1. Relationship of log BAF vs. log E^, under low growth conditions
after 20 days exposure to 40 PCB congeners. For congeners with
log KO,, < 7, r2 = 0.84 and slope = 0.78. Reference line is slope 1.
with the two-step mechanism for bioaccumulation in phytoplankton. It is our belief that previous
investigators misinterpreted the decrease in the initial uptake rate of their experiments as
equilibrium, and terminated the exposures before observing the slow but continued uptake. Our
observed BAFs at 24 hours are similar to those reported in the literature, whereas the BAFs
after 30 days exposure are considerably greater. It is concluded that the overall uptake kinetics
are considerably slower than previously believed. For congeners with log K^ > 7, there was no
direct relationship of BAF to KOW. Since their BAFs continually increased over the course of the
experiment, it is possible that they might eventually reach their K^-defined equilibrium value. It
is also possible that their size or shape caused a significant resistance to transfer into the cell
matrix.
When cultures were maintained under typical growth conditions, different results were
obtained. For most congeners, the log BAFs were independent of the log K^s (Figure 2). BAFs
were lower for active growth conditions compared to low growth conditions. These experiments
indicate that growth is a major factor controlling the uptake of the HOCs by phytoplankton, and
may be more important than K^ under non-equilibrium conditions. For growth conditions
typical of the Great Lakes (0.06 - 0.60 day1, Fahnensteil and Scavia 1987), it is likely that
equilibrium would not be reached.
A decrease in uptake of HOCs may be due to a continual dilution of the biomass due to
growth, or an increase in the elimination of HOCs that is related to an increase in excretion of
extracellular products resulting from increased cellular metabolism. This elimination pathway
may be enhanced as a function of HOC exposure (Gotham and Rhee 1982). It has been
81
-------
5.0
6.0
Log Kow
7.0
8.0
Figure 2. Relationship of log BAF vs. log K^, under normal growth conditions
after 30 days exposure to 40 PCB congeners. Data are average of
all time points, error bars are one standard deviation. For congeners
with log K,,,, > 5, slope is not statistically different from 0.
reported that excretion of glycolates by Selenastrum was increased concurrent with an increase in
photosynthesis, with no net change in growth at certain concentrations of DDT (Rhee 1988).
However, the importance of this elimination depends on the magnitude of the association
constants of the HOC and phytoplankton exudates, which are currently unknown.
Desorption of HOCs from phytoplankton has not been studied as much as uptake.
Despite differences in experimental design, phytoplankton species, and experimental compounds
under consideration, all reports agree that there is a hysteresis effect, with the desorption kinetics
being significantly slower than the sorption kinetics (Sodergren 1968, Harding and Phillips 1978,
Lederman and Rhee 1982). The hysteresis may be explained in part by the lack of equilibrium
conditions in their experiments.
The importance of kinetics in algal uptake of HOCs that is indicated in our laboratory
experiments is also borne out in .field observations (Swackhamer 1985). Figure 3 shows
measured BAFs for different size fractions of particulate matter in the water column of Emrick
Lake, Wisconsin, as a function of K^ for selected PCB congeners. These data were obtained
during summer when growth would be expected to be high. The 10-50 /i size fraction, consisting
of phytoplankton biomass as confirmed by microscopic analysis, shows a weak BAF-K^
relationship. The other size fractions, consisting of zooplankton and detritus, show a strong BAF-
K^, relationship. These data are interpreted to indicate that PCBs have not reached equilibrium
in phytoplankton from these lakes during this time period.
82
-------
o>
o
7.0-
6.0 +
5.0-
4.0
• Total particulate
° 10 - 50
4.0
5.0
6.0 7.0
L°g KOW
8.0
Figure 3. Relationship of log BAF vs. log K^ for total water column
particulate matter (slope = 0.26, r2 = 0.66) and the phytoplankton
size fraction (10-50 p; slope = 0.08, r2 = 0.04) in Emrick Lake,
Wisconsin.
The algal composition will be a determining factor in HOC bioaccumulation and fate In
the Great Lakes, the biomass is thought to be dominated by nannoplankton (< 20/t, Fahnensteil
et al. 1986). Their high rates of production and high levels of excretion may limit HOC
concentrations during productive periods. On the other hand, their large surface area-to-volume
ratio may promote uptake. The role of nannoplankton in HOC fate is largely unexplored and
requires future attention.
HOCs associated with nannoplankton may be more available to the foodchain because
zooplankton grazing favors the smaller forms (Porter 1973). They also have slower settling
velocities than larger phytoplankton, leading to longer residence times in the water column The
rates of phytoplankton production relative to zooplankton grazing can also determine the
importance of foodchain transfer compared to sedimentation. High grazing rates in areas of low
primary production such as open waters may favor HOC transfer to the foodchain; in areas of
high productivity with moderate grazing pressure, transfer to sediments may dominate HOC fate
over foodchain uptake.
These results show that the conventional assumption that phytoplankton HOC
concentrations are in equilibrium and can be estimated from K^ is incorrect. Accurate
estimations of HOC concentrations in the primary trophic level require a kinetics modeling
approach. *
83
-------
SUMMARY
Phytoplankton play a critical role in controlling the fate of HOG in the water column
because they are high in lipids and because they serve as the base of the food chain and as
vectors for transport to the bottom sediments. The relative importance of these pathways in a
given system is dependent on species composition and characteristics of the foodchain. Previous
studies that demonstrate a relationship of BAF to K^ and to phytbplankton surface properties,
as well as the data presented here, support the hypothesis that the mechanism of HOC uptake is
a rapid surface sorption followed by a slower transfer into lipids in the cell matrix. Our work on
the kinetics of uptake indicates that equilibrium is reached slowly and that the rate of uptake is
of similar magnitude as phytoplankton growth under normal field conditions. Thus a critical
factor that controls the bioaccumulation of HOCs is the growth rate of the phytoplankton itself.
It is unlikely that HOCs reach equilibrium in phytoplankton during active growth periods,
therefore indicating that models used to predict.HOC concentrations in phytoplankton must use
a dynamic approach.
REFERENCES
Baughman, G.L., and D.F. Paris. 1981. Microbial bioconcentration of organic pollutants from
aquatic systems - a critical review. CRC Grit. Rev. MicrobioL 1:205-228.
Fahnenstiel, G.L., and D. Scavia. 1987. Dynamics of Lake Michigan phytoplankton: primary
production and growth. Can. J. Fish. Aquat. Sci. 44:499-508.
Fahnenstiel, G.L., L. Sicko-Goad, D. Scavia, and E.F. Stoermer. 1986. Importance of
picoplankton in Lake Superior. Can.'J. Fish. Aquat. Sci. 43:235-240.
Geyer, H., G. Politzki, and D. Freitag. 1984. Prediction of ecotoxicological behavior of chemicals:
relationship between n-octanol/water partition coefficient and bioaccumulation of organic
chemicals by alga Chlorella. Chemosphere 13:269-284.
Gotham, IJ., and G.Y. Rhee. 1982. Effect of hexachlorobiphenyl and pentachlorophenol on
growth and synthesis of phytoplankton. J. Great Lakes Res. 8:328-335.
Hansen, P.D. 1979. Experiments on the accumulation of lindane by the primary producers
Chlorella sp. and Chlorella pyrenoidosa. Arch. Environ. Contam. Toxicol. 8:721-731.
Harding, L.W., and J.R. Phillips. 1978. Polychlorinated biphenyls: transfer from microparticulates
to marine phytoplankton and the effects on photosynthesis. Science 202:1189-1191.
Lederman, T.C., and G.Y. Rhee. 1982. Bioconcentration of hexachlorobiphenyl in Great Lakes
planktonic algae. Can J. Fish. Aquat. "Sci. 39:380-387.
Mailhot, H. 1987. Prediction of algal bioaccumulation and uptake rate of nine organic compounds
by ten physicochemical properties. Environ. Sci. Technol. 21:1009-1013.
84
-------
Neudorf, S., and M.A.Q. Khan. 1975. Pickup and metabolism of DDT, dieldrin and photodieldrin
by a fresh water alga (AnMstrodesmus amattoides) and a microcrustacean (Daphnia pulex).
Bull. Environ. Contam. Toxicol. 13:443-450.
Porter, K.C. 1973. Selective grazing and differential digestion of algae by zooplankton. Nature
244:179-180.
Reinert, R.E. 1972. Accumulation of dieldrin in an alga (Scenedesmus obliquus), Daphnia magna,
and the guppy (Poecttia reticulata). J. Fish. Res. Bd. Can. 29:1413-1418.
Rhee, G.Y. 1988. Persistent toxic substances and phytoplankton in the Great Lakes. In: Toxic
contaminants and ecosystem health: A Great Lakes focus. M.S. Evans (Ed.). John Wiley
and Sons, New York, NY.
Rice, C.P., and H.C. Sikka. 1973. Uptake and metabolism of DDT by six species of marine algae.
J. Agr. Food Chem. 21:148-152.
Sodergren, A. 1968. Uptake and accumulation of 14-C-DDT by Chlorella sp. (chlorophyceae).
Oikos 19:126-131.
Swackhamer, D.L. 1985. The role of water particle partitioning and sedimentation in controlling
the fate and transport of PCBs in lakes. Ph.D. Thesis, University of Wisconsin, Madison,
WI.
Swackhamer, D.L., and R.S. Skoglund. 1991. The role of phytoplankton in the partitioning of
hydrophobic organic contaminants in water. In: Organic substances and sediments in
water. R.A. Baker (Ed.), Lewis Publishers, CRC Press, Boca Raton, FL.
85
-------
-------
EFFECTS OF A NEW INSECTICIDE, CCU, ON CARP:
ACUTE TOXICITY AND METABOLISM IN VITRO
by
Xu, Li-Hong1, Zhang, Yong-Yuan1, Xu, Ying1, and Chen, Zhu-An1
ABSTRACT
The reaction system with 0.1 M Gly-NaOH (pH = 9 ± 0.01) was chosen and homogenate
of fish liver was used as enzyme. The metabolites were extracted with ethyl acetate, separated
by TLC, and identified by HPLC. When the reaction mixture containing 0.8 mg/L CCU was
incubated for 5 hr at 25 °C, CPU could be detected, which was the product of CCU cleaved
between C-l and N-l bond. The specific inhibitor of esterase profenofos could decrease the
production of CPU significantly. It indicated that esterase played a very important role in CCU
degradation in this experimental condition. The pathway of CCU primary degradation in fish
was the same as that of DFB. Acute toxicity tests showed that CCU, CPU, CBA, CBD, CA and
DFB were not toxic to grass carp fry.
INTRODUCTION
Chlorobenzuron, or l-(4-chlorophenyl)-3-(2-chlorobenzoyl) urea (CCU) is a new
insecticide developed by China. The high degree of toxicity exhibited by CCU toward many
destructive insects indicates that the compound may be extensively used for insect control.
Consequently, it is especially important to research its toxicity and behavior in the environment.
CCU is a structural analog of diflubenzuron (DFB), which has been used for many years
in many countries. The fate of DFB in water and soil, its toxicity to nontarget organisms, and
the mechanisms of detoxification in insects have been interesting topics of research for scientists.
The conclusion is that DFB was highly selective and highly toxic to pests and had low toxicity to
mammals (Ivie 1978). So far the results obtained from research showed that properties of CCU
were the same as that of DFB, which was expected for an insecticide.
CCU can be degraded in aerobic water to produce CPU and CBA (Xu Ying et al. 1990).
Although the solubility of CCU in water was very low, the high accumulation in fish tissue could
not be obtained (Hass-Jobolius et al. 1990). It seems clear that there might be an enzyme in
fish which could metabolize CCU. Such a situation had been found with DFB. Apparently it is
significant to demonstrate this for evaluating the safety of CCU in water. The present study was
preliminary research in an effort to provide information on this project. In vitro metabolisms of
CCU are studied using carp liver as materials. By using TLC and HPLC, metabolites are
separated and identified. The enzyme that plays an important role in CCU degradation is
determined by using, a specific inhibitor. The acute toxicity of CCU and its metabolites and
DFB to grass carp fry is also determined.
Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, Hubei Province, PRC.
87
-------
MATERIALS AND METHODS
CHEMICALS
The CCU standard (> 95%) was provided by Eco-Environment Sciences, Chinese
Academy of Sciences. Authentic standards of DFB (> 99%, Fluka,'Switzerland),
4-chIorophenylurea (CPU, > 99%, Merck, BRD), ortho-chlorobenzoic acid (CBA, > 99%,
Merck, BRD), 2-chIorobenzamide (CBD, > 99%, Merck, BRD) and 4-chloroaniline (CA,
> 99%, Merck, BRD), inhibitor profenofos (> 90.6%, Ciba Geigy, Switzerland), and TLC plate,
silica gel 60 F254 (Merck, BRD) were gifts from The Institute of Ecological Chemistry, GSF,
BRD. Superpure water was prepared by MilliQ water purification system (Millipore, USA).
Other chemicals were analytic reagent or chromatography reagent.
TOXICITYTEST
Grass carp fry, age < 48 hr, were obtained from breed pond of the Institute. Tap water
filtered by active carbon was used as dilute water, which was fully aerated, pH = 8, D.O.
> 8 mg/L. Stock solutions of tested chemicals were dissolved in alcohol. The same amount of
alcohol was applied to the control. There was no discernible effect to fish. Acute toxicity test
(96 hr) was adopted. Test solutions were changed daily.
IN VITRO METABOLISMS OF CCU
Grass carp (different size) were caught from breed pond of the Institute. Alive and
healthy common carp were bought from market. The fishes were killed and the livers were
removed quickly, weighed, minced with scissors and homogenized with cold water (1:8, W/V) in
an ice bath. The homogenate was centrifuged at 4000 rpm for 5 min, and the supernatant was
kept in refrigerator for the test.
Incubation mixture was 32 ml containing 16 ml 0.1 M Gly->NaOH buffer (pH =
9 ± 0.01), 8 ml water (or inhibitor) and 8 ml supernatant. The reaction was started by adding
25.6 jig CCU (in alcohol), the final concentration of which was 0.8 mg/L, and incubated for 5 hr
(25°C). Boiled supernatant was used as control.
Profenofos stock solution was 3400 mg/L in alcohol. Working solution was diluted with
50 ml water containing one drop (about 0.03 ml) Triton X-100. Preincubation mixture
contained 8 ml profenofos working solution and 8 ml of the supernatant, and the final
concentration of inhibitor was 8.5 mg/L. Buffer and CCU were added after 40 mih.. The-
amount of alcohol applied to the mixture had no appreciable effect on enzyme activity.
Metabolite in the incubation mixture was extracted two times with 64 ml ethyl acetate.
The combined extract was filtered by anhydrous sodium sulfate funnel, then evaporated to
dryness by rotary evaporator at 35°C water bath; the residue was finally dissolved in 0.5 ml
dichloromethane for TLC.
The sample dissolved in dichloromethane was applied to TLC glass sheets of silica gel 60
F254 and developed in benzene: tetrahydrofuran:glaeial acetic acid (80:20:1). A standard of the
metabolite was chromatographed on the same TLC plate. When chromatography was finished,
8&
-------
the spot of sample which had the same Rf with CPU was scraped and dissolved in methanoL,
treated with ultrasonic, filtered, dried with pure nitrogen gas, and then redissolved in 0.25 ml
methanohwater (30:70) for HPLC analysis.
RESULTS
IDENTIFICATION OF METABOLITE
The TLC spots of the enzyme test and control were identified by HPLC. The spots were
confirmed as CPU by comparing retention time (RT) of standard and that of spots (Figure 1).
Furthermore, Waters 490 detector was adopted, by which absolute spectral evidence for
metabolite identification could be obtained. The scannings of CPU standard and metabolite
spots of TLC in the enzyme test were carried out in the same wavelength range. The
absorbance spectrum of both were coincidental (Figure 2). The reliability of metabolite being
CPU was approved by spectroscopy.
ABS
STANDARD
SAMPLE
24 6 8 10 12 14 MIN
Figure 1. HPLC chromatograph of
CPU standard and
metabolite of CCU.
column: NOVA PAK CIS 3.9 x 150 mm
mobile phase: methanol/water (30/70)
flow rate: 1.0 mL/min AUFS: 0.02
temperature: 30*C standard: 30 ng
sample: 14.5 ng
STANDARD
SAMPLE
190 230
Figure 2. UV absorbance spectrum of
CPU standard and metabolite
of CCU.
abs. AUGS: 0.05
wavelength step: 1 nm
standard: 80 ng
scan AUGS: 0.10
step period: 0.5 s
sample: 42.5 ng
89
-------
IN VITRO METABOLISM OF CCU BY LIVERS OF GRASS CARP AND COMMON CARP
CCU was not stable under alkaline condition. The chemical degradation of CCU could
not be neglected because pH 9 was chosen in the enzyme test. Boiled supernatant was used as
control, in which only chemical degradation was present. Table 1 lists the ratios of CPU
produced in enzyme test (ETcpu) to that done in control (Ccpu) in several tests. It was found
that ETcpu was much higher than Ccpu. It was indicated in some way that enzyme in fish liver
did hydrolyze CCU in vitro.
Table 1. The ratios of ETcpu to Ccpu.
No.
ETcpu
Ccpu
1
4.00
2
6.94
3
3.93
4
5.23
5
5.87
6
4.22
The difference between ETcpu and Ccpu was the amount of CPU produced by enzyme.
Table 2 presents the specific activity of enzyme calculated from the difference between ETcpu
and Ccpu, grams of liver used in the test and reaction time.
Table 2. The specific activity of enzyme in fish liver.
Fish
Grass carp
Grass carp
Grass carp
Grass carp
Common carp
Weight
(g)
750
750
200-300
2000
250
Specific Activity
(CPU ng/g liver/hr)
12.05
15.96
19.90
17.18
27.42
Either grass carp or common carp could hydrolyze CCU. Although there was a
difference in specific activities of different fish, which might be created by the variance of
individual or homogenations in different tests, it showed the presence of enzymatic hydrolysis.
INHIBITION OF ENZYMATIC DEGRADATION OF CCU BY INHIBITOR
Profenofos was a specific inhibitor of esterase. The amounts of CPU in the enzyme test,
control and inhibitor tests (TTcpu) in which 8.5 mg/L profenofos was present, are shown in
Table 3. The value of (I-C)/(E-C) was 11%; it meant that in the concentration of 8.5 mg/L
profenofos the production of CPU was reduced almost 90%. Apparently esterase plays an
important role in the production of CPU in this test condition.
90
-------
Table 3. Inhibition of profenofos to the production of CPU.
CPU(ng)
C
.21.88
I
29.63
E
92.31
I-C/E-C
11%
ACUTE TOXICITY OF DFB, CCU, AND ITS METABOLITES TO GRASS CARP FRY
Table 4 and Table 5 present the results of 96-hr acute toxicity tests. Both DFB and CCU
as well as its metabolites had no acute toxicity to grass carp fry.
Table 4. Survival (%) of grass carp fry exposed to DFB and CCU.
Concentration
fag/L)
CCU
DFB
0
100
100
25
100
100
50
100
100
100
100
100
200
100
100
400
95
90
Table 5. Survival (%) of grass carp fry exposed to CPU, CBA, CBD and CA.
Concentratio
n
fcg/L
CPU
CBA
CBD
CA
0
100
100
100
100
2.5
95
95
95
100
25
95
95
100
100
250
90
95
100
100
2500
95
100
100
100
DISCUSSION
Generally benzoylphenylurea is a kind of selective control agent that acts on various
orders of insects with minimum detrimental effects on man or on the pests' natural enemies.
A chemical possessing such properties is very useful in controlling agricultural pests. The
mechanisms of this kind of insecticide had been considered to affect the ecdysis of insects.
Several hypotheses have been proposed, including inhibition of chitin synthetase, increasing
chitinase activity and inhibition of ecdysteroid metabolism through some soluble eri2ymes
(Saleem 1987). CCU was one of this kind of insecticides. The demonstration of its toxicity to
91
-------
organisms, bioconcentration and transfer in organisms could make a comprehensive evaluation
and provide valuable judgment for its wide use.
The chemical structures of CCU and DFB are very similar.
a
CCU
DFB
The variety of evidence obtained from research indicated that primary degradation of
DFB was almost the same in different animals including many insects, fishes, chicken, sheep and
swine; e.g., initiated by cleavage at both the N-l and C-2 bond to give 2,6-difluorobenzamide
and 4-chloroaniline and at the N-l and C-l bond to give 2,6-difluorobenzoic acid and
4-chlorophenylurea (Ivie 1978, Metcalf 1975, Schaefer 1979, Ishaaya 1988, Opdycke 1982).
The pathway of degradation of CCU should be similar to that of DFB. If CCU cleaved
between N-l and C-2 bond and N-l and C-l bond as did DFB, four primary metabolites could
be obtained; e.g., CPU, CBA, CBD and CA.
o
CPU
o
CBA
Xa
CBD
CA
It was found in a laboratory model aerobic aquatic system that CCU could be degraded to
produce CPU and CBA (Xa. Ying et al. 1990). CPU had been detected from the incubation
mixture of fish liver homogenate and CCU in this study. It indicates that such a metabolic
pathway was also present in fish as for DFB.
CCU should be highly accumulated in fish tissue like DDT because it has very low water
solubility (0.21 mg/L), but the results turned out contrary to prejudgment. CCU had a lower
bioconcentration factor (BCF) in fish exposed to CCU for 5 days (Hass-Jobolius 1990). This
characteristic of CCU was identical with that of DFB. Channel catfish exposed to 0.55 mg/L
92
-------
and 0.007 mg/L DFB did not bioaccumulate DFB (Booth 1977). Explanation could be made
from the results of the present paper. Although CCU could be accumulated in fish tissue to a
certain amount, the presence of the enzyme system made it possible that CCU deposited in
tissue was degraded effectively. So CCU could not be concentrated very high in fish body.
Guppy (Gambusia afftnis) was able to degrade DFB efficiently. Even though the fish was on the
top of the food chain in a laboratory model ecosystem, the ecological magnification in guppy
was only about a fortieth of that in mosquito larvae (Metcalf 1975). This is another generality
between DFB and CCU.
Insecticides are metabolized within the insect's system to less toxic or nontoxic polar
products by three major types of enzyme systems; e.g., mixed-function oxidase (MFO),
hydrolase, and glutathione-dependent transferase (GSAT) (Pimprikar 1982). As to DFB, variant
results were obtained from different insects. The investigation of detoxification of DFB in
Egyptian cotton leafworm (Spodoptera tittorolis) larvae indicated that esterase played a key role
in metabolism of DFB (Ishaaya 1988). In housefly (Musca domestica L.), esterase and GSAT
might be contributing to DFB resistance, but MFO played an important role (Pimprikar 1979,
1982). The results were also obtained from other animals. From experiment Opdycke (1982)
found that neither induction nor inhibition of MFO activity altered DFB metabolism. Although
the enzyme system controlling the degradation of DFB might be different, the way that DFB
was cleaved was almost the same. In the present study, the producing of CPU was decreased
when 8.5 mg/L profenofos was applied to the reaction mixture. There is no denying the fact
that in this test condition esterase played a key role in the production of CPU.
When the concentration of CCU and DFB was high to 400 p,g/L, no death happened in
grass carp fry. Metcalf (1975) reported that mouse oral LD50 was in excess of 3000 mg/kg and
LC50 to guppy was > 100 mg/L. The LC50 of DFB to bluegill and channel-catfish was also
> 100 mg/L (Johnson 1980). The results obtained from different researchers are identical.
CPU and CBA have been detected in laboratory model aerobic aquatic system and CPU
done in enzymatic reaction. The toxicity tests were conducted with four metabolites. None of
these was lethal to grass carp fry in the concentration range of 2.5-2500 /ig/L. Therefore,
neither CCU nor primary metabolites presented acute toxicity to grass carp fry.
The cleavage of C-N bond was only primary metabolic pathway. The four metabolites
would be the precursor of the latter metabolism. When bluegill exposed to 10 /ig/L DFB for
24 hr, in which tissue 264 jug/L DFB was measured, were placed into water for 24 hr, it showed
only 8.0 figfL DFB in tissue and no detectable DFB and 0.4 /tg/L CPU in water. Thus, fish
eliminated DFB at a high rate and neither DFB nor CPU was an important excretory product
(Schaefer 1979). It is obvious that further metabolism could happen in CPU or other
metabolites.
From the results above it could be concluded that: (1) grass carp and common carp liver
could metabolize CCU in vitro. The cleavage ways for both CCU and DFB are the same;
(2) the key enzyme which metabolized CCU in the experimental condition adopted in the test is
esterase; (3) there is no acute toxicity of CCU, CPU, CBA, CBD, CA and DFB to'grass carp
fry.
93
-------
ACKNOWLEDGMENT
Technical assistance provided by Miss Li Yang is gratefully acknowledged.
REFERENCES
Booth, G.M., and D. Ferrell. 1977. Degradation of Dimilin by aquatic foodwebs. In: Pesticide
in aquatic environment. M Mohammed Abdul Qudua Khan (ed). Plenum Press, New
York and London, pp. 221-242.
Hass-Jobolius, M., Zhang Yongyuan, Xu Ying, and Chen Zhuan. 1990. Comparison study on
the fate of diflubenzuron and l-(4-chlorophenyl)-3-(2-chlorobenoyl) urea in laboratory
water and fish. Acta Hydrobiology. In press.
Ishaay, L, and D. Degfeele. 1988. Property and toxicological and significance of diflubenzuron
hydrolase activity in Spodoptera litteroralis larvae. Pestic. Biochem. Physiol. 32: 180-187.
Ivie. G.W. 1978. Fate of diflubenzuron in cattle and sheep. J. Agric. Food Chem. 26: 81-89.
Johnson, W.W., and M.T. Finley. 1980. Handbook of acute toxicity of chemicals to fish and
aquatic invertebrates. Washington, D.C. pp. 30-31.
Metcalf, R.L., Po-Yung Lu, and S. Bowlus. 1975. Degradation and environmental fate of 1-
(2,6-diflurobenzoyl)-3-(4-chlorophenyl) urea. J. Agric. Food Chem. 23:359-364.
Opdycke, J.C., R.W. Miller, and R.E. Menzer. 1982. In vivo and liver microsomal metabolism
of diflubenzuron by two breeds of chickens. J. Agric. Food Chem. 30:1227-1233.
Opdycke, J.C., R.W. Miller, and R.E. Menzer. 1982. Metabolism and fate of diflubenzuron in
swine. J. Agric. Food Chem. 30:1223-1227.
Pimprikar, G.D., and G.D. Georghiou. 1979. Mechanism of resistance to diflubenzuron in the
house fly, Musca domestica (L.). Pestic. Biochem. Physiol. 12:10-22.
Pimprikar, G.D., and G.D. Georghiou. 1982. Effect of sesamex on the in vivo metabolism of
diflubenzuron in larvae of susceptible and resistant strains of the housefly, Musca
domestica L. J. Agric. Food Chem. 30:615-628.
Saleem, M.A., and AR. Shaloori. 1987. Joint effects of Dimilin and Ambush on enzyme
activities of Tribolium castaneum larvae. Pestic. Biochem. Physiol. 29:127-137.
Schaefer, C.H., E.F. Dupras, Jr., R.J. Stewart, L.W. Davidson, and A.E. Colwell. 1979. The
accumulation and elimination of diflubenzuron by fish. Bull. Environ Contain Toxicol
21:249-254.
Xu Ying, Xu Lihong, and Zhang Yongyuan. 1990. Preliminary studies on the degradation and
metabolism of chlorobenzuron under aerobic environment. Acta Hydrobiology. In press.
94
-------
RESPIRATORY OXYGEN REQUIREMENTS OF FISHES:
DESCRIPTION OF OXYREF, A DATA FILE BASED ON TEST
RESULTS REPORTED IN THE PUBLISHED LITERATURE
by
Robert V. Thurston1 and Peter C. Gehrke2
INTRODUCTION
We have compiled a file of data from the published literature on respiratory oxygen
requirements of fishes. We call the file OXYREF, and it contains data from over 6,800
individual laboratory tests. For each test, these data include fish species, fish weight, test water
conditions, degree of fish metabolic activity, and measured respiratory oxygen requirement. This
paper describes OXYREF, explains how it was compiled, and reports the results of a simple
analysis of the data in the file.
PROCEDURES
To compile OXYREF, we conducted a computerized literature search of three separate
journal referencing sources for the time period 1969-1986. These are the Biological Sciences
Information Service (BIOSIS PREVIEWS), Aquatic Sciences and Fisheries Abstracts (ASFA),
and Canadian Water Resources References (AQUAREF). Key words used in the search
included that for fishes as a group, and any combination with oxygen, respiration, ventilation,
and gases. We obtained over 5,000 hits, and we reviewed the print-out for each; these printouts
included publication title, key words, and, for the past several years, abstracts. From these
print-outs we decided that 676 publications should be reviewed in their entirety, and we began a
systematic library search for them. We have now located all but 13. From our own library we
reviewed an additional 94 publications. Of all the publications considered, 61 were not in
English and translations were not immediately available to us; these have not been reviewed. In
all, we eventually reviewed 696 publications. Of these, 404 were rejected because they were not
relevant in spite of title or abstract, did not contain primary data, or did not contain sufficient
documentation of test conditions to meet our criteria for inclusion in OXYREF. The remaining
292 publications each contained one or more tests that qualified for the file.
The minimum information needed for a test to qualify for OXYREF was author
identification of the test fish species, number of fish per test, fish weight, test temperature,
indication of fish activity (which we classified according to the conventional metabolic activity
categories: standard, routine, active, and burst), and oxygen consumption per unit of time. If
there was more than one fish in a test, the weight recorded in the file is the mean weight. If
test salinity was not reported, we recorded it as 35 o/oo for a marine species and as 0 o/oo for a
fisheries Bioassay Laboratory, Montana State University, Bozeman, Montana, USA.
*NSW Agriculture & Fisheries, Inland Fisheries Research Station, Narrandera, New South Wales
2700, Australia.
95
-------
freshwater species. We recorded partial pressure of oxygen in the test water (in mm Hg) as
reported, or if not reported it was.possible to make certain assumptions for most tests from the
description of the test methods. When we made partial pressure, assumptions, these included
corrections for test temperature and water salinity. Oxygen consumption was reported in a
variety of ways, but we converted all to mg O2 per kg of fish per hour (mg/kg/h).
Additional information recorded for each test, if available, included fish length, age (in
years), sex (female, male, both, or unknown), and swimming speed, either reported as or
converted to body lengths per second (bl/s). The file also includes the literature reference for
each test, and where appropriate, explanatory comments that sometimes describe the purpose of
the test
Although OXYREF contains information only from publications reporting primary data,
our definition of primary data includes coordinate measurements of data points read from
graphs if that was the only way the data were presented. If data were presented in both
tabulated and pictorial form we chose the tabulation. Frequently, however, an author would
present data only as a plot of oxygen consumption vs. some variable such as temperature, body
weight, or partial pressure of oxygen, along with an equation derived therefrom. If all other
minimum criteria for inclusion of data in the file were met, we would read data from the plot.
If necessary, we would enlarge the plot by photocopy to improve linear measurement precision,
and would then measure and tabulate the X, Y coordinate measurements for each data point in
the plot. We used a microcomputer program to convert the linear X, Y coordinate
measurements into the values reported on the axes of the plot. At the same time that we
inserted the scale of the X, Y axes into the program, we converted the reported oxygen
consumption to our standard units if it had been reported in other units.
Separate from the data file, we have compiled a bibliography of all publications cited.
We also have created a hard copy record for each publication that includes work sheets for all
conversions, and provides any additionally available information on test water pH, alkalinity,
hardness, and turbidity. Records have also been kept of all publications reviewed but rejected,
stating reasons for rejection, the two most common of which were failure to report fish weight
or test water temperature.
SUMMARY OF DATA FILE ENTRIES
OXYREF currently contains 6,840 separate test entries from a total of 292 publications
(mean - 23 tests per publication; range 1 to 355). There are 1,753 entries for fishes tested
under standard metabolic conditions, 3,573 under routine conditions, 1,514 under active
conditions, and none under burst conditions. Each test is identified by a publication ancTtest
number, e.g., 241.012 (bibliography publication number 241, test 12), along with a listing of
authors' surnames and year of publication. The file contains data on fishes from 101 families
and at least 237 species (Appendix); 17 tests are included where authors did not identify fishes
beyond family or genus.
To obtain an estimate of numbers of freshwater and marine fish species in OXYREF, we
chose the following divisions for test water salinity: freshwater, < 5 o/oo; estuarine, 5-25 o/oo;
and marine, > 25 o/oo. Using these divisions, we searched the file and found 4,427 tests on 123
"freshwater11 species, 579 tests on 30 "estuarine* species, and 1,834 tests on 137 "marine" species,
96
-------
totalling 290 species. This number is greater than the actual number of fish species in the file,
indicating that we report tests for some species in more than one of the three salinity divisions.
INITIAL DATA ANALYSIS
The aim of this initial analysis of OXYREF is to generate descriptive equations of
respiratory oxygen requirements of fishes under different metabolic activity conditions.
Eventually, we propose to use equations to construct a definitive, statistical model of oxygen
uptake by fishes over a range of water quality conditions. For the present illustrative purposes,
however, only the effect of body mass was considered. Data were sorted into categories of
standard, routine, and active metabolism as listed in the data file, and each category was then
analyzed separately.
The following allometric equation was fitted to data from each metabolic category:
M = aW> (where M = oxygen uptake, W = body weight, and a and b are constants). Results
are listed in Table 1 and illustrated in Figures 1-3.
Table 1. Summary of regression analyses of oxygen uptake against body weight
Activity Level
Standard
Routine
Active
N
1753
3573
1514
r*
0.918
0.838
0.912
In a
-1.96
-0.754
-0.760
b
0.839
0.646
0.877
tOOOOr
1000
o> 100
o
£
o
o
10
0.1
0.01
0.001
0.01
0.1
10
100
n = 1753
In a = -1.96
b = 0.839
r* = 0.918
1000
10000
WEIGHT (g)
Figure 1. Oxygen consumption rates at standard metabolic test conditions.
97
-------
o
£
10000
1000
10
o o.i
o.oi
O.OO1
n = 3573
In o - -0.754
b - 0.646
r« - 0.838
0.01
0.1
W
100
1000
10000
WEIGHT (g)
Figure 2. Oxygen consumption rates at routine metabolic test conditions.
o
£
10000
noo
wo
o 0.1
0.01
n - 1514
In a = -0.760
b " 0.877
r» = 0.912
0.001'—
0.01
' ••"' 1—i i i mil
'
0.1
10 100
WEIGHT (g)
WOO
WOOD
Figure 3. Oxygen consumption rates at active metabolic test conditions.
The equations for standard, routine, and active oxygen consumption all have high
coefficients of determination (r2), indicating that 84% or more of the variance in oxygen uptake
is directly attributable to changes in fish size. Oxygen uptake in swimming (active metabolism)
fish is clearly greater than in spontaneously active (routine metabolism) or resting (standard
metabolism) fish.
98
-------
These initial results suggest that by selecting equations of a suitable form, and including
other environmental factors from the data file, it should be possible to obtain a definitive model
explaining over 90% of the total variance in oxygen uptake in each metabolic activity category.
Work on OXYREF is continuing, both in expanding the file and in our analysis of the data it
contains. We expect the next version of OXYREF to be available by 1994.
ACKNOWLEDGMENT
John F. Neuman, Fisheries Bioassay Laboratory, organized the data file for this
presentation. This research was funded by the U.S. Environmental Protection Agency
Environmental Research Laboratory, Duluth, Minnesota under Cooperative Agreement
CR811958, and Environmental Research Laboratory, Athens, Georgia under Cooperative
Agreement CR813424.
99
-------
APPENDIX
LISTING BY FAMILY AND SPECIES OF ENTRIES IN OXYREF
FAMILY
Acanthuridae
Acipenseridae
ftlepocephalidae
Aluteridae
Atnbassidae
Anabantidae
Anguillidae
ftnoplogasterida
Anostomdae
Astronesthidae
Bagridae
GENUS and SPECIES
Acantbunusp.
Adpensertransmontemus
Bajacaliforma burragd
Obeslaascripta
Ambasasintemtpta
Anabosscondens
Andbos testudincus
CoSsa fosciota
Trichogaster trichopteru
AnguiOaangidOa
jtngtrilfft nttstrntif
AngiaOa japomca
Angtuttarostrata
Anoptogastercamuta
Leporinus fasciatus
Borostomas panamenas
Mystusarmatus
Myatuscavasau
MyatusguBo
Mystusvittatus
FRESH- ESTU- MARINE S
HATER ARINE
1
10
1
3
262
71
16
48
3
98 41
2
34
10 9
15
3
2
10
17
2
61
iPECIES
TOTAL
1
10
1
3
10
71
16
48
3
139
2
34
19
15
3
2
10
17
2
61
100
-------
Balistidae
Bathylagidae
Blennndae
Bothidae
Callionymidae
Carangidae
Carapidae
Catostomidae
Centrarchidae
Chanidae
Cnann i en t hu i dae
Channidae
Chanmde
Characidae
Batistes caprisaus
Batistes sp.
*^~
BathytagusmSleri
BathyUtgusochotensis
Bathyiaguswesetki
gKflraW*
Blemtius photis
Scophthabnusmtudmus
Ca&onymuslyra
Caranxsexfasdatus
Elegatisbipinrsulata
Naucmtes doctor
Seriolaquinqueradiata
Campus hornet
Catostomus commersom
Catostomustahoensu
Le ponds cyanettus
Leponds gtbbosus
Lepomusmacrochmts
Micro ptems sabnoides
Chanoschanos
Chaenoceohalusacenaus
pruaus
Chama gachua
Channamaru&a
Chama punctatus
Channa striatus
CnOfttUl StfKftUS
Exodon panidoxus
Proch&odusscrofa
6
18
2
243
76
56
52
16
84
2
19
4
2
3
39
i
1
1
1
i
46 59
30
26
1
3
50
1
6
-
3
39
i
1
1
1
1
105
30
26
1
3
50
1
6
44
6
18
2
243
76
56
21
52
16
84
2
19
4
2
101
-------
Cichlidae
Clarudae
Clupeidae
Cobitidae
Coryphaenidae
*
Cottidae
Cypnnidae
Aegtddera pordegrmas
HapJochromudegans
Ofcochfoitos mosxintbicus
Onochramuraloticiu
Pterophythan scalar*
Sarotherodonri&oticus
TUapiarendaK
TUapiaziUu
Claries batrachus
Brevoortiatyramus
Donaomacepediamun
Lcpidocephaha guntea
Coryphaena eqtdsetis
Coryphaena Mppunu
Myoxocephahu octodecemsp
Abranosbrama
Barbusaenau
Barbtuticto
Canuaiusauratus
Carossius carassus
Ctrrhtnus tnrigala
CtenopharyngodonideOa
Cyprinuscarpio
Esomus dannca
HybognathusnuchaUs
Labeocalbaat
Lobeocapenos
Labeo erythrinus
Labeororata
NotendgDnuscryaateucas
NotropuhOrensu
Orthodonmcrolepidotus
Pimephaks pramdas
Pimephalesvigflax
Puntuubarbus
Tinea tinea
28
10
208
4 12 4
31
1
7
24
9 .
3 18
52
20
8
79
1
16
37
28
1
274
24
88
6 2
175
25
2
8
50
1
9
8
2
17
22
4
2
16
7
. 28
10
208
20
31
1
7
24
9
21
52
20
8
79
1
16
37
28
1
274
24
88
8
175
25
2
8
50
1
9
g
2
17
22
4
2
16
7
102
-------
Cyprinodontidae
Oactylopteridae
Dasyatidae
Diodontidae
Echeneidae
Ernbiotocidae
Erythrimdae
Esocidae
Exocoetidae
Gadidae
Galaxudae
Gasterosteidae
Aphamtudiapar
CyprinodonvariegfOus
Funduhis grandis
Ftotdulus sitstSis
DactytoptenuvoBtans
Dtuyatissabina
Diodanhutris
Echmusnaucrates
Rttnofa mnofu
EmbiotocalateraKs
BhacochHusvacca
Hoplerythrmusumtaetna
Hopltasmalabaricus
Eaoxtucau
CheSopogon nigriams
ChsUopogonsp.
CheUopogon suttom
Cypsdunut opathopus
Cypsdums poecUopterus
Cypsebmusp.
Euleptorhamphusviridis
Exocoetusmonocirrhus
Exocoetusvotitmu
Hirundichthys speculiger
Progrdchthysaealei
Gadtumarhua
MetanogramnKu aegJefinus
Thmtgmchalcogramma
Neochanaa btarrowaus
Gostemstauaculeatus
5
4 13
4 18
1 *5 1*3
1 O lj
4 18
14
10
2
8
3
2
1
55
41
18
1
1
1
2
2
1
1
1
1
1
3
127
10
23
1
71
5
17
22
4-7
l/
22
14
10
2
8
3
2
1
55
41
18
1
1
1
2
2
1
1
1
1
1
3
127
10
23
1
01
103
-------
Gobudae
Gonostomatidae
Haercuiidae
Heteropneustida
Hexagrammdae
Holocentridae
Ictaluridae
Kuhliidae
Lepidogalaxi ida
Lethrinidae
Loricanidae
Lutjanidae
Hacrouridae
ttalacosteidae
Hastaceinbelidae
Helanphaidae
CoryphoptenavSehata.
YTJiMt ntnjp rriinntlla i
GJossogobius gfttrus 25 18
Gobiodonhatrio
OKgolepuacu&pattds 1 13
Typhlogobau caKfonuen
Cydothoneacdandens
i Ottuuiasys cotnittersonni
Heteropneustes fossHis 44
Optaodon dongatus 1
Holocentmsdiadema
Myxocephahuoctodecemsp
Ictdhuusmdas 2
Ictahuusnebalosus 90
Ictakims punctatus 31
Kuhliasandvicensis 80
Tfridotyilcndnxisalotnandf 3 1
*
Lethrimtssp.
Andstnudwgnsi 29
Lot janus catnpcchanus
PhomboptitesauTontbens
Coryphaenoides armatus
AistpstonaassdntiOan
Afoffffgnatftus aculfatus 43
Mdamphaesacanthotnus
Pofonutm cnusKtpcs
Scopelogadusndzolepis
104
23
86
1
1
41
9
on
3D
13
1
5
4
2
18
14
3
1
2
1
^•M
23
87
44
1
15
41
9
1
i
30
44
14
1
5
2
90
31
84
31
2
29
18
14
3
1
43
2
1
•^M
-------
Mondae
flugilidae
Mullidae
Myctophidae
Myximdae
Neuuptendae
Neoscopelidae
Nomeidae
Notothenndae
Oneirodidae
Oryziatidae
Melanoma zugmayeri
IhathaneriK
Lizamacmtepis
Luarichmdsotd
Mugftaumtus
Mugflcephahis
MugH corona
Mugftmacmlepu
Mugflsp.
lUunomugflcoraula
MuHoidichthysattriflamm
4 species
Diapfautheta
Lampanyctusregalu
Lampanyctusritteri
MyctophutnoufoUttertuitutn
Myctophummdduhun
Myctophumsptsosum
Porvihtxingens
Stenobrochttuleucopsow
SymbolophoruacoK forme
Symbdophorus evemuutni
Tonetonbeomo cffnulon
Triphotuntsmeocomts
Afyxbie gtotososa
Pentapodusndcrodon
Scopefaigystristis
Nameussp.
Psenes cyanophrya
Notothemarosa
OndrodesGconihias
Oryzuutadpes
1
19 26
7
22
30
19 54
12
36 18
96
18
2
67
2
1
2
1
1
2
1
2
1
2
2
2
2
1
1
18
9
8
1
59
o
1
45
7
22
30
73
12
54
96
18
2
67
2
1
2
1
1
2
1
. 2
1
2
2
2
2
1
1
18
9
8
1
59
105
-------
Osteoglossidae
Oxyporamphidae
Percichthyidae
Percidae
Petronyzontidae
Pleuronectidae
Poecilndae
Polyptendae
Ponacentridae
Poinatonidae
Aropoitna gtgof
Oxypofotnphus tntcroptcnts
Moroneatnericana
Morone saxatilis
Ethcostomaboschungi
Etheoatomaduryi
Etheoatoma flabettare
Etheostoma fusi forme
Etheostoma Tufilincotum
Etheostoma squanuceps
Pcrca flavescens
Parana caprodes
Stizoste&anhtdoperca
Stizostedionvitreutn
Ichthyomyzan tatbiba
Lampetra fhtviatiUs
Lampetra ftaneri
Petromyzon marmus
Hippogtosxndesplatesso
LimandaUmanda
Microstomusldtt
Platichthys flesus
PlatKhthyssteOatus
Pleuronectes platessa
Pscudopleuronectesameri
Gambusa of finis
Piatypofdhis sp.
PoedSalatipama
Poccttia reticulata
XiphophonuheOeri
Calanioichthyscalabanat
PomacatSnuaatjtfaadatu
Pomatomussaftertnx
41
8
50
62 98 19
2
2
2
2
2
2
20
2
8
35
39
65
17
4
5
6
6
7 4
22
4 94
25 43
28
13
13
46
5
1
3
22
41
8
50
179
2
2
2
2
2
2
20
2
8
35
39
65
17
4
5
6
6
11
22
98
68
28
13
13
46
5
1
3
22
106
-------
Protopteridae
Salmomdae
Sciaemdae
Scombri dae
Scorpaenidae
Scyliorhimdae
Searsudae
Serranidae
Serrasalnudae
Span dae
Protoptenuannectcus
Congomtsautumnalis
Cofegonusaatdinetta
Cofegonus schinud
Congonussp.
Oncorhynchtakisutch
Oncorhynchusnerka
Saknogmdneri
Saimoaatar
Sabnotrutta
SalveSmutalpinus
Saivetinus fontinatis
Safoetinusnamaycush
ThymaBusarvticus
Laostomus xanthurus
Euthymau off ads
Euthymuu pdattds
Thurmusalatunga
No. species not known
StbeutestRptopro
Sebastoiobusaltivetts
HemiscyQaan ptagiosum
Scytiorhinuscanicula
ScyHoMnus ste&aris
Saganachthys abei
No. species not known
Centfopfutu stnoto
Epmephekutakaara
Epinephekusp.
Colossoma macro pomum
Lagodonrhcmboides
AfySomacmcephahis
20
3
13 7
28
8
14
88
532 44 19
37
53
2
120
122
16
85
1
.48
6
3
19
6
2 1
33
10
1
4
4
31
4
114
10
23
20
3
20
28
8
14
88
595
37
53
2
120
122
16
85
1
48
6
3
19
6
3
33
10
1
4
4
31
4
114
10
23
107
-------
Squall dae
Stomidae
Synbrandndae
Teraponidae
Torpedimdae
Uranoscopidae
Xiphndae
Zoarcidae
SfonftftM nftmtttifiit
Xaitaln* mvtrtfvt
Stomasatriventer
AmptepnouscwMa 10
Synbranchusmarmoratus 18
TcToponthffops
Torpedo marmorata
Genyagnus monoftaygfus
Xiphias gjltulius
Afetanostigma pammdas
Kfu gopttila detubonu
Zoarcesviviparus
i
19
1
2
3
21
1
8
15
8
4
19
1
10
18
2
3
21
1
8
15
8
101 Families
254 Species
4427
579
1834
6840
108
-------
TOXICANT UPTAKE ACROSS FISH GILLS
by
David J. Randall1 and Colin J. Brauner1
Xenobiotic chemicals are not produced by living organisms but synthesized by abiotic
processes. Most of these compounds are synthesized by the pharmaceutical, petrochemical
pesticide and plastic industries (Connell 1990). These chemicals can be surface active or enter
the body and exert a toxic action within the body. Chemicals enter animals with the food or
directly through the body surface.
In aquatic animals many compounds are transferred from the water into the body across
the body surface. This is achieved either by the compound passing directly through cells or via
paracellular channels between cells. Epithelial cells are often cemented together by tight
junctions reducing paracellular transport to a minimum. Substances can pass through the cell if
they are hpid soluble; if not, either they must be bound to a membrane transport molecule or
pass through a limited number of small water-filled channels in the membrane. In general,
transport proteins in the membrane only bind specific compounds, and molecules must be very
small to pass through membrane channels. The surrounding lipid membrane of epithelial cells
joined together by tight junctions, forms a significant barrier around the animal to lipid-insoluble
compounds. It is not surprising, therefore, that the entry of many xenobiotics into animals
depends on lipid solubility;
If a chemical must be lipid soluble to enter an animal, then most of the chemical, once
mside the animal, will be dissolved in the fat. Many xenobiotic compounds have octanol/water
partition coefficients (K^) between 100 and 1,000,000. An animal may be 10% fat and if log
KOW is 6, the ratio of chemical dissolved in body fat versus body water will be 10s, assuming that
K,,w reflects the distribution between body fat and body water. This is a reasonable assumption
because Dobbs and Williams (1983) showed that fat solubility of xenobiotics varied little with
the type of fat. If the chemical is in equilibrium between the water and the animal, then clearly
the animal will contain much more of the chemical because of body fat content. This is referred
to as bioaccumulation, and in aquatic animals where the concentration of a chemical in the body
is compared with that in the water, it is termed "bioconcentration." In terrestrial animals, where
comparison is made, not between the animal and the environment, but between the
concentration of the chemical in the food and the animal, then the term "biomagnification" is
used (Connell 1990). &
There is a linear relationship between the rate of absorption by fish and the octanol/water
partition coefficient of many chemicals. The higher Kow, the greater the rate of influx of the
compound into the fish (Figure 1; McKim et al. 1985, Saarikoski et al. 1986, McKim and
Erickson 1990). At high values of log K^, above 4-6, the relationship breaks down and no
further increase in rate of influx is seen with increasing K™, (Figure 1), in fact McKim and
Erickson (1990) report a decreased uptake with increasing log K«w above 5.
'Zoology Department, University of British Columbia, Vancouver, B.C., Canada.
109
-------
z
Log K
1
•
O1 234 56
LogKow
Figure 1. The relationship between the rate of absorption (Log k) by guppies
(Poecilia reticulata) from acidic water (pH < pK,), and the n-
octanol/water partition coefficient (K^), for various phenolics and
carboxylic acids (from Saarikoski et al. 1986).
The gills represent the major portion of the body surface area of fish and also present
only a 5 to 10 ji barrier between water and blood (Hughes 1984, Laurent 1984); consequently,
most of the chemical transferred between the fish and the environment occurs across the gills.
Chemicals are delivered to the gill surface by a unidirectional water flow over the gills; they
diffuse across the gills and are then distributed to the tissues by the blood flow. Potentially
water flow, blood flow and/or diffusion across the gills could limit uptake of the chemical by the
fish. Thus, transfer of chemicals across the gills can be discussed in terms of ventilation, blood
perfusion, and/or diffusion limitations.
In general the flows of water and blood are matched to their oxygen contents, that is, flow
times oxygen content are the same for both blood and water. The blood contains an order of
magnitude more oxygen than water, and so water flow is usually an order of magnitude greater
than blood flow in fish. Blood of vertebrates contains about 5% lipid, and as long as log K^ is
above about 2 the carrying capacity of the blood to remove the chemical will exceed the ability
of the water to deliver the chemical to the gills. In addition, many xenobiotics are bound to
plasma proteins (Schmieder and Henry 1988), and any ionizable xenobiotic with a pK below 8~
the pH of the blood—will be retained in the blood because gill membranes are impermeable to
the ionized form. That is, binding to plasma proteins, high fat solubility, and ionization will tend
to keep the concentration of the undissociated form in aqueous solution in the blood low, and
will also enhance the carrying capacity of the blood to transport the compound away from the
gills. Thus, transfer of xenobiotics into the fish will be dependent on water flow and/or rate of
diffusion across the gills and largely independent of blood flow.
The linear increase in absorption with log K^, indicates the changes in lipid and/or water
solubility effects the rate of chemical uptake. The use of K^ can be misleading, however,
because a high K,*, could mean either a high lipid solubility or a very low water solubility.
Dobbs and Williams (1983) showed that there was a linear relationship between water and fat
110
-------
solubility, but a steep inverse relationship between water solubility and log K^ That is, high log
KOW values were associated with very low water solubilities. The ability to deliver a chemical to
the gills will depend on gill ventilation and the water solubility of the chemical, whereas diffusion
across the gill epithelium will depend on the lipid solubility. At low log K,^ water solubility is
high and the ability to deliver the chemical to the gills is high compared with transfer rates
across the lipid membrane, indicating that uptake is diffusion limited, perhaps over the range of
log KOT, 1 to 4 (Figure 1; Saarikoski et al. 1986). At high values of K^ the reverse is true, and
although fat solubility is reduced, compared with compounds having a lower K^, (Dobbs and
Williams 1983), water solubility is even more reduced. This means that the capacity to deliver
the chemical to the gill surface is much more impaired than the gill diffusing capacity, and at
high values of log K^ the uptake process may be limited by gill ventilation. Thus we conclude
that uptake of lipid soluble chemicals tend to be diffusion limited below a log K^ of about 4,
but limited above that by gill ventilation.
Water pH can have a marked effect on the uptake of weak acids because cell membranes
are often permeable to only the undissociated forms of weak acids (Saarikoski et al. 1986).
Thus, if water pH < pK of the weak acid, absorption of the compound will be rapid but uptake
will decrease with increasing pH. The actual pH at which uptake will be reduced will be related
to the pK of the acid in question. Saarikoski et al. (1986) clearly demonstrated this for a
number of carboxylic acids (Figure 2).
100
10
0.1
Ph,
DBNP
5 6
PH
8 9
Phe
DCP
245-TCP
246-TCP
PCP
PheBuA
DCBeA
DBNP
Phenol
2,4-Dichlorophenol
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
Pentachlorophenol
4-Phenylbutyric acid
3,4-Dichlorobenzoic acid
2,6-Dibromo-4-nitrophenol
10.05
7.85
7.07
6.22
4.71
4.76
3.7
3.7
Figure 2. The relationship between rate of absorption k(%) by guppies (PoecUia reticulatd),
expressed as a percentage of that measured at low experimental pH where uptake
is greatest, and the pH of the bulk water. Chemical abbreviations are defined and
their respective pK»'s listed (modified from Saarikoski et al. 1986).
Ill
-------
Thus the uptake of these weak acids is related to the concentration of the undissociated
form rather than the total concentration of the organic compound in question. To calculate the
concentration of the undissociated form in solution it is necessary to know the pH of the
solution in contact with the gills (Lin and Randall 1990; Randall et al. 1990) and the pK of the
compound in question. It may be difficult to determine the exact aqueous concentration of the
undissociated form in water containing particulate matter that binds the compound. This is a
complex problem because binding has been shown to increase with log !£<„, of xenobiotic
compounds (Black and McCarthy 1988).
The gill diffusing capacity is a measure of the ability of the gills to transfer material, and
will increase with gill area and the inverse of gill thickness. Diffusing capacity can also change
within an individual due to increased ventilation:perfusion matching; i.e., more even water flow
over, and blood flow through, the gill epithelium, and a thinning of the epithelium due to
elevated blood pressure. Diffusing capacity may change by a factor of 2 or 3 during exercise or
hypoxia. Water flow over the gills also increases with exercise and hypoxia. Exercise therefore,
can be expected to enhance the rate of uptake of a chemical because of increases in both
diffusing capacity and gill ventilation. We observed this in rainbow trout, where plasma
tetrachlorobenzene (TCB) levels increased with swimming speed (Figure 3) and oxygen uptake
(Figure 4), presumably due to an increase in both water flow over the gills and gill diffusing
capacity.
10
o
o
o
m <=>
8--
6--
o> .,
a 4..
2--
1.0
1.5 2.0
SWIMMING VELOCITY (Bl/s)
2.5
Figure 3. The relationship between plasma TCB (1,2,4,5-tetrachlorobenzene)
concentration and swimming speed in rainbow trout. (Unpublished
data from Thurston, Brauner, Neuman, and Randall.)
Measurement of gill dimensions and water flow is technically difficult and time-consuming.
Water flow over the gills and gill diffusing capacity is altered to maintain required levels of
oxygen uptake. Oxygen uptake is, in fact, an indication of the conditions for transfer across the
gills. As oxygen uptake increases so does toxicant transfer (Figure 5; Murphy and Murphy 1972,
112
-------
Rodgers and Beamish 1981). Thus oxygen uptake can be used as an indicator of conditions for
toxicant delivery and transfer across the gills.
10
o
I
OQ O
P E
8--
6--
o> .
3 4
2-
200 400
OXYGEN UPTAKE (mg/kg/hr)
—I—
600
Figure 4.
The relationship between plasma TCB (1,2,4,5-tetrachlorobenzene)
concentration and oxygen uptake in rainbow trout. (Unpublished
data from Thurston, Brauner, Neuman, and Randall.)
Toxicant uptake increased with oxygen uptake in both goldfish and trout (Figure 5)
Comparisons of regressions of the two groups of data indicated no significant difference between
the two fish species. In subsequent manipulations the data have been treated as a single set
During the initial stages of transfer, although plasma content will be higher than that in
the water, most of the chemical will be dissolved in the fat (blood being about 5% lipid) bound
to plasma proteins (Schmieder and Henry 1988), and/or ionized and the concentration of the
undissociated chemical in aqueous solution will be negligible, and can be assumed to be zero
During the early phases of toxicant uptake (for many chemicals the first few hours), therefore
the level in the water is a direct indication of the gradient between water and blood We
calculated toxicant uptake per unit gradient, from data presented in Figure 5 and values of
water toxicant concentration (Figure 6).
The slope of the regression on the data presented in Figure 6 represents the toxicant
transfer coefficient (A) for a toxicant with a log K^ above 4.3. This coefficient is expressed in
mg Oa/kg fish/hr exposure per mg/L gradient (water to blood), i.e., the uptake of toxicant per
unit toxicant gradient per unit oxygen uptake (mg/kg/hr).
A = 0,17 x oxygen uptake - 11.7
113
-------
OL/
60-
Ld
f£
Q.
:D_O> 40-
11
0 20-
i—
n-
?•
i
n D
i 1 D
A A D
ii «
1 : 1
, 0 200 400
OXYGEN UPTAKE (mg/kg/hr)
Figure 5. Total body burden of TCB in ~6 g goldfish (A) and ~9 g trout (») after
two hours exposure to low (open symbols) and high (closed symbols)
concentrations of TCB at different rates of oxygen uptake.
(Unpublished data from Thurston, Brauner, Neuman, and Randall.)
LJ
X
O
200 400
OXYGEN UPTAKE (mg/kg/hr)
600
Figure 6. The relationship between the uptake of TCB and oxygen uptake of
goldfish and trout; see Figure 5 and text for further details.
(Unpublished data from Thurston, Brauner, Neuman, and Randall.)
114
-------
Uptake rates do not change with log K^, values above about 4 (Figure 1), so this
relationship is applicable to toxicant uptake for chemicals with a log KOW greater than 4.3.
Variations in uptake of chemicals of different log K^, below 4.3 are described in Figure 1. Thus,
the relationship between the toxicant transfer coefficient for a chemical with a log K^ below 4.3
(X') and oxygen uptake can be described by the following equation based on data in Figures 1
and 6.
X' = X[10(°-429I°gKow<)-184)]
Thus chemical uptake rate for this group of chemicals can be predicted from log K,^ of
the compound in question and the oxygen uptake of the fish. There is a large data bank of
oxygen uptake measurements in fish, and in many instances uptake data can be retrieved from
the literature. It is possible, therefore, that toxicant uptake rates can be predicted for a large
number of aquatic animals and a large number of chemicals, based on this simple relationship.
The validity of this approach is presently being tested.
Hypoxic conditions lead to an increase in gill ventilation and gill diffusing capacity with
only small changes in oxygen uptake in most fish (Randall and Daxboeck 1984). Under hypoxic
conditions one would expect a much larger increase in toxicant uptake than oxygen uptake
compared with normoxic values, as observed by McKim and Goeden (1982). Fish species
respond to hypoxia in different ways, so it is difficult to generalize about the effect of hypoxia on
toxicant uptake, except that it will tend to be higher for a given oxygen uptake than in normoxia.
Toxicant will be taken up by the skin directly from the water and will not be transported
by the blood. This uptake pathway is probably rapid and could account for a large proportion
of the initial uptake, reaching equilibrium with the water far more rapidly than the rest of the
body. Nothing is known of the dynamics of skin uptake but it can constitute a large proportion
of the initial body burden, perhaps even as high as 40 to 50% of the initial uptake (Saarikoski et
al. 1986).
This manuscript attempts to derive a transfer coefficient for uptake of xenobiotic chemi-
cals across fish gills correlated to oxygen uptake. The rate of uptake of the toxicant by the fish
will be influenced by the concentration of the undissociated form in the water and the size of
the toxicant reservoir in the fish, which will depend on the size and fat content of the animal, as
will the distribution of the toxicant to various tissues. The rate at which the toxicant level in the
fish approaches equilibrium with that in the water will also depend on the rate of loss of the
toxicant (depuration) from the fish and its rate of metabolism. None of these factors is analyzed
here.
ACKNOWLEDGMENT
This work was supported, in part, by Cooperative Agreement CR816369 from the U.S.
Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia.
115
-------
REFERENCES
Black, CM., and J.R McCarthy. 1988. Dissolved organic macroinolecules reduce the uptake of
hydrophobic organic contaminants by the gills of rainbow trout (Salmo gairdneri).
Environ. Tox. Chem. 7: 593-600.
Connell, D.W. 1990. Bioaccumulation of xenobiotic compounds. CRC Press, Inc., Boca Raton,
Florida. 213 p.
Dobbs, AJ., and N. Williams. 1983. Fat solubility - A property of environmental relevance?
Chemosphere 12(1): 97.104.
Hughes, G.M. 1984. General anatomy of the gills. Pages 1-72 In: Fish Physiology, vol. 10A
W.S. Hoar and D.J. Randall (Eds). Academic Press Inc., New York.
Laurent, P. 1984. Gill internal morphology. Pages 73-183 In: Fish Physiology, Vol. 10A W.S.
Hoar and DJ. Randall (Eds). Academic Press Inc., New York.
Lin, H., and D J. Randall. 1990. The effect of varying water pH on the acidification of expired
water in rainbow trout J. Exp. Biol. 149: 149-160.
McKim, J.M., and RJ. Erickson. 1991. Environmentalampacts on the physiological mechanisms
controlling xenobiotic transfer across fish gills. Physiol. Zool. 64(1): 39-67.
McKim, J.M., and H.M. Goeden. 1982. A direct measure of the uptake efficiency of a
xenobiotic chemical across the gills of brook trout (Salveliniis fontinalis) under, normoxic
and hypoxic conditions. Comp. Biochem. Physiol. 72C(1): 65-74.
McKim, J.M., P.K. Schmieder, 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.
Murphy, P.G., and J.V. Murphy. 1971. Correlations between respiration and direct uptake of
DDT in the mosquito fish Gambusia affinis. Bull Environ. Contain. Toxicol. 6(6): 581-
588.
Randall, DJ., and C. Daxboeck. 1984. Oxygen and carbon dioxide transfer across fish gills.
Pages 263-307 In: Fish Physiology, VoL 10A W.S. Hoar and DJ. Randall (Eds).
Academic Press Inc., New York. '
Randall, D J., H. Lin, and P.A. Wright 1991. Gill water flow and the chemistry of the
boundary layer. Physiol. Zool. 64(1): 26-38.
Rodgers, D.W., and F.W.H. Beamish. 1981. Uptake of waterborne methylmercury by rainbow
trout (Salmo gairdneri) in relation to oxygen consumption and methylmercury
concentration. Can. J. Fish. Aquat. Sci. 38(11): 1309-1315.
Saarikoski, J.,R. Lindstrom, M. Tyynela, and M. Wuksela. 1986. Factors affecting the
absorption of phenolics and carboxylic acids in the guppy (Poedlia reticulata). Ecotoxicol.
Environ. Saf. 11:158-173.
Schmieder, P.K, and T.R. Henry. 1988. Plasma binding of a-butanol, phenol, nitrobenzene and
pentachlorophenol in the rainbow trout and rat: A comparative study. Comp. Biochem.
Physiol. 91C(2): 413-418.
116
-------
TOXICITY AND DISTRIBUTION OF COPPER IN AN
AQUATIC MICROCOSM UNDER DIFFERENT CONDITIONS OF
ALKALINITY AND HARDNESS
by
Jin, Hong-Jun1, Zhang, Yi-Min1, and Yang, Rong1
INTRODUCTION
The traditional single species toxicity tests have formed the major toxicological methods
available for predicting the adverse biological effects of chemicals and industrial effluents on ;
aquatic ecosystems ever since the 1940s. Microcosms are increasingly being used as surrogates
for natural systems due to their increased replicability, lower resource requirements for
investigations, and the lack of environmental damage as might occur in exposing a natural
system (Hedtke 1984). The synthetic microcosm, constructed by various laboratory-pure
cultured aquatic organisms, medium and artificial sediment, can be used to determine the
variables of microcosm response to stress. Moreover, it is regarded that this type of test system
can be developed into a standard aquatic microcosm independent of local organisms, water and
sediment (Taub and Crow 1982, Taub 1985). But it is also considered that the artificial
community set up in this way does not represent the natural biotic assembly already adapted to
the real environment.
To date, a good deal of publications on copper toxicity to aquatic life have come out
based on single species testing. The data available indicate that the toxicity of copper can be
affected by water quality such as hardness, alkalinity, pH, and the existence of chelator (Rama
Rao 1985, USEPA 1980, 1984). With regard to research on the toxicity of copper to mixed-
culture microcosms, there have been some descriptions of synthetic microcosm testing (Harrass
and Taub 1985, Taub et al. 1986, Shannon et al. 1986) and other aquatic microcosm studies
(Sugiura et al. 1982, Hedtke 1984, Yasuno 1985) over the past few years. But few papers on the
relationship between water quality and microcosm toxicity of copper have been found in the
literature.
The distribution of heavy metals in aquatic ecosystems depends on the physico-chemical
properties of water (Forstner 1981). The copper uptake and retention by aquatic organisms is
related to water quality factors (Winner 1985). Microcosms are useful for investigating the fate
and effect of chemical contaminants in a more realistic context (Giddings 1983).
In this study, with appropriate test concentrations selected by referring to the microcosm
toxicity data available for copper, the effects of copper added at a given concentration on algal
community, Daphnia magna population, and the metabolism of microcosm were comparatively
investigated under different conditions of alkalinity and hardness. Furthermore, the distribution
of copper added at three concentrations were also studied in the microcosms with different
alkalinity and hardness. The test systems were similar to Taub's synthetic microcosm mentioned
above.
Department of Environmental Sciences, Nanjing University, Nanjing, Jiangsu, PRC.
117
-------
MATERIALS AND METHODS
MICROCOSMS
The microcosms were set up in acid-washed 5-L glass jars. The artificial lake water was
T63MV medium (Taub and Crow 1980), prepared with deionized water (0.02 jimhos). The
standard sediment (Taub and Crow 1980) was used, the components of which were rinsed with
10% HC1 and deionized water in advance. The artificial community consisted of pure cultured
organisms inoculated, including eight species of algae, two species of protozoa, and two species
of metazoa, as well as the natural microbial community existing in the laboratory (Table 1). All
species of algae were stored in solid agar and then transferred to the liquid T63MV medium for
expanding culture before constructing the microcosms. Prior to the inoculation, Tetrahymena
pyriformis Ehrenberg was proliferatively cultured in peptone medium, while Paramecium,
Dap i,iia and rotifer were acclimated in T63MV medium. Initially, standard sediment was
placed at the bottom of the vessel and 3 L of medium added. Then eight species of algae were
inoculated and incubated at 20 ± 1 °C with illumination of fluorescent light (1993 ± 29 lux) on
a 12-h light-dark cycle. This algal community was allowed to grow and develop in the absence
of grazers for 6 days under the aforesaid condition. Then protozoa and metazoa were
inoculated, and the vitamin components added simutaneously. The densities of all pure cultured
organisms inoculated are shown in Table 1. Copper was delivered 3 days later and microcosms
were kept for further culture under the same conditions. Test duration lasted 28 to 35 days
from the day copper was delivered to the end of the copper exposure. All variables were
determined during this period.
Table 1. Initial concentrations of organisms in microcosms.
Trophic Level Group
Species
Concentration
(individuals/ml)
Producer
Algae
Consumer
Protozoa
Metazoa
Decomposer Bacteria
Selenastrum capricomutum
Ankistrodesmus falcatus
Chlamydomonas microsphaera
Scenedesmus obliquus
Stigeoclonium sp.
Chlorella pyrenoidosa
Anabaena 7120
Nitzschia sp.
Tetrahymena pyriformis Ehrenberg
Paramecium sp. Hill
Daphnia magna
Rotifera sp.
Natural microbial community
existing in the lab
104
104
103
103
104
103
3
0.68
*
0.25
*4 adults with embryos/3 L; 5 young/3 L; 10 larvae/3 L.
118
-------
WATER QUALITY AND COPPER TREATMENT
The solutions of CaCl22H20 (1.470 x 10s mg/L) and NaHCO3 (4.200 x 104 mg/L) were
respectively employed to adjust alkalinity and hardness. The two water quality parameters were
determined separately by means of EDTA titrimeteric method and hydrochloric acid titrimetric
method at both the beginning and the end of the experiment (APHA-AWWA-WPCF 1985).
For copper treatment, standard stock copper solution (1000 mg/L) was employed (China EPA
1983). Overall, the microcosm copper exposure tests were divided into three runs with different
water quality simulated. The first run was designed to observe the microcosm responses to
50 ng Cu/L added and the copper distribution in the microcosms under low alkalinity conditions
at low and high hardness. The second run was similarly set as the first one apart from the high
alkalinity condition. The third run aimed at surveying the distribution of copper added at
200 and 2000 ng/L in the microcosms under low alkalinity at low and high hardness, and under
low hardness at low and high alkalinity.
MICROCOSM RESPONSE VARIABLES
In aquatic ecosystems, the ratio of gross primary productivity (GPP) to total ecosystem
respiration (Re), usually referred to as P/R, is an important index to evaluate the ecosystem
metabolism, and can be estimated by the amount of dissolved oxygen (D.O.) change during the
12-h light and dark periods. The daytime net primary productivity (PI) and the overnight
respiration (Rl) were calculated (Giddings 1986). On the basis of ecological principles, the GPP
is the sum of daytime net primary productivity (PI) and daytime respiration. Thus, P = P1+R1,
R = 2R1; and P/R = (Pl/Rl+l)/2, on the assumption that daytime respiration is equivalent to
overnight respiration (Rl) in aquatic ecosystems. The units of P and R are mg O2/L-24h. The
algal community biomass was expressed as chlorophyll a concentration (jig chlorophyll a/L) ,
which was measured periodically during the test (APHA-AWWA-WPCF 1985). The density of
Daphnia was represented as individuals per 100 ml water sample.
TOTAL COPPER MEASUREMENT AND THE PARTITION/DISTRIBUTION OF COPPER IN THE
MICROCOSMS
At the end of the test, certain amounts of water, organisms, and sediment were sampled
and then treated with nitric acid-hydrochloric acid digestion. The total copper concentrations
were measured by atomic absorption spectrometric method (APHA-AWWA-WPCF 1985), with
a Shengyang SF-402A atomic absorption spectrometer or a Phillips PU-900 atomic absorption
spectrometer.
The distribution of copper was expressed as partition coefficients (K) among the
compartments in the synthesized aquatic microcosms. These factors included organism/water
partition coefficient, Kbw; i.e., biological accumulation factor (BAF); sediment/water partition
coefficient, Ksew; and Daphnia/algae partition coefficient, Kda; i.e., biological magnification
factor (BMF) along food chain (algae-Daphnio), which was the ratio of total copper
concentrations in the two biological compartments of Daphnia and algae.
DATA STATISTICAL ANALYSIS
Data on all variables were compared using two-way analysis of variance (ANOVA), and
Duncan's Multiple Range Test was employed if the BAFs for algae and Daphnia differed among
treatments.
119
-------
MICROCOSM RESPONSES TO 50
ALKALINITY AND HARDNESS
RESULTS
Cu/L ADDED UNDER DIFFERENT CONDITIONS OF
Figures 1, 2, and 3 show respectively the variations of chlorophyll a concentration, Daphnia
density, and P/R ratio with exposure time to 50 \ig Cu/L added under low alkalinity (34.94 ± 4.53
mg CaCOa/L) at two levels of hardness (95.76 ± 6.98 and 475.81 ± 11.19 mg CaCOa/L). The
chlorophyll a concentrations of both control and treatment groups increased 5 days after copper
was delivered and then gradually decreased, reaching the minimum values by day 11 (Figure 1).
Thereafter, chlorophyll a concentrations rose up and all attained the climax by day 25. The
Daphnia density of each group appeared to fall moderately 4 days from copper treatment, then
increased by degrees until day 10, and again dropped continuously before picking up from day 16
(Figure 2). The comparison of Figures 1 and 2 suggested the predator-prey bearing between
Daphnia population and algal community in the development of microcosms. The P/R ratio of
each group reduced 10 days after copper treatment, and the minimum values emerged under low
alkalinity at low hardness for both copper treatment and control groups. From then on, the P/R
ratios increased, reaching about 1 by day 22. Along with the microcosm succession, P/R ratios on
day 30 attained approximately 0.9. The data on chlorophyll a concentration, Daphnia density, and
P/R ratio were compared using ANOVA to describe the effects of 50 jig Cu/L added alone,
hardness alone, and the interaction between these two factors on aquatic microcosm under low
alkalinity. The result suggested that copper treatment, hardness, and their interaction didn't
significantly influence the three biological parameters in the development of microcosms under
low alkalinity conditions (p > 0.05).
1000 -,
(tf
rH
900 -
¥ 800
a
O
3 700
ttf
I 600
-------
3 10°
o
2 90
CO
3
•a
to
o
•H
T3
-------
Figures 4, 5, and 6 showed, respectively, the variations of chlorophyll a concentration,
Daphnla density and P/R ratio with exposure time to 50 ug Cu/L added under high alkalinity
(13516 ± 6.55 mg CaCOs/L) at two levels of hardness (95.92 ± 6.48 mg/L and 452.50 ± 14.71
me CaCOa/L). Comparison of Figures 4 and 1 indicated that the peaks of chlorophyll a
concentrations in both 50 ug Cu/L treatment and control groups under high alkalinity at
different hardness appeared earlier than those under low alkalinity conditions. However, the
chlorophyll a concentrations on day 25 under low alkalinity were higher than that under high
alkalinity, which seemed to suggest that high alkalinity was unfavorable to the growth and
reproduction of algae in microcosm succession. The chorophyll a concentration climax of each
group occurred 9-15 days after copper was delivered under high alkalinity conditions. From
then on the chlorophyll a concentration of each group dropped gradually to the initial levels
(Figure 4). According to the results of ANOVA, the chlorophyll a concentration of 50 v-g Cu/L
treated group was significantly higher than that of control on day 15 (p < 0.05). The other
factor hardness increase, could lead to the significant decrease of chlorophyll a concentration on
day 9 compared with control (p < 0.05). The chlorophyll a concentrations on day 4 and 27
were significantly influenced by the interaction between copper treatment and hardness
(p < 0 05). Under high alkalinity, the Daphnia density of each treatment and control group
increased significantly (Figure 5) from day 5 and the peak appeared at the time algal community
biomass decreased a lot (Figure 4), which illustrated the predator-prey relationship between
Daphnia and algae.
1200-
1
g 1000-
o
3
oi
c 800-
o
o
a
o
600-
p.
§ 400-1
200
O hiSfc alkalinity & low hardness, control
A high alkalinity & low hardness, 50ug cu/l
• high alkalinity & high hardness, control
D high alkalinity & high hardness, JO ujj Cu/l
15
Time (day)
22
2?
Figure 4. Variations of chlorophyll a concentrations with exposure time
to 50 |ig Cu/L added under high alkalinity conditions.
122
-------
'O high alkalinity & low hardness, control
A high alkalinity & low hardness, 50ug cu/1
• high alkalinity & high hardness, control
D high al«calinity & high hardness, 50ug Cu/1
Time (day)
Figure 5. Variations olDaphnia densities with exposure time
to 50 \ig Cu/L added under high alkalinity conditions.
1.2-
to
S l.OH
ttf
0.8-
CM 0.6-
O./f-
0.2-
O high alkalinity & low hardness, control
A high alkalinity & low hardness, pOug Cu/1
0 high alkalinity & high hardness, control
alkalinity & high hardness, ^Oug Cu/1
10
18
22
Time (day)
26
30
Figure 6. Variations of P/R ratios in the exposure duration
of 50 \ig Cu/L added under high alkalinity conditions.
123
-------
Statistical analysis indicated that Daphnia density on day 22 in 50 ug Cu/L treated group
was significantly lower than that in control (p < 0.05), but it was not significantly affected by
hardness (p > 0.05). The interaction between copper treatment and hardness, however, had
significant effect on Dapknia density 5 and 10 days after copper treatment (p < 0.05 and
p < 0.01, respectively). Comparison of Figures 3 and 6 suggested that the dynamic patterns of
P/R ratio were different in all groups under low and high alkalinity conditions in the microcosm
test duration. The lower P/R ratio under high alkalinity in the later stage of the test was
probably due to low levels of chlorphyll a concentration and primary production. It is also
revealed that copper and hardness, as well as the interaction between these two factors, didn't
significantly influence the mean values of P/R ratios in both the last 10 days and the whole test
duration under high alkalinity (p > 0.05).
COPPER PARTITION/DISTRIBUTION IN MICROCOSMS UNDER DIFFERENT CONDITIONS OF
ALKALINITY AND HARDNESS
Tables 2 and 3 show the partition of copper at three concentrations among the
compartments of microcosms under different conditions of alkalinity and hardness. The
algae/water partition coefficient of copper; i.e., the biological accumulation factor (BAF) of
copper for algae, was 103-104; the Daphnia/\vater partition coefficient; Le., BAF of copper for
Daphnia, was 10M03; the Daphnia/algae partition coefficient; i.e., the biological magnification
factor (BMP) along food chain (algae-Dophnia), was lO'MO'2; and the sediment/water partition
coefficient of copper had the order of 101.
Table 2. Partition coefficients of 50 ug Cu/L added between the different compartments in the
microcosm under various water quality conditions.
Alkalinity & hardness
(mg/L, CaCO3)
Low alkalinity (34.94±4.53)*
Low hardness High hardness
(95.76±6.98) (475.81+11.91)
High alkalinity (135.16±6.55)
Low hardness High hardness
(95.92±6.48) (45230+14.71)
Algae/Water (xl(P) 8.488±3.024 11.09±4.27 5.187±1.935 6.255±1.911
jDop/uiw/Water (xlO3) 1.243±0.344 1.139+.0.694 0.6734±0.2146 0.4930±0.2836
Sediment/Water 22.90±6.96 29.16±7.21 22.58+6.69 22.84±3.16
Daphnia/Algae 0.1603±0.0651 0.09809+0.06980 0.1213+0.0050 0.0921 2± 0.00310
•Mean ± SD of 3 replicate systems.
The relationship between the BAF for algae or Daphnia and copper concentrations at
different alkalinity and hardness was analyzed. The ANOVA suggested that under low alkalinity
the concentration of copper added didn't significantly influence the BAF for Daphnia or algae
(p > 0.05), while the change of hardness could result in significant effect on the BAF for algae
(p < 0.05) but no effect on the BAF for Daphnia (p > 0.05). Moreover, the interaction
between copper treatment and hardness wouldn't affect the BAF for Daphnia or algae
significantly (p > 0,05). Uader low hardness, copper exposure at three concentrations had
significant effect on the BAF for Daphnia (p < 0.05), but made no difference among treatments
124
-------
for algae (p > 0.05). However, the change of alkalinity, as well as the interaction between
copper treatment and alkalinity, could lead to significant effects on the BAF for algae
(p < 0.01), but had no effects on the BAF for Daphnia (p > 0.05). Duncan's Multiple Range
Test showed that, under low hardness, there were significant differences in the BAFs for
Daphnia between the 50 and 200 ug Cu/L, and 200 and 2000 ug Cu/L groups (p < 0.01).
However, the difference between 50 and 2000 jig Cu/L groups was not significant (p > 0.05).
Table 3. Partition coefficients of 200 ug and 2000 jig Cu/L added between the different
compartments in the microcosm under various water quality conditions.
Alkalinity & hardness
(mg/L CaCO3)
Copper added (|ig Cu/L)
Algae/Water (xlO3)
Dop/i/ua/Water (xlO3)
Sediment/Water
Daphnia/Algae
High alkalinity
(151.30±9.18)*
Low hardness
(61.68±6.55)
200 2000
15.56 . 14.74
±0.87 ±8.76
1.766 0.5676
±0.320 ±0.2350
38.48 27.92
±6.80 ±5.51
0.1135 0.05173
±0.0193 ±0.01848
Low alkalinity
(20.53±2.73)
High hardness
(453.67±2.73)
200 2000
a455 10.35
±5.209 ±1.26
1.352 2.437
±0.243 ±1.197
28.26 22.70
±4.05 ±1.63
0.1572 0.2359
±0.0792 ±0.1119
Low hardness
(61.85±6.55)
200 2000
5.537 5.448
±0.553 ±1.623
1.151 0.6087
±0.768 ±0.1710
12.42 18.43
±1.79 ±0.69
0.1226 0.1281
±0.1142 ±0.0832
*Mean ± SD of 3 replicate systems.
DISCUSSION
In recent years, there have been many publications on the toxicity of copper to aquatic
microcosms, and some of the studies concerned synthesized aquatic microcosms (Harrass and
Taub 1985, Shannon et al. 1986, Taub et al. 1986). However, no reports have been found about
the relationship between copper toxicity on microcosms and water quality. The change of water
quality, including hardness, pH, D.O., temperature, etc., should be taken into account in the risk
assessment of chemicals, for the response of aquatic organisms to chemicals can be affected by
the physico-chemical properties of water. Today, knowledge of the factors influencing the
toxicity of chemicals is based almost entirely on laboratory single-species toxicity tests and
virtually without validation in natural aquatic ecosystems. The development of multispecies
toxicity testing procedures provides a means of evaluating responses of more complex systems
than single species. The results of this study demonstrated that in addition to hardness,
alkalinity was also an important factor affecting the toxicity of copper to microcosms, and a
great increase of alkalinity could reinforce copper effects.
The microcosms employed in this study were similar to Taub's aquatic microcosm,
although there were differences in the biotic components and the laboratory conditions (Taub
and Crow 1980). Taub et al. (1986) found that algal biovolume peaks in treated microcosms
with 500, 1000, and 2000 ug Cu/L were always delayed relative to the controls. In this study, the
patterns of chlorophyll a concentration and Daphnia density dynamics during the test period
under low alkalinity, whatever the hardness, are quite different from those under high alkalinity
125
-------
conditions. The delay of chlorophyll a concentration peak was only seen in 50 jig Cu/L
exposure systems under high alkalinity at low hardness. Under high alkalinity, whatever the
hardness, algal community biomass decreased rapidly as Daphnia population was growing during
the later period in the treated and control microcosms, as compared with the results of tests
under low alkalinity. It seems that such a water quality condition was unfavorable to algal
community but beneficial to grazers.
The microcosm protocol used in this study calls for the development of a defined
community from pure stock cultures of algae and small aquatic animals. The species assemblage
does not simulate a specific community, but serves as a generalized model aquatic ecosystem
that includes species at the primary and secondary trophic levels as well as decomposers.
Hammons (1981) expressed concern that artificial communities may not be representative of
"nature, co-adopted species assemblage" and suggested that they may not be reliable for studies
of ecosystem-level properties. Stay and Katko (1988) pointed out that microcosm bioassays are
an ecosystem level analog of single species chronic tests. At present, however, there does not
appear to be a generally acceptable method for using microcosm or ecosystem level response
variables in the development of toxicity criteria for chemicals introduced into the environment.
Although some aquatic microcosm tests offer ecosystem level information that cannot be
inferred from conventional bioassays (Shannon et al. 1986, Stay and Katko 1988), the results of
this study indicated that ecosystem level parameters in microcosms were not always sensitive to
copper and other stress. The lack of correspondence between P/R ratio and changes in Daphnia
population and algal community supports the hypothesis that the response of population-
community hierarchy cannot be used to infer changes in the process-functional level.
The results presented here showed that chlorophyll a concentration and Daphnia density
in microcosms were quite sensitive to copper stress. Because of the reduced species numbers,
microcosm tests may be more sensitive to stress than more diverse nature communities
(Giddings 1983). Also, the protocol used in this study allowed only 9 days before treatment;
copper exposure occurred during a highly sensitive stage of ecosystem development.
Consequently, the magnitude of effects may be greater than would be observed in aged, stable
systems. Despite the possible problem, this protocol appears to be appropriate to the
objectives, namely, the comparative studies on the toxicity and distribution of copper in the
microcosms under different alkalinity and hardness conditions. Prolonging the test duration,
however, would be better for more reasonable results.
Microcosm bioassays should accurately reflect the concentrations of chemicals that cause
changes in natural systems, and hazard assessments should provide information at all levels in
ecosystems. Taub's microcosm protocol was developed with copper for several years, but no
minimum effect level of copper to the microcosm was determined (Shannon et al. 1986). The
sensitivity of the microcosm constructed in this laboratory, to copper or other stress, remains to
be investigated.
Microcosm tests are different from single species tests. The response of a microcosm to a
toxicant is the result of interactions among many biotic and abiotic factors. Once a toxicant is
added to a microcosm, it is quickly partitioned among the biota, the liquid phase, and other
abiotic components of the system (Shannon et al. 1986). In the present study, the aqueous
copper levels in the treated microcosms were not routinely measured because the measurement
of copper in the liquid phase is of little value in determining the amount of copper in the
exposure systems. The partition of copper among the biota, liquid phase, and sediment of
microcosms should be investigated, however.
126
-------
Copper is sorbed rapidly to sediments, resulting in high residue levels. The rate of sorption
varies with the type of clay/sediment, pH, competing cations, and the presence of ligands and
Fe/Mn oxides. In the sediments of some fresh and marine waters, there are very high levels of
total copper (Moore and Ramamoorthy 1984). The low coefficients of copper between sediment
and water under different conditions of alkalinity and hardness were probably associated with the
characteristics mentioned above of the sediment used in this study.
Large amounts of data are now available on the bioconcentration/bioaccumulation of copper
in single-species test systems and natural waters, but there are relatively few food chain studies on
the transfer of copper to higher trophic levels (Moore and Ramamoorthy 1984). This study
demonstrated that alkalinity and hardness not only affected the toxicity of copper, but also
influenced the partition/distribution of the metal in microcosms. The results of this study also
showed that the uptake of copper from water is more important than ingestion from food, and
biomagnification via the food chain (algae-Daphnia) was insignificant
Although this microcosm protocol is useful to investigate the toxicity and fate of toxicants in
aquatic ecosystems, the amount of time or manpower required to conduct the microcosm tests in
this study was substantially greater than that needed for a conventional 21-day chronic test with
Daphnia or partial life-cycle tests with other species. This type of microcosm test is very expensive
for routine use and may be more appropriately used at later stages of a hazard assessment
procedure. Nevertheless, there are still certain issues to be further studied.
ACKNOWLEDGEMENT
This study was supported by the National Natural Science Foundation of China.
REFERENCES
APHA-AWWA-WPCF. 1985. Standard Methods for the Examination of Water and Wastewater.
16th Edition. American Public Health Association, Washington, D.C. pp 199-201- 204-207-
269-271; 1067-1072. . r '
China EPA 1983. Pages 67-68 In: Analytical Methods for Environmental Monitoring. China
Environmental Protection Agency, Beijing, pp. 67-68.
Forstner, U., and G. Miller. 1981. Pages 271-318 In: Heavy metals in aquatic organisms. In:
Metals Pollution on the Aquatic Environment (2nd Ed.) Springer-Verlag, Berlin.
Giddings, J.M. 1983. Microcosms for assessment of chemical effects on the properties of aquatic
ecosystems. Pages 45-94 In: Hazard Assessment of Chemicals: Current Developments
Vol. 2. J. Saxena (Ed.). Academic Press, New York.
Giddings, J.M. 1986. Microcosm procedure for determining safe levels of chemical exposure in
shallow water communities. Pages 121-134 In: Community Toxicity Testing, ASTM STP920.
J. Cairns, Jr. (Ed.). American Society for Testing and Materials, Philadelphia.
127
-------
Hammons, AS. (Ed.). 1981. Methods for Ecological Toxicology: A Critical Review of Laboratory
Multispecies Tests. ORNL-5708, EPA-560/11-80-026. Ann Arbor Science, Ann Arbor,
Michigan, 307 p.
Harrass, M., and F. Taub. 1985. Comparison of laboratory microcosms and field responses to
copper. Pages 57-74 In: Validation and Predictability of Laboratory Methods for Assessing
the Fate and Effects of Contaminants in Aquatic Ecosystems, ASTM STP865. T. Boyle
(Ed.). American Society for Testing and Materials, Philadelphia.
Hedtke, S.T. 1984. Structure and function of copper-stressed aquatic microcosms. Aquat. Toxicol.
5:227-244.
Moore, J.W., and S. Ramamoorthy (Ed.). 1984. Pages 77-99 In: Heavy Metals in Natural Waters,
Springer-Verlag New York Inc., New York.
Rama, Rao, S.V. 1985. Effect of metal chelation on the toxicity of some environmentally
hazardous trace metals to Daphnia magna. Intern. J. Environ. Studies. 26: 87-90.
Shannon, L.J., et al. 1986. A comparison of mixed flask culture and standardized laboratory model
ecosystems for toxicity testing. Pages 135-157 In: Community Toxicity Testing, ASTM
STP920. J. Cairns, Jr. (Ed.). American Society for Testing and Materials, Philadelphia.
Stay, F.S., and A. Katko. 1988. Effects of fluorene on microcosms developed from four natural
communities. Environ. Toxicol. Chem. 7:635-644.
Sugiura, K., et al. 1982. Effect of Cu stress on an aquatic microcosm: a holistic study. Environ.
Res. 27: 307-375.
Taub, F.B. 1985. Toward interlaboratory (round-robin) testing of a standardized aquatic
microcosm. Pages 165-186 In: Multispecies Toxicity Testing. J. Cairns, Jr. (Ed.). Pergamon
Press Inc., New York.
Taub, F.B., and M.F. Crow. 1980. Synthesizing aquatic microcosms. Pages 69-104 In: Microcosms
in Ecological Research. J. P. Giesy (Ed.). DOE Symp. Series 52, Dept. of Energy,
Springfield, Virginia.
Taub, F.B. et al. 1986. Preliminary results of interlaboratory testing of a standarized aquatic
microcosm. Pages 93-120 In: Community Toxicity Testing, ASTM STP920. J. Cairns, Jr.
(Ed.). American Society for Testing and Materials, Philadelphia.
USEPA. 1980. Ambient water quality criteria for copper. U.S. Environmental Protection Agency,
440/5-80-036. U.S. Government Printing Office, Washington, D.C.
USEPA. 1984. Ambient water quality criteria for copper. U. S. Environmental Protection Agency
440/5-84-031. National Technical Information Service. Springfield, Virginia.
Winner, R.W. 1985. Bioaccumulation and toxicity of copper as affected by interactions between
humic acid and water hardness. Water Res. 19: 449-455.
Yasuno, M. 1985. Hazard assessment of toxic substances using model aquatic ecosystems.
Pages 56-73 In: Symp. Biomonitoring State Environ.
128
-------
ENVIRONMENTAL IMPLICATIONS OF METAL COMPLEXATION
Liu, Jennie Jingyi1
ABSTRACT
A vast number of metal complexes are existing and forming in the environment; their
formation and interaction of metal ion and ligands are governed by various prevailing
environmental conditions. Some environmental chemical aspects are briefly mentioned and
emphasized.
L The importance of metal speciation and its estimation for understanding trace metal
transport in aquatic systems are emphasized. Species in heterogeneous systems need to
be developed.
n. Hydrolytic and polymeric species of metals are complicated and need to be studied; e.g.,
those of aluminum in soil. Distribution, transport and fate of metals in aquatic systems
are significantly influenced by those dissolved organic matters which are capable of
increasing solubility or mobility of metals in the aquatic environment Humic
substances, humic acid and fulvic acid are the important chelators, but the more basic
data are far from sufficient.
III. Mobility and bioavailability of sediment-bound metals are correlated in the
environmental processes, naturally or anthropogenically. Sediment-water exchange can
be considered a predictor of metal bioavailability from the sediment. A multi-
chambered device to study metal sediment with different surface interaction is
introduced.
IV. Metal complexation plays an important role in the transport, transformation and fate of
metals in the aquatic systems. Studies on the environmental chemical behavior of
mercury complexes in Ji River and surface binding of Cd in Xiang River sediment can
be taken as examples.
INTRODUCTION
Metals and their compounds are part of the environment and exert their beneficial or
detrimental effects on biological systems through specific, but largely unknown, chemical or
physico-chemical interactions with molecules essential for life or natural substances in the
environment.
Research Center for Eco-environmental Sciences, Academia Sinica, Beijing, PRC.
129
-------
Aqueous solution or natural aquatic environments provide media in which chemical and
biochemical reactions proceed under the influence and control of other dissolved substances.
Solid matters-suspended in water or deposited in sediment-offer adsorption sites for dissolved
substances and provide active'surface to influence the rate of reactions. Biota with their
variability in size, function and biochemistry assume a multitude of interaction with metals and
relevant chemical compounds.
The thirty essential and toxic elements (Table 1) might join with the available ligands to
produce several million coordination compounds, and their formation is governed under various
prevailing environmental conditions. In recent years, metal with more organic compounds in
nature and interaction of metals with relevant compounds or ligands in heterogeneous systems
together with their models have begun to receive much attention.
Table 1. Essential and toxic elements in the environment.
Essential Elements
H, Na, K
Mg,Ca
V
Cr, Mo
Mn
Fe
Co
Ni
Cu
Zh
Cd
C, Si, Sn
N, P, As
O, S, Se
F, Cl, I
Toxic Elements
Be, Ba
Ag
Hg
A1,T1
Pb
As, Sb
Se
UnconGrmed Growth
Stimulation
Cr
Cd
Sn
(Irgolic and Martell 1985)
As the topic on environmental implications of metal complexation is quite a large field,
some environmental chemical aspects of metal complexation, particularly in heterogeneous
phase, will be briefly reviewed here.
130
-------
CHEMICAL SPECIAT1ON
In solutions, metals occur as a suite of species distributed among soluble ligands, and
aquatic biota are exposed to all species that occur in water (Figure 1). In natural waters metals
are partitioned among aquatic, suspended particulates and bottom sediment phases. In
sediments, metals also partition among different types of ligands associated with various
compounds of the particulates.
Free metal
ions
Examples:
Cu(II).aq.
Fe(ll).aq.
Inorganic
ion pairs,
Inorganic
complexes
Cu2(OH)22+
Zn(OH)3-
CdCl+
Organic
complexes,
Chelates
Me.SR
ME.OOCR
CH2-C=O
/ \
NH2 O
N. f
XCu
/\
°\r
0=C-CH
Metal
species
bound to
high
molecular
weight
organic
Me-lipids
Me-humic
acid,
Polymers
Metal
species
in the
form of
dispersed
colloids
FeOOR
Fe(OH)
Mn(IV)
oxides
Metal
species
sorbed on
colloids
Me^OH),
on clays
FeOOH or
Mn(IV)on
oxides
Precipi-
tates,
Organic
particles,
Remains of
living
organisms
(Stumm and Morgan 1981)
Figure 1. Forms of occurrence of metal species.
It is now well established that speciation measurements are necessary for an
understanding of trace metal transport in aquatic systems and study of aquatic toxicity.
Most studies of the metal toxicity toward aquatic biota have shown that the hydrated
metal ion is the most toxic form; e.g., in the case of Cu, hydroxy complexes are believed to have
some toxicity. Most stable complexes and species associated with colloidal particles are not toxic
except those complexes that are lipid soluble as they can penetrate into the biomembrane .
rapidly. Based on the free ion activity (not total concentration), the order of toxicity was found
to be: Hg > Ag > Cu > Pb > Cd > Zn > Tl. In the presence of complexing agents, this
order changes dramatically. For instance, if the river water contains 50 ug/L Cu, it is all
complexed with fulvic acid or tannic acid, and the water will be completely non-toxic toward
aquatic organisms. If the river is polluted by ethylxanthogenate complex, a flotation agent from
ore'processing plants, 50 ug/L Cu will exist as the highly toxic Cu-ethylxanthogenate complex
Using fish and algae in bioassays, toxicity is related to free metal ion activity.
131
-------
ESTIMATION OF CHEMICAL SPECIATION
Experimental determination of the nature and concentration of all metal complexes
potentially present in the environment is practically impossible. Reasonable estimates of the
stability of metal complexes can be obtained even in multi-component systems, provided the
stability constants for individual complexes are available.
Knowledge of specific reactions within solution is generally more advanced than that in
sediments. The metal speciation process following thermodynamic principles and models to
describe metal species-determination are available. However, either analytical methodology or
computational methods are not satisfactorily developed for metal speciation in natural
sediments. Experimental approaches reviewed a number of operational methods and titration
approaches to evaluate the stability constants and complexing capacities (Table 2).
Table 2. Methods for assisting in the specific identification of individual species.
Method or Principle
Physical-mechanical separation based
on size (MW), density or charge
Auxiliary equilibria
A familiar equilibrium system
introduced to provide indication
for the species
Equilibrium potentiometric methods
Electrode kinetics
Interdependence of current, potential
& time for given electrode processes
Direct detection of electrode or
atomic structure
Bioassay & catalytic effect
Examples
Membrane filtration
Gel filtration
Effect of complex formation on acid-base
equilibrium, adsorption, redox reaction
Ion selective electrodes (ISE)
Polarography
Optical methods
Batch or continuous culture
(Stumm and Morgan 1981)
Sequential extraction procedures are useful in predicting metal availability among
geochemically similar sediments or soils, but no universal extractant procedure alone would
closely define the availability of all metals. It is available for characterizing sediments but
unable to separate them into specific species. Relevant modeling studies are being developed,
but quantitative application seems premature at present Current accurate stability constants
are available only for approximately 200 organic and inorganic ligands with metal ions present in
132
-------
environmental systems. New ligands are generated in the environment mainly from organics as
a result of chemical, photochemical and biological processes. Expansion of these data is
urgently needed. Theoretical and empirical approaches that have been employed for the
estimation of stability constants in homogeneous systems are far from satisfactory. Metal-ligand
stabilities may be predicted from simple monodentate donors and may be extended only to
bidentate donors.
Species in heterogeneous systems are much more complicated. In natural systems,
equilibria are established among dissolved metal complexes and metal complexes adsorbed on
colloidal materials, soil particles and clays. Metal complexes may exchange ligands, which they
bind in solution for donor atoms in the solid materials upon contact with such phases. The
concentration of metals is typically much higher in the solid phase, and the higher the affinity of
a metal for coordinating sites in particulate, the more completely will the metal be removed
from the aqueous phase. At present there are several concepts developed on surface and
colloidal chemistry. The surface complexation concept has been developed rather quickly, and
still needs to be developed further.
SOME METAL IONS AND LIGANDS IN THE ENVIRONMENT
There are many data available on the occurrence of heavy metals; e.g., Cd, Pb, Cu, in
large varieties of environmental samples and fair knowledge on the fate and accumulation of
heavy metals,in aquatic systems. As soil is a dynamic system and heterogeneous in nature, the
transferring processes from soil to plants and animals are much more complicated, and basic
knowledge on interactions of undergoing processes in terrestrial ecosystems is insufficiently
known. In recent years attention on soil kinetics is growing, partly due to the increase of
different kinds of stress to which the soil is subjected. Model studies involving kinetics may
contribute to the understanding of relevant processes. Much more attention to sediments, soils,
solid materials (including dredged materials, sewage sludge and mine tailings) is under
consideration.
In addition, a number of environmental problems of global concern have been raised;
e.g., Al toxicity in aquatic systems and soils, ground water contamination and solid wastes. Trace
metal contamination together with relevant environmental processes including metal
complexation remains an important topic to be dealt with.
HYDROXYL AND POLYMERIC ALUMINUM
Soluble Al and colloidal A13+-OH- complexes have been shown to cause injury at ppm
level to fish and some other aquatic organisms maintained under controlled laboratory
conditions. However, the toxicity of Al is not readily related to any observed or potential
environmental hazard, and it appears likely that the effects of Al on aquatic life are greatly
ameliorated by circumstances of exposure in natural waters.
The forms of occurrence of Al in water are related to pH and age of solution as well as
to the presence of complexing ions and organic substances.
133
-------
The lethal limit for many aquatic organisms occurs at pH 4. Progressive hydrolysis of
A1(III) leads to cation and final colloidal forms. Distribution of chemical species of Al is
complicated because of numerous hydrolytic intermediates formed prior to the precipitation of
x A13+ + y H2O »* AI,
A1(OH)2+
y H+
There is a strong tendency for Al(ni) to form dimeric, oligomeric and polymeric species,
enhanced as Al/OH decreases. For example:
3 Al2(OH)2
-------
Those hydroxides of Fe(DI), Al(m), Mn(IV) forming polymeric complexes vary in
composition with environmental conditions. Metal ions' interaction with those colloids,
participates and organic matters, will be adsorbed or chemically bound to inorganic colloids or
surface of biota, and play an important role in the transport and transformation of metal ions in
the environment.
Table 3. Naturally occurring organic substances.
Life Substances
Decomposition Intermediates
Intermediates & Products
Typically Found in Natural
Waters
(1) Proteins
Polypeptides -> aminoacids ->
RCOOH, RCH2OH, RCH3, RCH2NH2
Fatty acids + glycerol -> RCH, OH,
RCOOH, RH, acids
Mono -> polysaccharides -> hexoses,
pentoses --> polydroxy-carboxylic acids
NH.4+, CO* CH4, HPO^, amino
acids, phenols, fatty acids,
mercaptans
CO2, CHx, acetic, lactic, citric,
malic, oleic acids, carbohydrates,
hydrocarbons
, CO2, CH4, glucose, etc.
(2) Lipids, fcts,
waxes, oils,
hydrocarbons
(3) Carbohydrates
Cellulose
Starch
Lignin
(4) Porphyrins &
plant pigments
Chlorophyll
Hemin
Carotenes &
xantophylls
(5) Complex substances formed from breakdown intermediates, phenols, amino compounds -->
humic acid, fulvic acid, "gelbstoffe", etc.
~> hydrocarbons
Phytane
carotenoids
alcohols, ketones, acids
porphyrins
(Stumm and Morgan 1981)
Much less understood are the interactions among ions, natural organic matter and colloidal
surface. With those synthetic organic chelators; e.g., EDTA, NTA, most metals are known to
form stable complexes and less stable complexes with HA, FA As many organic matters exist
in colloidal forms, it does seem to be very effective for occluding and adsorbing metals. A
considerable fraction of Cu, Cd, Pb was found to be bound to organic colloidal matter in
seawater.
. At present various chemical and biological methods are available to measure the
complexing capacity of these ligands. Metal (e.g., Cu) complexation and detoxification by
certain organic substrates liberated by some algae were reported. Algae extracellular products
might produce the ameliorating effects on metal toxicity in natural aquatic environments. Some
blue-green algae can produce some weak acids with strong complexing abilities; e.g., hydroxamic
acids, amino acids, siderphores, algae polysaccharides, carragenates which can complex with
metals (Cd, Pb) and reduce their toxicity. A series of algae assays with Zn and a series of
chelators; e.g., EDTA, oxydisuccinic acid, carboxymethyloxvsuccinic acid and carboxymethyl-
tartaric acid have been studied. Algae toxicity is related to the free metal ion and not to the
total metal ion. In those soluble complexes form, heavy metals are rendered unavailable to
algae, and their toxicity decreases. This is also true in the sedimentary phases.
135
-------
Simple models have been established to evaluate the complexation reaction between the
coordination sites of the soluble ligands and those of the surface for metal ions; e.g.,
[GH] + M2+ ^ [GM M2+] +H*'
GH = polymeric organic group; e.g., [sAlOH]n, [sFeOH]B, etc.
HUMIC AND RELATED SUBSTANCES
Transformation of biogenic substances within the soil; water and sediment environment
leads to humic substances which represent a significant fraction of the bulk of organic matter in
water and soil.
Humic substances include humic acid (HA), soluble in alkaline solution but precipitated by
acidification, and fulvic acid (FA), soluble over the entire pH range, and are made up of
phenolic and benzene carboxylic acids. HA and FA constitute weak acids which tend to form
complexes with metal ions:
+ ivr
OH
'o
M
FA, HA and humate coated with particles have been suggested as the principal complexing
agents, and FA is the dominant organic compound in natural waters.
Recently, considerable progress has been made in attempting to characterize quantitatively
the metal ion binding properties of FA It was shown that the stability of FA-metal complexes is
similar in magnitude to that of the corresponding complexes of its monomeric units, believed to
contain functional groups similar to those in FA; e.g., FA-Fe complexes.
Humic substances also have a strong tendency to become adsorbed on hydrous oxides, clays
and other surfaces. This adsorption may be interpreted in terms of surface complexation model:
R-C + HO-Fe==
X0~
R-C + OH'
xO-FeS=
HO-Fe* represents the surface of colloidal iron oxides (similarly, HO-Als represents hydrous
Al oxides).
136
-------
Pb, Cu and humic substances complex with high stability and are immobilizable in clay
minerals, soils and forests. Their stability relates to pH, concentration of humic substances (HS)
and (M)/(HS). Tables 4(a) and 4(b) give stability constants of HA-Metal complexes.
Effect of acidification* on mobility of metals and stability constants of metal complexes of
humic substances, chemical and mineralogical nature of particles in natural aquatic systems
including clays, HA, FA, semipolymerized oxyhydroxides of Al, Fe, Mn, Si have been studied.
Metal-HA(FA)-clay composites may be of central importance to natural transport processes of
metal ions. The nature of interactions of metal ions and complexes with the particle surface,
thermodynamic investigations of complexation occurring at interfaces, and model systems of
ternary complexes with organic and inorganic ligands and simultaneous interaction of the ligands
were studied.
Kinetic aspects in heterogeneous systems: Rate of dissolution of metal complexes under
environmental conditions at which equilibrium was attained were determined.
Table 4(a). Stability constants (log k) of metal-humic acid complexes.
Source
Peat
Lake water
River water
Songhua River (China)*
Ji River*
Estuarine
Sediment (Estuarine)
Songhua River*
Ji River*
Soil
Sample No.
FA
HA
FP
FPI
FP (Irish)
FC
CFI-1
CFI-2
CFI-3
BAL
DEE
CON
HA
FA
FA
ET
ET
ET-3
ET-4
ET-5
ET-6
FA
HA
FA
HA
FA
FA
FA
Ca
3.65
3.95
3.13
4.07
3.56
3.65
3.27
4.65
3.40
Mg.
3.81
4.00
3.67
3.74
3.26
3.5
3.41
4.09
3.92
2.20
Cu
7.85
8.29
8.40
8.27
8.30
8.28
9.83
8.42
9.35
930
9.48
9.59
8.89
10.21
1137
10.43
10.14
9.91
4.00
Zn
4.83
5.14
5.05
5.13
5.25
5.36
5.41
3.14
2.68
5.89
4.89
2.76
3.13
3.70
3.04
Cd
4.57
4.57
4.70
3.01
2.54
2.66
3.00
5.08
Hg
183
19.4
20.1
18.4
193
19.7
21.9
16.74
16.02
1638
18.1
213
1651
16.39
16.38
16.41
5.20
(Mantoura 1978, *Peng and Wang 1981)
137
-------
Table 4(b). Relative stability of metal complexes with humic substances.
Sample, Source
Simulation
experiments
Sewage sludge
Natural water
Lake water
Fresh water
seawater
Aquatic humic
substances
Soils
Metal Stability
Pb > Cu > Cd > Zn
Fe, Cu > Zn > Ni > Cd
Cu > Pb > Zn > Cd
Pb > Cu > Cd
Hg > Cu > Ni > Zn >
Co, Mn s Cd > Ca > Mg
UO2(n) > Cu > Zn > Ni
Pb>Ca
pH3: Fe(IT) > Hg > Al > Cu >
Ni > Pb, Cd > Zn > Mn(U)
pHS: Ni, Co > Cu, Zn, Mn(IT) >
Ca > Cd > Mg
Methodology*
Dialysis, AAS, pH
Gel filtration
AAS.ISE
ISE
Ion exchange
ISE
Reference
Can. J. Chem. 54:
2600 (1976)
J. Environ. Qual.
7:181 (1978)
Environ. Geology
2:257 (1978)
Nature 256:399
(1975)
Soil Sci. 119:98
(1975)
Soil Sci. 109:333
(1970)
•AAS, Atomic Absorption Spectroscopy; ISE: Ion Selective Electrode.
METAL BINDING IN HETEROGENEOUS SYSTEMS
Metals in natural waters are often associated with suspended particulate matter, physically
or chemically bound (Figure 2). This association affects the mobility or transport of metals,
Aqueous phase
Hydrated metal ions,
HydraKO compounds,
Inorganic complexes,
Organic complexes
Oxyanions
Aquatic solids
Clays
Silicates
Carbonates
Detrital organic matter,
Fe(Mn) hydrous cocides,
Sulfides
Bacteria
Algae
I
Physico-chemical conditions:
pH, Eh, ion concentration, redcoc potential
Figure 2. Physico-chemical interactions and metal species in aquatic systems.
which is largely responsible for removal of metallic contaminants from aquatic systems.
Dissolved organic matter can inhibit the adsorption of metals on inorganic solids through
formation of metal-organic complexes. Stability of colloidal particles and association with metals
138
-------
can thus be enhanced, and transport of metals in aqueous systems is facilitated. The mean
lifetime of a metal complex in the environment is regulated by the rate of complex formation,
dissociation and exchange with the solid phase. Absorption of organics and metal-organic
complex formation at solid-solution interfaces are important in the transport of metals in the
environment. Complexation of metals by soluble organic chelators can also increase the
biological availability of trace metals to aqueous organisms, especially when org./M is high, but
the relationship between complexation capacity and toxicity of Cu, Cd to algae is found to be
inversed.
MOBILITY AND BIOAVAILABILITY OF SEDIMENT-BOUND METALS.
Transfer of heavy metals from contaminated sediments to aquatic organisms involving
mobility and bioavailability of sediment-bound heavy metals are important in the processes of
metal recovery of dredged materials, ore tailings from mining activities and sewage sludge
treatments. Absorption/desorption, suspension, resuspension, transport and disposal of dredged
materials and tailings in aquatic systems are effected by changing of physico-chemical conditions,
and metal binding (or complexation) is the important process to enhance the mobility of the
metals and to effect metal availability.
The sediment-solution distribution ratio reflects the strength of metal binding to sediment,
and that metal's uptake was controlled by the binding intensity. Sediment-water exchange must'
be considered as a predictor of metal bioavailability from the sediment. Using weak acidic
chelators to extract metals from sediments in natural systems (e.g., DTPA, diethylene-
triamminepentaacetic acid) may improve correlation with metal bioavailability. The amount of
Cu-DTPA complex reflects available Cu in plant-available soils. Geochemical characterization
of sediment shows some coincidence with bioavailability of some metals. However, a lot of -
mechanistic questions concerning control of metal bioavailability from sediments are beginning
to develop better models of interaction in sediments, together with better analytical methods
and improved biological understanding. A surface complexation model for trace metal toxicity
to fish was suggested (Pagenkopf 1983).
A multi-chamber (Figure 3) was designed to study metal sorption/desorption, metal
binding on model sediment compartment with different surface binding interaction. It competes
for metals in solution with important parameters controlling the interaction of the dissolved and
relevant solid metal species in question.
Metal complexation together with sediment adsorption and redox reactions can be
studied; e.g., Cu on bentonites, Fe oxyhydrates and algae cells. Determination of metal sorption
on solid phases resulted in a significant enrichment on the algal cell walls, particularly for Cu,
Cd (Figure 4). It was found that sorption depends not only on ion exchange but also on
complexing reactions which lead to relatively stable surface binding. The dominant role of
organic substrates in the binding of metals such as Cu, Cd is of particular relevance for the
transfer of these elements into biological systems. Even relatively small percentages of organic
substrates, if involved in metabolic processes, may constitute a pathway by which metals are
transported within the food chain. This study related to the effect of pH and redox changes and
the influence of organic chelators on the distribution of metals among the individual
compartments. It is useful for predicting intensities and reversibilities of metals association
under different physicochemical conditions.
139
-------
redox
magnetic stirrer
model
component
membrane
Figure 3. Schematic view of the multi-chamber device. This experimental device consists
of a central chamber connected with six external chambers and separated by
0.451*, diameter membrane. Algae cell walls, clay minerals and metals were
used as model components.
Remobilization or desorption of heavy metals from sedimentary phases is strongly
dependent on the type of chemical bonding and on the environmental conditions. However, the
presence of dissolved organic ligands in pore water of the sediment suspension to complex the
metals and its possible sorption on algae surfaces should be considered.
MOBILITY OF METALS FROM SOLID MATERIALS
Mobility/bioavailability of metals needs to be assessed in sludge treatment. Metal binding
characteristics of raw, activated and anaerobically digested sewage sludges need to be studied.
Quantitative assessment of metal sludge characteristics may be obtained by the determination of
the conditional stability constants of the complexes formed and the concentration of metal
binding sites in a given sample for effective regulation of its complexation capacity. It was found
that the stability of the complexes formed in each sludge type followed the general order:
Cu > Pb > Cd. Both Cu and Pb formed more stable complexes in the treated sludges than in
140
-------
Uptake of Cd, Zn. Cu. and Mn from sediments in Brachiomonas suamanna a; ; .5% sa::r:;Sy *
Uptake o( Cd, Zn. Cu, and Mn from sediments in cell walls of ^ccncdcsmtis quadricauda at
1.5% salinity
Uptake of Cd. Cu. and Zn from sediments in Cordylophora caspia ai i 5% salim
Rgure 4. Uptake of metals from sediments in algae.
141
-------
the raw sludge. The capacities of the raw and digested sludges to complex all metals decrease
with increase of particle size. The reverse was true for the activated sludge. Digestion of sludge
plays an important role in controlling the mobility and environmental dispersion of these
metals. Recent studies include Cd complexation in sewage sludge, sewage effluent, natural
water samples and coal humic materials. Results show that Cd binding to FA increases as pH
increases, and labile complexes were formed.
SURFACE COMPLEXATION
In natural waters, the dispersed phase consists predominantly of inorganic colloids and
organic colloidal matter of detrital origin as well as living microorganisms. Because of the large
extent of interfaces available, metal oxides and clays are of relevance in regulation of water
composition. Mechanistic studies on the interfaces of water with naturally occurring solids have
been studied in the past few decades. Fundamentals for the surface complexation concept are
developed from the equilibrium analysis bearing the potential of unifying thermodynamics and
binding principles supported by spectroscopic studies. The colloids are aggregates of defined
chemical structure; the primary charge of the surface arises from the ionization of complex
inorganic groups present on the surface of the particles, and the destabilization of colloids is due
to such interactions as complexation formation and proton transfer. Ligands charged with
coordinating anions lead to a release of OH" from the surface.
Adsorption of H+ and OH' is based on the protonation and deprotonation of surface OH"
S-OH+H* *• S-
S-O- + H (+H20)
Adsorption of metal ions:
S-OH + M2+ ** S-
2 S-OH + M2* ^ (S-O)2 M + 2 H*
S-OH + M2+ + $L -* S-OML'1)* + H*
Ligand exchange:
S-OH + L «* SL* + OH~ (ternary surface complex)
S-OH + L + M2* ** S-L-M<<+1>+ + OH-
Surface complexation permits us to handle adsorption equilibrium as in solution.
Conditional stability constants of the surface species and the intrinsic surface acidity constants
can be evaluated by simple linear extrapolation of the experimentally available conditional
constants.
142
-------
Better insight into the mechanism of surface complexation has been gained by combining
thermodynamic information with that on the structure of both binary and ternary surface
complexes obtained by molecular spectroscopy and from kinetic studies on the mechanism of
absorption rate constants. For example: molecular structure in surface complex of linear
alkylbenzene sulfonate (LAS, a detergent) association by EPR spectra of Cu(H) surface complex
(Motschi and McEvoy 1985).
Metal ions adsorbed on functionalized cellulose derivatives show that:
Cu(H) forms a cationic inner sphere complex capable of forming an ion pair with LAS and
forms an anionic outer sphere complex with a surface PO^, which prevents additional
interaction with the anionic surface and hence shows no influence upon its adsorption
(FigureS). K
300 -r
250
200 --
algal cell walls
Cu [jjg/gJ-150.--
100
so --
in 100 g sludge: 32.200 ug
treatment with seawater
mobilized copper
dissolved: 156
5 9 3 g 0.5 g 0.2 g 0.5 g
Figure 5. Mobilization of Cu from sediments and transfer to different
model substrates after treatment with seawater.
143
-------
Models have been developed to describe the kinetics of dissolution reactions of Al
(hydrous) oxide by the surface complexation concept, the surface structure and multisurface
speciation (Figure 6).
Surface
groups ^/
Doubly coordinated I /
Singly coordinated
Reaction-Complex
Large circle: OH group
Small circle: Al ion
Surface group: singly or doubly coordinated
Reaction complex: an Al center near the surface surrounded by ligands
Dissolution: (1) an Al four-fold bound with the solid is released,
(2) release of Al ion, which is only doubly linked
with the solid.
Figure 6. Two dimensional representation of the structure of Al hydrous oxides
(Hiemstra and Van Riemsdijk 1990).
i
For interpretation of metal complexation in heterogeneous systems, although dynamic
considerations are very important in natural environmetal processes, our present knowledge of
geochemistry of aquatic systems is based predominantly on the concept of equilibrium -
thermodynamics (classical and statistical), coordination chemistry and non-relativistic physics.
REFERENCES
Ahlf, W., W. Calmano, and U. Forstner. 1986. The effects of sediment-bound heavy metals on
algae and the importance of salinity. Pages 319-324 In: Sediment and water interaction.
P.O. Sly (Ed.) Springer-Verlag, New York, NY.
Andersson, M. 1988. Toxicity and tolerance of Al in vascular plant Water, Air, Soil Pollut
39:439-462.
144
-------
Allen, H.E. 1983. Correlation of Cu distribution in freshwater sediment system to bioavailability.
Bull. Environ. Contain. Toxicol. 30:37-43.
Allen, H.E., R.H. Hall, and T.D. Brisbin. 1980. Metal speciation. Effect on aquatic toricity.
Environ. Sci. Technol. 14:441.
Bernhard, M., RE. Brinckman, and P.S. Sadler. (Eds). 1986. The importance of chemical
speciation in environmental processes. Springer-Verlag, Berlin.
Bourg, A.C.M. 1988. Metals in aquatic and terrestrial systems: Sorption, speciation and
mobilization. Pages 3-30 In: Chemistry and biology of solid wastes. W. Salomons and U.
Forstner (Eds). Springer-Verlag, Berlin.
Buffle, J. 1980. A critical comparison of studies of complex formation between Cu(H) and fulvic
substances of natural waters. Anal. Chim. Acta 118:29-44.
Buffle, J., and R.S. Altmann. 1985. Interpretation of metal complexation by heterogeneous
complexants. Pages 351-384 In: Aquatic surface chemistry. W. Stumm (Ed.). John
Wiley, New York, NY.
Burrows, M.D. 1977. Aquatic aluminum: Chemistry, toxicology and environmental prevalence.
CRC Crit. Rev. Environ. Control 7(2): 167.
Calmano, W., W. Ahlf, and U. Forstner. 1988. Study of metal sorption/desorption processes on
competing sediment components with a multichamber device. Environ. Geol. Water Sci.
Daum, K.A., and L.W. Newland. 1982. Complexing effects on behavior of some metals. Pages
129-137 In: The handbook of environmental chemistry. Vol.2, Part B. Reaction and
processes. O. Hutzinger (Ed.). Springer-Verlag, Berlin.
Florence, T.M., and G.E. Batley. 1980. Chemical speciation in natural waters. Crit. Rev. Anal.
Chem. 9:219-296.
Florence, T.M. 1982. Effects of model compounds on the removal of Zn, Cd, from water by
chelation. Anal. Chim. Acta 141:73-94.
Hem, J.D., and C.E. Roberson. 1967, 1972. Form and stability of aluminum hydroxide complexes
in dilute solution; Effect of ageing in aluminum hydroxide complexes in dilute solutions.
USGS Water Suppl. Paper No. 1827 A.D.
Irgolic, J.K., and A.E. Martell. 1985. Environmental inorganic chemistry workshop summary.
Pages 1-59 In: Environmental inorganic chemistry. Irgolic, KJ. and A.E. Martell (Eds.).
VCH Publishers.
Kang, D.M., and S.W. Pang. 1987. Leaching pattern of Al from soil collected from southwestern
China by acid rain. Acta Scient. Circum. 7(2):230 [In Chinese].
Lake, D.L., P.W.W. Kirk, and J.N. Lester. 1980. Heavy metal solids association in sewage sludge.
Water Res. 23(3):285-91.
145
-------
Li, C.S., and C.I. Liu. 1982. Absorption of mercury in fresh water. Environ. Chem. 1:301 [In
Chinese].
Lin, Y.H., D.M. Kang, and C.I. Liu. 1983. Speciation and distribution of mercury in Ji Yan River
sediment. Environ. Chem. 2:10 [In Chinese].
Liu, C.I., and H.X. Tang. 1985. Chemical studies of aquatic pollution by heavy metals in China.
Pages 359-371 In: Environmental inorganic chemistry. Irgolic, KJ. and A.E. Martell
(Eds.). VCH Publishers.
Lu, Z.P., M.Z. Mao, and J.X. Wei. 1988. Determination of the chemical forms of toxic elements
in aquatic samples from China. Pages 373-391 In: Inorganic environmental chemistry.
Irgolic, J.K. and A.E. Martell (Eds.). VCH Publishers.
Luan, Z.,. 1987. Aquatic chemical reaction of Al and its chemical speciation. Environ. Chem
6(1):64 [Tn Chinese].
Luan, Z.K., and H.X. Tang. 1990. Chemical species distribution and transformation in
polyaluminum chemical solution. J. Environ. Sci.
Mantoura, R.F.C., A. Dickson, and J.P. Riley. 1978. The complexation of metals with humic
materials in natural waters. Estuar. Coast. Mar. Sci. 6:387-408.
Motschi, H., and J. McEvoy. 1985. Aspects of the molecular structure in surface complexes:
Spectroscopic investigation. Pages 111-124 In: Aquatic surface chemistry. W. Stumm
(Ed.). John Wiley, New York, NY.
Mouvet, C, and A.C.M. Bourg. 1983. Speciation of Cu, Pb, Ni, Zn. Water Res. 17:641-649.
Neville, CM., and P.G.C. Campell. 1988. Possible mechanisms of Al toxicity. Water Air Soil
PolluL 42:311-347.
O'Donnel, J.R., B.M. Kaplan, and H.E. Allen. 1985. Bioavailability of trace metals in natural
waters. Pages 485-501 In: Aquatic toxicology and hazard assessments. 7th Symposium
ASTM STP 854. Cardwell, R.D., R. Purdy and R.C. Bahner (Eds.).
Pagenkopf, O.K. 1983. GUI surface interaction model for trace-metal toxicity to fishes: role of
complexation, pH, and water hardness. Environ. Sci. Technol. 17:342-347.
Pang, S.W., and D.M. Kang. 1986. Leaching of active Al from soil and distribution of Al species
by chemical extraction. Environ. Chem. 5(3):68 [In Chinese].
Parthasarathy, H., and J. Buffle. 1985. Study of polymeric A1(HI)-OH solution. Application in
water treatment Properties of the polymer. Water Res. 1:25.
Peng, A., and W.H. Wang. 1981. Extraction and characterization of humic acids from Ji Yun
River. Acta Scient. Circum. 1:126 [In Chinese].
Peng, A., and ZJ. Wang. 1985. Mercury in river sediments. Pages 393-400 In: Inorganic
environmental chemistry. Irgolic, J.K. and A.E. Martell (Eds.). VCH Publishers.
146
-------
Peter, G. 1983. Speciation of Al in acidic fresh water. Anal. Chem. 55:2246.
Salomons W., and U. Forstner (Eds.). 1988. Chemistry and biology of solid wastes. Springer-
Verlag, Berlin. . K 6
Schindler, P.W., and W. Stumm. 1985. The surface chemistry of oxide, hydroxides and oxide
PagCS 83"1°7 IH: A^uatic surface chemistry. W. Stumm (Ed.). John Wiley, New
Schnitzer, M. and S.U. Khan. 1972. Humic substance in environment Marcel Dekker, New
X OF Kj XN X •
Sigg, L 1985. Surface chemistry aspects of distribution and fate of metal ions in lakes Pages
319-346 In-. Aquatic surface chemistry. W. Stumm (Ed.), John Wiley, New York, NY.
Stumm, W. and JJ. Morgan. 1981. Aquatic Chemistry. 2nd Ed. John Wiley, New York, NY.
Stumm, W. (Ed.). 1985. Aquatic surface chemistry. John Wiley, New York, NY.
Tang, H.X., H.B. Xue, M.Z. Mao, and Z.K. Luan. 1989. The chemical stability of heavy metals
in natural water system. Environ. Sci. 1(1):55. .
Tyler, G. 1989. Effect of heavy metals in forest ecosystems. Water, Air, Soil Pollut 47f3/4V
Special issue. v '"
Vuceta, J., and JJ. Morgan. 1978. Chemical modelling of trace metals in fresh water Role of
complexation and adsorption. Environ. Sci. Technol. 12:1302-1309.
Van den Berg, C.M.G., and J.R. Kramer. 1979. Determination of complexing capacities of
hgands in nature waters and conditional stabilities of the Cu complexes by means of MnO
Anal. Chim. Acta 106:113.
147.
-------
-------
PREDICTION OF METAL CONTAMINANT EXPOSURE
IN NATURAL WATERS USING GEOCHEMICAL
EQUILIBRIUM MODELING
by
Nicholas T. Loux1 and David S. Brown1
INTRODUCTION
The Environmental Research Laboratory-Athens is expanding the MINTEQA2
geochemical speciation model (Brown and Allison 1987, Allison et al. 1990) to incorporate
recent advances in the understanding of those processes influencing the behavior of metal
contaminants in the environment. MINTEQA2, when coupled with a mass transport model, has
found a useful application in estimating potential human drinking water exposure to metal
contaminants in aquifers underlying land disposal facilities (Brown et al. 1986, Mulkey et al
1989). More recently, MINTEQA2 has been used to assess the potential immobilization of
metal contaminants in freshwater sediments containing an anoxic (sulfidic) underlayer (Loux et
al. 1990).
With the current understanding of those processes governing the existing and potential
bioavailability of metal contaminants, it is clear that chemical speciation as described by
thermodynamic equilibrium theory will play an increasing role ^quantifying the risks associated
with potential metal contaminant bioavailability. The focus of the present paper is to present
the findings from recent innovations in the ability to describe the behavior of metals in the
aquatic environment and to identify key research requirements pertinent to quantifying potential
bioavailability of metals.
EQUILIBRIUM PROCESSES INFLUENCING METALS EXPOSURE
TO THE BIOLOGICAL COMMUNITY
Figure 1 is a generic illustration of the reservoirs of a metal contaminant in an aquatic
system. The central species, Me2+, represents the free dissolved metal ion that is considered by
the scientific community (Luoma 1983, Sanders et al. 1983, Jenne et al. 1986, Morel 1988,
Mount and Anderson-Carnahan 1988, DiToro et al. 1988, Morrison 1989, Mason and Jenkins
1990) either to be directly bioavailable or to be strongly correlated with the most bioavailable
form of the metal ion. Because the experimental measurement of a free dissolved ion
concentration (or chemical activity) in environmental samples is a difficult task more in the
realm of research than of standard practice (Florence 1989), the sum of the soluble species in
'Environmental Research Laboratory, U.S. Environmental Protection Agency, Athens, Georgia,
149
-------
r
Figure 1 (the free ion in addition to the soluble inorganic and organic complexes)-represents a
convenient means of interpreting experimentally derived total dissolved metals concentrations
(via water sample pretreajment with filtration or centrifugation) in terms of thermodynamic
equilibrium model predictions. The lower two compartments in Figure 1 illustrate the types of
chemical processes that may lead to the partitioning of a dissolved metal to natural sediments.
Figure 1. Compartmentalization of a generic metal contaminant in a natural
aquatic system. Water column and porewater exposure to the
biological community may (arguably) be the dominant exposure
pathway to the lower trophic levels.
150
-------
The soluble and participate metal species depicted in Figure 1 can be viewed as existing in
simultaneous equilibrium with a large number of the compartments in the figure. This
equilibrium, in turn, is best described using conventional thermodynamics.
According to thermodynamic equilibrium theory, the speciation of a metal can be
estimated using a sequence of chemical equilibrium reactions described by the following generic
type of equilibrium constant:
b~B + c-C + d-D... <--> e-E + f-F + g-G...
(1)
K =
B ac aD...
(2)
where the subscripted a's represent chemical activities of both the reactants and the products
involved in the reaction. The superscripts designate the stoichiometry (number of individual
species) involved in the overall reaction. The chemical thermodynamic activities in Equation 2
can be related to experimental concentrations using the following type of equality:
*Y
(3)
where the bracketed species represents an actual molar concentration of a given species and the
Y term represents an "activity coefficient" that can be estimated for ionizable species using
mathematical relationships such as the Davies or extended Debye-Huckel equations (Brown and
Allison 1987) that relate activity coefficients to ion valence and soluble dissolved salt content
(ionic strength; IS).
Progress in understanding the chemical processes leading to equilibrium among the
various compartments in Figure 1 is often a case of recognizing that these processes represent
special cases of Equations 2 and 3. The remainder of this paper will illustrate current advances
and limitations in mathematically describing these chemical reactions.
COMPLEXATION OF METALS IN NATURAL WATER
The reactions governing the soluble inorganic complexation processes in Figure 1 are
among the best understood of all reactions. Generally, compendia of the formation constants
describing these reactions are widely available (e.g., Smith and Martell 1976) and are often built-
in to the databases of models such as MINTEQA2. For a more detailed discussion of the
significance and/or rigor of inorganic complexation reactions in natural waters, the reader is
referred to Eberle (1989).
151
-------
Complexation reactions between metal ions and organic ligands generally involve both
anthropogenic and natural organic ligands. Similar to the situation with inorganic complexation
reactions, compendia of formation constants with anthropogenic ligands are widely available. In
terms of the significance of complexation to metals bioavailability, the Toxicity Identification
Evaluation Program (Mount and Anderson-Carnahan, 1988) uses the addition of the synthetic
ligand EDTA (ethylenediaminetetraacetic acid) to determine whether a toxicant in an industrial
effluent is a cationic metal; specifically, if toxicity decreases upon addition of EDTA, then a
presumption of toxicity due to metals is assumed.
Because natural dissolved organic matter does not exist as a series of unique, identifiable
monomers, a continuous, multiligand model has been developed recently to describe the
complexation of metals with naturally occurring organic matter (Dobbs et al. 1989). Based on
an assumed Gaussian distribution of carboxylate-type binding sites of variable metal affinity, this
model uses a spectrum of proton exchange reactions of the type:
Ka
[JfflfiLjj
(4)
The model implementation in MDSfTEQA2 then sums the concentration of metal ions bound to
the sites of variable concentration and energy.
Figures 2a and 2b illustrate an application of this model in describing Pb complexation
with naturally occurring organic matter. These diagrams were constructed to interpret chemical
and lexicological data obtained from water samples collected at the Kassouf-Kimmerling
National Priority Listing (Superfund) site in Florida (USEPA 1989, Loux et al. 1990). Total (i.e.,
unfiltered) water samples were analyzed to identify common elements and contaminants
(especially Pb).
Figure 2a illustrates the results from simulations when complexation of Pb with naturally
occurring dissolved organic carbon is ignored. Three points can be discerned from Figure 2a: 1)
the majority of Pb in the samples is predicted to be soluble; 2) especially at lower pH values, the
majority of dissolved Pb is predicted to exist as the "free" (bioavailable) ion; and 3) at higher pH
conditions, the fraction of dissolved Pb existing as inorganic complexes increases. Based purely
on Figure 2a, Pb in the water column could possibly be expected to be present at toxic
concentrations.
Figure 2b was used to interpret any possible toxicologieal effects observed with these
samples. In these MINTEQA2 simulations, potential Pb complexation with naturally occurring
•dissolved organic carbon was included. In contrast to Figure 2a, the major dissolved species of
Pb shifts from the free ion (Figure 2a) to a soluble Pb-organic matter complex, although the
total quantity of dissolved Pb remains roughly equivalent at all pH conditions. Hence, the
potential bioavailability (and toxicity) of the free Pb2* ion in the water column may be
decreased.
152
-------
Q.
5
100
90
80
70
60
50
40
30
20
10
0
(a) Pb Complexation with Inorganic Ligands Only
tree •btoavall-
abte-Pb
Pb-lnorgante
complexes
Pb-adsorbed on
suspended Fe
—I—
7
8
.a
Q_
15
5
100-
90-
80-
70-
60-
50-
40-
30-
20-
10-
o-
c
(b) Complexation with Natural Dissolved Organic Matter
Ptwvwnplovort ,*Hh
natural org. matter ~~^^
N.
\y
\
\
\
Pb-lnorganfc X
(XKnptexes ^__^ /
^>» /
/
free •bioavail- S Pt>adsorbed on
abte'Pb ~ — __..---' suspendedFe(
I T 1 , , , -,_
'6789
pH
Figure 2. MINTEQA2 predictions of Pb speciation derived from data obtained
at the Kassouf-Kimmerling Superfund site. Total Pb concentration =
0.099 mg/L. (2a) In the absence of Complexation with dissolved
organic carbon, free "bioavailable" Pb is the dominant species in the
pH range 6-7. (2b) With organic carbon Complexation, free Pb is less
than 10% of the total at all conditions. (Loux et al. 1990).
153
-------
PARTITIONING: ADSORPTION OF METAL IONS ON NATURAL SURFACES
Virtually all of the "hydrous" oxide surfaces encountered in environmental sediments
(Figure 1) experience "surface complexation" reactions "of the following type:
=FeOH + Me2* <--> =FeOMe+ + H+
(5)
Reaction 5 describes the potential adsorption of a metal on an amorphous iron oxide surface
site. Similar expressions for manganese oxide, aluminum oxide, silicon oxide, and "edge"
aluminosilicate (clay) sites also may be developed. A formation constant describing this reaction
is given by Dzombak (1986) and Dzombak and Morel (1990):
[=FEOMe+][H+]EXP(-e-Psilkl)
[=FeOH\[Mez+lEXP(-2'e-Psilttf)
(6)
(the bracketed species in Equation 6 represent concentrations; EXP = exponential to the base
of the natural logarithm, e = charge of the electron, Psi = the electrostatic potential in the
plane where the adsorbing ion resides, k = Boltzmann constant, and T = the absolute
temperature). Although beyond the scope of the present document, the exponential terms in
Equation 6 represent "corrections" to the mobile ion (H+ and Me2+) concentrations (activities)
resulting from moving ions between a charged surface and an uncharged bulk solution. From
preliminary work in the literature and at this laboratory, these mobile ion electrostatic
corrections tend to be of lesser significance in describing metal partitioning than are other
variables such as pH.
Figure 3 illustrates an application of this model to describe Pb, Zn and Ni partitioning
behavior on a sandy aquifer material. MINTEQA2 model predictions were made by
incorporating the MIT-Diffuse Layer Model (Dzombak 1986, Dzombak and Morel 1990) for
sedimentary amorphous iron oxide (Loux et al. 1989). Several points were made in the study:
(1) model predictions were not "fitted" to the data; model predictions were based on the
sediment amorphous iron content and "intrinsic" adsorption constants found in the thesis by
Dzombak (1986); (2) the predicted "adsorption envelope" for these elements in the intermediate
pH range better describes their partitioning behavior than did precipitation mechanisms; and (3)
the model overestimates of dissolved metal concentration at lower pH conditions for all three
elements likely result from ignoring other adsorptively significant phases in the sediment.
PARTITIONING: PRECIPITATION OF METAL IONS
Historically, the precipitation behavior of pure phase solids is described using a special
case of Equation 2. By convention, the chemical activity of the solid phase is defined as unity
and Equation 2 reduces to:
v bed (7)
K = aB ac aD...
154
-------
100
Figure 3. Comparison of MIT Diffuse Layer Model adsorption predictions (lines)
with experimental partitioning data (blocks) for Pb, Zn and Ni on a
sandy Wisconsin aquifer material. Overpredicted soluble metal
concentrations in the low pH ranges are believed to result from
ignoring all sediment sorptive phases other than amorphous iron oxide
(Loux et al. 1989).
Pure phase precipitation reactions are well understood in general, and hence their application in
environmental systems requires only a confirmation that these reactions occur within the time
frame under investigation in actual sedimentary deposits.
A special case in the application of precipitation includes the precipitation of highly
amorphous, high-specific surface area solids. It is generally known that given sufficient time, an
initial ensemble of small crystals will eventually be sacrificed to form one large crystal. The
process governing this phenomenon is the solubility product sensitivity to crystalline specific
surface area (Stumm and Morgan 1981). In a practical sense, this can translate into a
potentially increased solubilization by an order of magnitude or more (CdCO3--Lindsay 1979,
Loux et al. 1989; BaSO4-Fruchter et al. 1988). Specifically, fresh, amorphous pure phase
minerals may be more soluble than their highly crystalline counterparts.
155
-------
A second special case includes the possibility of solid-solution formation. Generally, a
solid-solution occurs when one of the components in a crystalline solid is replaced by an
analogous ion; e.g., replacement of Ca2+ by Ba2+ or other metals in crystalline CaCO3 (Davis et
al. 1987), replacement of Fe3+ by Cr3+ in Fe(OH)3 (Rai and Zachara 1986). The thermodynamic
description of solid-solution formation differs from pure phase precipitation reactions in that the
solubility product(s) generated are sensitive to the mole fraction(s) of the replacing ions in the
crystalline solid. Solid-solution processes are much less documented in the literature than are
pure phase precipitation reactions. Because the actual mechanisms limiting soluble metal
concentrations with environmental sediments are seldom experimentally verified, and because
both adsorptive and solid-solution processes may depress soluble metal concentrations below
limits predicted by pure phase precipitates, considerable debate exists among geochemists and
soil scientists as to which mechanism is most significant with environmental systems. From work
conducted at the U.S. Environmental Protection Agency (EPA) Athens Environmental Research
Laboratory, consideration of both mechanisms will likely be required to improve accuracy in
predicting soluble metals exposure to the biological community.
PARTITIONING: METAL SULFIDES AND EQUILIBRIUM-PARTITIONING
SEDIMENT QUALITY CRITERIA
The EPA is currently investigating various methods for use in developing sediment quality
criteria in order to manage contaminated sediments in the national waterways. From sediment
bioassay experiments conducted at EPA laboratories in Duluth, Minnesota and Narragansett,
Rhode Island and at Manhattan College, there is evidence to support the contention that
sedimentary sulfides may limit the bioavailability of metal contaminants. A recent geochemical
assessment of sulfide metals immobilization (Loux et al. 1990) addressed two questions:
(1) What metals currently regulated by EPA would be susceptible to a sulfide immobilization
process? (2) Because soluble sulfides are themselves potentially toxic, how may they inhibit
metals bioavailability when the benthic biological community will presumably avoid contact with
sulfides?
The first question was examined within the context of conventional metal sulfide
precipitation:
Me
2*
<--> MeS.
(8)
where the underline indicates a solid phase and
Assuming excess dissolved sulfides exist in sedimentary porewaters, MINTEQA2 can be used to
estimate the resulting maximum porewater metal concentrations.
Figure 4 compares MINTEQA2 predicted porewater concentrations of Cd2* and Se2' with
water quality criteria protective of freshwater aquatic life (Table 1). These simulations were
performed assuming a dissolved porewater bisulfide (HS-) concentration of 1 mg/L and a
sediment iron content of 6,400 mg/kg (Loux et al. 1990). Based on these simulations, sulfidic
porewater Pb2+ and Se2' concentrations are expected to remain below water quality criteria
values due to PbS and FeSe formation respectively.
156
-------
o
o
o
i
o
Q.
T3
O
jg
O
O>
o
0-r
-2
-4-
-6-
-8-
-10-
-12-
-14-
-16-
-18-
-20-
-22-
-24-
-26-
3
(a)
WQC.
Soluble Cd/*
concentration
Toxicity
Unexpected
(hardness
IS = 0.01 M T = 13 °C
sediment = 1.000 mg/Kg
CdS solubility control
—r-
9
11
0-
. -2-
c -4-
8 6_
1 -8-
| -10'-
o-12-
*-14-
--16-
co
o-18-
0.22-
-24-
-264
(b)
WQC (35 ug/L)
-«»>_^^
Soluble Se X
concentration
^^^^
— *"
Toxicrty
Unexpected
^ — —
,
IS = 0.01 M T = 13^ FeSe solubility control
pH = 9.0 [Se^ed.mem= 1,000 mg/kg
35 7 ' 9 ' 1
PH
Figure 4.
Comparison of MINTEQA2-predicted soluble Cd (a) and Se (b)
concentrations in a typical freshwater porewater with water quality
criteria protective of freshwater aquatic life. Selenium \& assumed to
exist as the selenide species. In the PH range 6.5 to 9.0, porewater
concentrations are predicted to be below criteria values. Selenium may
exist in other oxidation states and hence may exhibit higher
concentrations in sulfidic regions (Loux et'aL 1990).
157
-------
Table 1. Water quality criteria (WQC) protective of freshwater aquatic life for selected
inorganic contaminants currently regulated by EPA.
Metals & pH
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Thallium
Silver
Zinc
HjS
pH
WEQ (ng/L)
610
190
1,000
53
e(nteeu)]+l.M)
35
20
g(1.72pn(hirdneu)}
-------
Table 2. Status of sedimentary sulfide immobilization of metal contaminants
(Loux et al. 1990).
1V1CU11
Ag
As
Ba
Be
Cd
Cr
Cu
Hg
Ni
Pb
n
Sb
Se
Zh
=====
aumae imooiiization Comments ||
Yes
X
X
X
X
X
X
-
Sometimes
?
X
X
X
9
X
No I
X
X
X
X
Dissolved silver is rarely detected in environmental samples;
an AgzS phase can decrease porewater concentrations to
below WQC protective of aquatic life; silver also possesses a
strong affinity for other sedimentary phases.
Kinetic limitations, biomethylation and/or insufficiently low
redox conditions to form A%S3.
BaS may decompose in water; reducing conditions will
release Ba from sulfete- and possibly carbonate- limiting
solubility controls.
BeS may decompose in water under normal environmental
conditions.
CdS forms rapidly in water
Formation of a Cr2Q3 or Cr(OH)3 solubility control requires
reduction of the more mobile Cr(VI); this may impose kinetic
limitations.
CuS forms rapidly in water
HgS forms rapidly in watery biometnylated species may be
significant and not exhibit sulfide immobilization. -
NiS forms rapidly in water.
PbS forms rapidly in water.
TIzS is sufficiently soluble in water to exceed Water Quality
Criteria.
Sb sulfide complexation may lead to concentrations exceeding
Water Quality Criteria.
FeSe can form rapidly in water; alternatively, reduction of
selenite and selenate may be significantly slow; organic Se
compounds may be significant.
ZnS forms rapidly in water
Assuming that sulfide minerals limit sedimentary porewater concentrations to below water
quality criteria values, the question arises as to how these processes might limit bioavailability
This question may be partially answered by the fact that many metals are more soluble in
oxidized (sulfide-free) sedimentary porewaters, hence a permanent oxidic layer lying atop a
sulfidic underlayer may slowly be cleansed of its metal contaminant burden by diffusion of
dissolved metals into the underlying sulfidic region. The porewater exposure to metal
contaminants m the oxidic surface layer may decrease with time due to this diffusive mechanism
159
-------
CONCLUSIONS
The ability to estimate the potential bioavailability of metal contaminants via the water
column/porewater exposure route will require improvements in the present understanding of
those processes governing both the environmental distribution of the contaminant and the
mechanisms of toxicity and/or transport of soluble metal to the cell membrane and/or
intracellular regions. For example, sediments/suspended participates often represent the most
significant reservoir of contaminants that "buffers" (maintains) the concentration of soluble metal
at quasi-steady-state levels; hence, those mechanisms most significant in partitioning also will be
most significant in terms of long term bioassimilation. From the preceding discussions, one sees
that those equilibrium processes significant to metals bioavailability include (in order of
decreasing understanding): inorganic complexation > pure phase precipitation > natural
organic matter complexation > adsorption and solid-solution formation.
Those processes governing toxicity may include binding to biological membranes with
possible subsequent transport of metals to intracellular regions. In general, these processes are
much less well understood. For example, Morel (1988) proposed that existing EPA water quality
criteria be interpreted in terms of bioavailability mediated by initial metal complexation with
sites on the surface of biological membranes. Specifically, a membrane sorption mechanism
sensitive to both free metal contaminant ion activity and competition from other common ions in
water is wholly consistent with current EPA hardness-dependent water quality criteria (Table 1).
Furthermore, more recent work (Gardea-Torresdey et al. 1990, Majidi et al. 1990) with algal
cells suggests that the binding sites for selected cationic metals on algal cells may be carboxylate
in nature (R-COOH). Clearly, however, considerable work remains to be performed in
identifying all significant exposure pathways/mechanisms for each significant species at the
ecosystem level.
ACKNOWLEDGMENT
The authors have drawn heavily from previously published work, hence would like to
acknowledge the previous contributions from Mr. Jerry D. Allison, Ms. Claudia R. Chafin and
Dr. Sayed M. Hassan. The authors also would like to thank Dr. Leo V. Azarraga and co-
workers for assistance in providing a test version of the continuous distribution metal
complexation model. The authors gratefully acknowledge the editorial and technical comments
from Mr. Robert C. Ryans, Dr. Arthur W. Garrison and Dr. John E. Rogers.
REFERENCES
Allison, J. D., D. S. Brown, and K. J. Novo-Gradac. 1990. MINTEQA2/PRODEFA2, A
geochemical assessment model for environmental systems: Version 3.0 User's Manual.
United States Environmental Protection Agency, Athens, Georgia. EPA/600/3-91/021.117p.
Brown, D. S. and J. D. Allison. 1987. MINTEQAl, An equilibrium metal speciation model:
User's Manual EPA/600/3-87/012. United States Environmental Protection Agency. Athens,
Georgia. 103p.
160
-------
Brown, D. S., R. E. Carlton and L. A. Mulkey. 1986. Project Summary: Development of land
disposal decisions for metals using MINTEQ sensitivity analyses. EPA/600/S-86/030 United
States Environmental Protection Agency, Athens, Georgia. 4p. '
Davis, J. A., C. C. Fuller, and A. D. Cook. 1987. A model for trace metal sorption processes at
the caldte surface: adsorption of Cd*+ and subsequent solid-solution formation. Geochemica
et Cosmochimica Acta 51:1477-1487. '
c v - > K C°wan' D- J' Hansen' R R' Pa s- p- Pavlou, A. E. Steen
R. C. Swartz, N. A. Thomas and C. S. Zarba. 1988. Briefing report to the EPA Science '
Advisory Board on the equilibrium partitioning approach to generating sediment quality
criteria. (Unpublished Report)
Dobbs, J. C, W. Susetyo, L. A. Carreira and L. A Azarraga. 1989. Competitive binding of
hUmi° SUbStanCCS by lanthanide i°n probe spectroscopy. Analytical
Dzombak^ D. A. 1986. Toward a uniform model for the sorption of inorganic ions on hydrous
colloids. Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA.
Eberle, S. H. 1989. A correctness test of the computation of chemical speciation for the main
constituents of natural waters. Water Research, 23:1373-1382.
Florence, T M. 1989. Electrochemical techniques for trace element speciation in water In
Irace element speciation: methods and problems. Batley, G.E. (Ed.). CRC Press, Boca
Raton, FL.
Fruchter, J. S., D. Rai, J. M. Zachara, and R. L. Schmidt. 1988. Leachate chemistry at the
Montour fly ash test cell. Document Number EA-5922, Electric Power Research Institute
Palo Alto, CA. '
Gardea-Torresdey, J. L., M. K. Becker-Hapak, J. M. Hosea, and D. W. DarnalL 1990. Effect of
chemical modifications of algal carboxyl groups on metal ion binding. Environmental
Science and Technology 24:1372-1378.
Honeyman,B.D.andP.H.Santschi. 1988. Metals in aquatic systems. Environmental Science
and Technology 22:862-871.
Jenne, E A., D. M. DiToro, H. E. Allen, and C. S. Zarba. 1986. An activity-based model for
developing sediment criteria for metals. In Proceedings of the International Conference on
Chemicals m the Environment, Lisbon, July 1-3, 1986. Lester, J.N., R. Perry, and R.M.
Sterntt (Eds.).
Lindsay, W. L. 1979. Chemical equilibria in soils. John Wiley and Sons, New York, NY.
161
-------
Loux, N. T., D. S. Brown, C. R. Chafin, J. D. Allison and S. M. Hassan. 1989. Chemical
speciation and competitive cationic partitioning on a sandy aquifer material. Journal of
Chemical Speciation and Bioavailability 1:111-125.
Loux, N. T., C. T. Jafvert, J. D. Allison, S. M. Hassan and C. R. Chafin. 1990. A Geochemical
assessment of potential porewater exposure to EPA-regulated metal and ionizable organic
contaminants for use in developing equilibrium-partitioning sediment quality criteria. United
States Environmental Protection Agency, Athens, GA. (EPA internal report)
Luoma, S. N. 1983. Bioavailability of trace metals to aquatic organisms - A Review. Science of
the Total Environment 28:1-22.
Majidi, V. D., A. Laaude, and J. A. Holcombe. 1990. Investigation of the metal-algae binding
site with 113Cd nuclear magnetic resonance. Environmental Science and Technology 24:1309-
1324.
Mason, A. Z. and K. D. Jenkins. 1990. Effects of feeding on zinc and cadmium accumulation
by the polychaete: Neanthes arenaceodentata. Chemical Speciation and Bioavailability 2:33-
47.
Morel, F. M. M. 1988. Trace element speciation and aquatic toxicity. Society of Environmental
Toxicology and Chemistry, Ninth Annual Meeting, Washington, D.C., Nov. 13-17.
Morrison, G. M. P. 1989. Trace element speciation and its relationship to bioavailability and
toxicity in natural waters. In Trace element speciation: Methods and Problems. Batley, G.E.
(Ed.). CRC Press, Boca Raton, FL.
Mount, D. I. and L. Anderson-Carnahan. 1988. Methods for aquatic toxicity identification
evaluations: Phase I Toxicity characterization procedures. EPA-600/3-88/034. United States
Environmental Protection Agency, Duluth, MN. 68p.
Mulkey, L. A., A. S. Donigian, T. L. Allison and C. S. Rajn. 1989. Evaluation of source term
initial conditions for modeling leachate migration from landfills. Internal report submitted to
the EPA Office of Solid Waste and Emergency Response. U.S. Environmental Protection
Agency, Athens, GA. (Unpublished report).
Rai, D. and J. M. Zachara. 1986. Geochemical behavior of chromium species. EPRI Report
Number EA-4544. Electric Power Research Institute, Palo Alto, CA.
Sanders, B. M., K. D. Jenkins and W. G. Sunda. 1983. Free cupric ion activity in seawater,
effects on metallothionein and growth in crab larvae. Science 222:53-54.
Smith, R. M and A. E. Martell. 1976. Critical stability constants: Volume 4- Inorganic
complexes. Plenum Press, New York, NY.
Stumm, W. and J. J. Morgan. 1981. Aquatic Chemistry. 2nd Edition. John Wiley and Sons,
New York, NY.
162
-------
USEPA. 1980. Ambient water quality criteria for thallium. EPA 440/5-80-074 United State*
Environmental Protection Agency, Washington, D.C.
USEPA. 1986a. Quality criteria for water, 1986. EPA 440/5-86-001. United States
Environmental Protection Agency, Washington, D.C.
C?ldl?lium; water 1ualitv standards criteria summaries: A compilation of
ar- EPA 440/5-88/014 United ^ E-ronmenta, Projection Agency,
C?hr°miuTm'waterquality standards criteria summaries: A compilation of
aa '*• ^ 44°/5-88/°26- United States Environmental ProtLion 4ency,
' Water quality standards criteria summaries: A compilation of
a- EPA44°/6-88/027- United States Environmental P^eS^n Agency,
USE
EO/5 summaries: A compilation of state/federal
EPA 440/5-88/030. United States Environmental Protection Agency, Washington,
USE 188 ¥er?ury' water Vuality standards criteria summaries: A compilation of
EPA 44°/5-88/°05- United States Environmental Protection Agency,
USEPA. 1988h Zinc, water quality standards criteria summaries: A compilation of state/federal
o-uena. EPA 440/5-88/019. United States Environmental Protection Agency!
Site (Plant City, Florida): Wetland impact
o P , Protection Agency, Region IV, Environmental Services
ivision, Ecological Support Branch, Atlanta, GA. (Unpublished Report)
163
-------
-------
IMPACTS OF ACID MINE DRAINAGE ON WATER
QUALITY OF THE LO AN RIVER, POYANG LAKE AREA, PRC
by
Lin, Yu-Huan1
INTRODUCTION
In the sulfide copper mining process, when certain ore surface and associated strata, waste ore,
and covering stone are exposed to the oxidizing environment, acid mine drainage generates. The
drainage carries HjSC^ and a variety of soluble metals, including Fe, Mn, Al, Cu, Pb, Zn, Cd, Ni, As.
The impacts of high acidity and heavy metals to aquatic organisms have been well documented
(Ramsey and Branoon 1988, Baker and Schofield 1982, Moore and Ramamoorthy 1984).
Dexing Copper Mine, the largest opencast mine in China, is located in the southern part
of China, Jiangxi Province. Impacts of acid drainage on water quality of Lo An River and
Poyang Lake are being studied. It is emphasized that the concentration, species, and transport
of heavy metals in aquatic ecosystem of the Lo An River and Poyang Lake will be investigated,
and the potential risk on water quality assessed.
STUDY SITE DESCRIPTION
The study is being conducted on the Lo An River system in the northeastern part of
Jiangxi Province (Figure 1). The river forms at Huiyueshan Mountain (Latitude 29° 11'),
an elevation of 860 m, which drains 9616 km2 and 279 km long running into Poyang Lake
Poyang Lake
Haikou
Gukou
Zhongzhou
Xiangtun
Daicun
Sampling sites
Lo6 Hushan
Lo7 Jiedu
Lo8 Hanjiadu
Lo9 Shizhenjie
LolO Caijiawan
Loll Huanglongmiao
Lol2 Shuanggang
Lol3 Longkou
Lol3
Poyang Lake
Figure 1. Le An River-Poyang Lake area.
Research Centre for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, PRC.
165
-------
through Dexing, Loping, and Poyang Counties (latitude 29° 2') at an elevation of 32 m. The
stream gradient varies from 1 to 0.065 m/km. It may be divided into three segments as follows
(Table 1).
Table 1. The segment of Lo An River.
Segment
Upstream
Middle
Downstream
Length (km)
61.4
83.6
72.5
Gradient (%)
1.0
0.34
0.065
Width (m)
50-100
100-200
> 200
Depth (m)
1-2
2-5
5-10
Its river bed is full of rocks, pebbles, and sand, and not rich in sediments except a
segment near Poyang Lake. The average temperature is 17.1 C (high 29.6 C, low 0.64 C).
Average rainfall is 1882 mm, 150 raining days, from April to June. The dry season starts from
October to February each year. River flows are estimated by the record from hydrological
stations for 10 years as follows (Table 2):
Table 2. Average hydrological records (10 years).
Station
Xiangtun
Hushan
Shizhenji
Avg. runflow
(m3/y)
40.7 x 108 '
68.8 x 108
115.0 x 108
Avg. flow
(m3/s)
129
218
298
Avg. velocity
(m/s)
0.06-3.0
0.06-1.0
0.01-0.46
The upper 46.5 km of the river is lightly buffered with an alkalinity of 0-3 mg/L CaCO3
and is classified as being sensitive or highly sensitive to acid pollution. The middle segment of
47.5 km is buffered with 3-8 mg/L CaCO3, classified as being sensitive as well. The downstream
72.5 km is alluvial plain, classified as sensitive with 8-10 mg/L CaCO3, with not much limestone
in the red soil area.
Poyang Lake is the largest freshwater lake in China, about 150 km long from north to
south, 50 to 60 km wide, 3900 km2 water area in the rain season; but in the dry seasonitis
about 500 km2, 20 m depth in flood period, and water level about 22 m. The capacity of the
lake is 3.00 billion m3. There are plenty of aquatic products in Poyang Lake, e.g., fishes, wild
animals, and wild plants. Fishes include carp, silver carp, grass carp, crucian carp, salmon, eel,
shrimp, snail, fish and mussels, etc. The annual fishing production is about 36 to 45 thousand
tons in the 1960s; 25 to 30 thousand tons in the 1970s.
Unfortunately, a great amount of wastewater has been discharged to the lake from cities
and counties nearby. The annual amount of wastewater received is about 34.19 million m3,
containing organics (COD, BOD), CN, CrO4-2, Cu, Zn, and sulfur-bonding matter etc.,
166
-------
particularly heavy metals. The metal concentration in sediment near estuaries reached tens to
hundreds ppm.
The study is aimed at the Lo An River from Dawu River mouth at Dexing mining area to
Poyang Lake, about 210 £m long.
-'
FIELD STUDY, SAMPLING AND ANALYTICAL PROCEDURES
Lo An River in Dexing County includes two tributaries, named Dawu River and Jishui
River, surrounded by four mines. The largest opencast copper mine in China is located at the
upstream of Dawu River, while the other mines are located at the Jishui riverbank, named the
Fusiayu Copper Mine, Damoushan Copper Mine, and one lead smelter plant, and Yishan Lead-
Zinc Mine as well.
It is necessary to study mine drainage from the mines entering into the Lo An River and
assess the water quality at various sections along the river. Sampling was conducted at 13
sections and four sources of acid mine drainage in the rainy season and dry season from 1987 to
1990 (Figure 1). Both pH and Eh were determined with a portable pH meter. Flows at the
mouth of the tributary were measured with a flowmeter. River water, suspended matter, and
sediments were sampled. Duplicate water samples were collected at each section, one treated
with 2 ml HNO3, the other filtered with 0.45 p membrane. The filtrate was used to measure
anions and dissolved metal ions. Suspended matter was collected and filtered with 0.45 /*
membrane in 10 or 20 liters. The samples were refrigerated and digested with HNO3 and
analyzed with ICP and AAS, and with Dionex ion chromatagraph for anions.
RESULTS AND DISCUSSION
COMPOSITION AND ACID MINE DRAINAGES
The acid mine drainages sampled at four mines from 1987 to 1990 were analyzed. The
pH of combined seepage at Dexing Copper Mine ranged from 2.22 to 2.75, Fe concentration
from 8250 to 35724 ppm, sulfate 8250 to 11150 ppm, copper 72.53 to 129.2 ppm, Al from 741.8
to 1300.4 ppm, Pb, Zn, Cd, Ni, and As, etc., ranged from several to tens ppm. The amount of
drainage is about 6-7x107 tons/yr.
The pH of Fujiayu Copper Mine drainage is 2.54 to 3.72, the concentration of the heavy
metals are similar to Dexing Mine drainage, the amount of waste water is around 6-7x106
tons/yr.
The pH of drainages in Damoushan Mine ranged from 1.97 to 2.34. That is more acidic
than other drainages as it combines with the wastewater from pyrite and the copper mine, with
higher concentration of Fe, Cu, SO4-2, As; the amount of drainage is 2-3x107 tons/yr.
The Yinshan lead-zinc acid drainage is higher in concentration of zinc, lead, and As, and
the amount ranged from 5-6xl06 tons/yr. The typical results (on the average) are listed in
Table 3.
167
-------
Table 3. Typical composition of acid mine drainages (concentrations in ppm).
Elements
pH
r*
Ca
Me
•***o
K
Na
Cu
Pb
Zn
Cd ;
Ni
Fe
Mn
Al
As
SCX
vj«^4
HCO;
NO'j
CU
Dexing
2.51
235.17
538.63
• 47.07
3.11
102.15
0.427
238
0.131
4.47
23478.90
65.94
979.46
2.268
9716.67
237
0.73
033
Fujiayu
3.59
201.54
61.93
14.71
4.11
64.91 !
0.124
0.677
0.103 |
0.444
36.63
7.746
14.414 ;
0.002
769.50
0.00
5.40
0.60
Damoushan
2.41
147.89
119.72
5.93
3.51
328.09
0.171
3.820
0.176
2.027
5700.17
19.17
243.22
0.002
10825.00
0.00
13.05
0.28
Yishan
2.50
63.12
102.30
5.27
15.97
2.94
0.445
454.10
1.266
1.351
171.40
106.57
4937
0.424
1950.00
0.00
1.66
0.00
All of these acid mine drainages are running into the Lo An River through two
tributaries; the acidity and concentration of elements in Lo An River water decreases due to
dilution precipitation and neutralization. Some chemical and physical processes in the
tributaries have been studied (Tang and Chen 1990, Liu et al. 1989). The typical results in the
Dawu River and Jishui River mouths are shown in Table 4. The water flows are 65-85.8 and
586.5 million tons/yr, respectively-
Table 4. Typical composition of elements in tributaries (concentrations in ppm)
Elements
pH ;
Ca
Me
*»*o
K
Na
Cu
Pb
Zn
Cd •
Ni ;
3Fe I
Mn ;
Al
SGv*
HCOi
NO',
cr
Dawu
(1)
5.28
180.91
47.16 ;
88.95
21.77
4.27 :
0:00 ;
1.03 ;
0.04
0.46
2L03 ;
452
17.O6
24030
1.03
230
6.20
(2)
4.60
157.60
7330 i
2.49 :
6.83
11.75
0.12
0.98
0.05
0.22
218.89
4.14
80.96
467.57
2.12
1.73
4.65
(3)
5.05
139.08
79.09
9.21
632
8.59
0.08
0.31
0.014
0.20
22.23
6.14
61.70
177.0
0.98
1.00
1.10
Jishui
(4)
7.15
19.70
4.15
3.28
8.70
0.003
0.052
0.112
0.005
0.005
0.007
0.002
0.032
19.76
6.60 \
1.33 ;
2.46
(1) Dry, (2) Rainy and (3) Normal season at Dawu River. (4) Dry season at Jishui River.
168
-------
WATER QUALITY OF LO AN RIVER
The impacts of acid mine drainage on the Lo An River were assessed. The field
investigation shows that the pH and concentration of Fe, Mn, Al, SCv2, Cu, Pb, Zn, etc. in
surface water changed dramatically; the contents of suspended matter and composition of
sediments changed as well.
The concentration of metals in surface water (Table 5) indicated that the pollution of
heavy metals on water quality is evident. Copper, zinc, lead, and cadmium are the main
elements accumulated in the sediments and suspended matter (Tables 6,7). The concentration
of metals reached several thousands to hundreds mg/kg, hundreds of times higher than the
background levels in this area.
Table 5. Total concentration of metals in river water (mg/L).
No.
Cu
Pb
Zn
Cd
Fe
Al
Mn
Cr
Lo. 1
.002
.023
.080
.002
.170
.069
.020
.000
Lo. 2
7.497
'
.283-
.007
80.753
43.3
5.415
.008
Lo. 3
.205
.015
.072
.003
2.279
.236
.138
.128
Lo. 4
.035
.015
1.104
.004
.882
.585
.252
.067
Lo. 5
.015
.011
.108
.003
.313
.222
.087
.115
Lo. 6 .
.016
.034
.089
.001
.716
387
.109
.003
Lo. 7
.015
.015
.145
.002
321
.238
.056
.127
Lo. 8
.014
.023
.104
.003
.321
.328
.047
.055
Lo. 9
.020
.023
.041
.003
.231
.163
.046
.000
Lo. 10
.035
.015
.033
.004
322
.078
.015
.000
Table 6. Content of suspended matter and its metal concentration (mg/kg)
No.
Susp.
Cu
Pb
Zn
Cd
Fe
Al
Mn
Cr
Lo. 1
14.8
34.9
470.9
24453
31.4
57836.1
220371.6
5023
257.4
Lo. 2
160
10199.7
217.0
5921.6
72.4
1503523
157954.5
1822.4
966.4
Lo. 4
18.4
613.6
240.0
2598.3
25.9
29048.5
154753.7
476.4
271.0
Lo.5
16.4
292.1
327.1
33141.5
30.5
18718.4
164038.5
528.4
96.4
Lo. 6
13.4
12513
356.9
3724.7
24.2
72785.1
140706
8713
3023
Lo. 7
6.1
1097.6
722.8
5583.6
61.7
64467.2
377936
1244.2
2263.0
Lo. 8
65
1098.5
1044.1
12606.9
182.6
108208
1046462
24613
751.9
Lo.9
12.1
7285
11103
981.7
133
316583
23114.6
895.2
2795
Lo. 10
18.0
392.7
59.7
1150.7
11.0
10987.4
4417.2
205.4
903
Table 7. Metal concentration in sediments, 1989 in normal season (mg/kg)
No.
Cu
Pb
Zn
Cd
Fe
Mn
Al
Cr
Lo. 1
33.5
45.4
245.0
3.5
31237.5
713.8
44370
95.5
Lo. 2
3109
0.9
53.2
1.7
4105
18.7
28657.5
47.0
Lo. 3
2689.3
15.9
92.4
2.4
20967.5
231.7
211303
104.8
Lo. 4
1648.0
38.3
183.5
3.0
38575
748.9
13826
79.4
Lo.5
670
139.1
1040.0
4.6
35890
604.2
12642.8
72.7
Lo. 6
1400.1
72.8
707.0
4.4
373225
1442.9
103663
84.0
Lo. 7
586.2
116.1
787.1
5.8
45295
1354.0
946075
445
Lo. 8
5922
75.7
608.0
45
64025
1246.7
27460
112.1
Lo.9
517.4
67.4
496.8
43
30845
1360.7
31003
81.4
Lo.10
526.7
79.9
7579.7
5.7
25775
1728.8
111393
110.0
169
-------
It is shown that the pH of the river water changes from Dawu River mouth where acid
mine drainage is discharged downwards about 30 km (Figure 2). The pH is higher in the dry
season than in the rainy season, about 0.5 pH unit. The concentration of metals and sulfate
increases dramatically, especially in sediment and suspended matter. The impacts of acid
drainage on water quality in the dry season is less than in flood season, as the amount of acid
drainage in dry season is less than in rainy season. The concentration of metal decreased
steeply along the river. The impacts of acid drainage on water quality was affected by a series
of physical and chemical processes; neutralization, precipitation, sedimentation, and adsorption
of suspended matter are the possible important processes.
3C-
o,/
19.5 20.5 40 60 84 % 106 125
Distance, km
138
159
Figure 2. pH changes along Le An River
(1 = 1989, 4; 2 = 1989, 12; 3 = 1990, 4).
MODELING OF THE WATER QUALITY OF LO AN RIVER
It is very complicated to simulate the fate, transformation, and transportation of acid mine
drainage in the Lo An River. A series of effects should be considered in water quality modeling
for heavy metals. There are some processes that depend on the pathway and time in which the
pollutants would be transformed and transported. Kinetically controlled processes are often
very important The sedimentation of suspended solids, adsorption of metal ions on particles,
and kinetically controlled dissolution and precipitation are slow processes. In phenomenology,
some processes behave like the dispensation of concentration in a reactor. The relative
coefficients can simulate these above processes. In the model, the sedimentation coefficient and
partition coefficient were used to simulate these processes in steady state. The following model
was assigned (Salomon and Bril 1990):
6C
"5F
= D
+ V4T - K * C + R
(1)
where C = concentration of metals in suspended particles (g/m3)
D = diffusion coefficient (L/s)
V = average velocity of flow (m/s)
K = sedimentation coefficient (L/s)
R = nonpoint source discharge (g/s)
170
-------
In the model, the concentration of metal in water was supposed to divide into two parts as
dissolved and participate species; the paniculate species include active phase and inert phase
The partition coefficient of metal between the suspended particulates and water is defined-
where Cse
Cwd
K'p •. Q»
active part concentration of metal in suspended particulates
(mg/kg).
dissolved concentration of metal in water (mg/L).
partition coefficient of metal (L/kg).
(2)
The sedimentation coefficient is defined as follows:
.
where Sp = velocity of the suspended particles sedimented to
bottom (m/s).
depth of water in segment (m).
sedimentation coefficient (1/s).
(3)
H
Ks
The partition coefficient above can be estimated from the speciation distribution of metals
in suspended matter and sediment, and the dissolved concentration of metal in water The
sedimentation coefficient may be obtained from the field investigation and experiment (Lin et aL
In the model, the balance of the suspended matter in water is
where
SSt = SSi + SSp - SSout - SSd
(4)
SSt
SSi
SSp
SSout
SSd
and SS = SSt/Vq
where SS
Vq
total mass of suspended matter in water compartment (g)
mass of suspended matter from upstream (g)
mass of suspended matter from lateral boundary (g)
output mass of suspended matter to downstream (g)
mass of suspended matter sedimented to bottom (g)
= concentration of suspended matter in water (g/m)
= volume of water at the compartment (m3)
(5)
Model calculations were conducted using an IBM/PS-2,70; the program for the model was
compiled by Salomon and Bril (1990).
The concentrations of Cu in water, sediments, and suspended matter calculated from the
above model of heavy metals are shown in Figures 3-4.
171
-------
U.
-------
The effect of washload on the riverbed promotes the transport of pollutants downstream into
the estuaries and Poyang Lake. Thus, water quality will be degraded, particularly in the
beginning of the flood season.
The thermodynamic equilibrium model MINTEQA2 (Allison et al. 1991) can be used to
calculate the species distribution of metals in the river water based on the monitoring data of
water quality in various segments. The concentrations of cations and anions and pH of the
water (Table 7) were entered into the input file of the computer program. The results showed
the! proportion of species and main minerals to be formed in water in saturated indices, given in
lable 8. The results show that the species distribution of metals in water will change
dramatically depending on the pH and species of anions; Eh and the dissolved oxygen in water
are almost similar in the whole river: Eh 190 mv and the pressure of oxygen in air is 103 atm.
The toxicity of metal ions is different from one species to another, depending on the
environmental conditions and the species distribution of metal.
1. Aluminum forms a series of species in water that are pH dependent Free Al
predominates below pH 5.0. The results (Table 8) show that the percentage of free Al ion
exceeds 50% of the total concentration in Lo.2 (64.2%) and in the Dawu River mouth (52 8%)
The secondary species is A12(S04)3 (31.9%) and in the Gukou and Dawu River mouths (38 9%)
T^uf^o^15 f°rmed might be allunite A1(°H)S04 in the Gukou and Dawu River mouths.
A1(OH)3 (58.3%) predominates in Lo.3 from pH 4.79 to pH 6.64. A113(OH)32 will be formed
between pH 6.0 and pH 6.5. A113(OH)32 and A1(OH)3 begins precipitating from Lo.3 to Poyang
Lake. The free Al ion concentration is very low in middle and downstream to Poyang Lake
A literature review of Al toxicity to the affected species reveals that dissolved concentrations
over 1.5 mg/L would cause drastic physiological changes for the warmwater fish, and 0.5 mg/L
would be set as a standard for them (Moore and Ramamoorthy 1984). It is a serious problem
that a large amount of Al is discharged from acid mine drainage to the Lo An River These Al
concentrations appear to exceed national levels for the fish and other aquatic organisms
especially in the flood time with lower pH; the most toxic effect of Al will occur from Lo.2 to
Lo.5. Thus, Al impacts on Lo An River should pay attention to the water quality.
2. Copper is highly toxic to aquatic plants, invertebrates, and fishes. The copper
concentration in Dawu River mouth and Lo.2 exceed the standard for fisheries and reaches
1.944 mg/L. Free Cu predominates below pH 5.0; the most toxic effects will occur from Lo.2
(Gukou) to Lo.3 (Zhongchow), and the total concentration in downstream almost exceeds the
standard for fisheries, 0.01 mg/L. Copper of acid drainage is the main pollution element in Lo
An River. This result was illustrated with the data of the Cu accumulation concentration in
fishes and algae (Research Centre for Eco-Environmental Sciences, unpublished data).
3. Other elements. The concentration of Cd, Zn, and Pb reach and almost exceed the
national standards for fisheries, especially in the dry season. But Cd and Pb are almost
eliminated with pH increase. Free Cd in Lo.3 predominates below pH 5.65; CdCO3 will
predominate in PH 7.25. The solubility of ZnCO3 and Zn(OH)2 is higher, so the dissolved Zn in
water is higher. In general, no effect will appear on the water quality in the Lo An River. All
toxicity data for iron are suspect and essentially no work has been done for Fe at low pH and
there is no standard for fisheries for it in China; but Fe national standard for drinking water is
0.3 mg/L. It is impossible to predict impacts of iron to aquatic biota, but iron exceeds the
standard of 1.5 mg/L for warm water fisheries. In addition, large amounts of iron will cause
water treatment problems for city water supplies. Recent data on the toxicity of Mn to aquatic
species are limited, but the maximum acceptable toxicant concentration of 0.77 mg/L for
173
-------
a
<4-<
O
a
.2
-*-*
.-2
i!
.S3
*o
.§
8.
CO
o
*^£
&0 ^
1 3
,2
tf)
^
*
^fi K/*
^ *
j
a
+
X-N
tf)
|
Na+(98.9),NaSO4+(l.l)
t— t
a
a
j
a
??|
^g
11
^x
fl
aa
^^ ^
t),CuSO,(11.9
.
3
CO
a
£
a
(OH)j(99.9)
a
3
sf
£ r"^
:03(41.1),Pb(0
OH)2(36.7),Pb(
COOi'CLT)
P« AH PL,
|
^
a.
P
1
O
1
s-
06
?
£
DH+(31.4),Pb(OH)2
.5),PbCO3(45.2)
sa
x>
P-,
s^
^r
O
^
^
f
fi
•
f
Zn+2(76.8),ZnS04(22.4)
+2(i5),Zn(OH)-I(4.5)
(OH)2(91.1),ZnC03(l.:
S.S
A
/"• x
V^Q
^21.0),O1(OH
OH)2(2.5),CdC
33
8
<=>
° ^
33
2+(54.4),Cd(OH)+(7.5)
CO3(36.6)
33
3
OH)3(1.7),
OH)
-------
3
•»„
5*
I?
II
2&
I
CO
o
CO
If 11
oo z 0
o of o « ^S
££os -S
&
8
o
*T
o'
i~
la
£o
of
co O
30)
I
3*sa
xd£~
foil
S/^ a -S SP c ^.
«-a -^ -5 o -^ u M
<<<« QS s
S
0
«? rJ
c,oo
s-co co
^^ rt w .4
Eife fi -far^
vf^ffxt -a u a « c
~ rS !S -H W -a s 'C S
1
),NiSO
'Qrd
^69.7
(S04)(1
S002-(
s
g
r
I
I
00
_u
3
K v--
O m~ O
%& %
<^ ^-^ ^J
1O ij
c-y 3
E^- ^
O ^ ^r
•^ T—I *—'
^ ^ CO
p
4 s^
o rt
O |
Ca
Di
Fe
§1
i1
O
8
i-6
43
S S
•§ g
£.s
•a
I.?
3-3
175
-------
rainbow trout was reported. The maximum recommended concentration for drinking water is
1.0 mg/L; it will cause dramatic effects on the people who live near the river and have impacts
on the water treatment.
CONCLUSIONS
(1) The pH, aluminum, and copper are important elements for impacts of acid mining
waste water on water quality in the Lo An River. The pH of water in Lo.2 to Lo.3 will decrease
one half of a pH unit as the acid mine waters discharge into the river. A large amount of heavy
metals will cause drastic physiological changes to the warm water fish and eliminate the products
of the fishery, which condition has occurred in the last several years in the Lo An River. The
copper and aluminum are highly toxic to aquatic plants, invertebrates, and fishes. Iron and
manganese will cause water treatment problems for city water supplies.
(2) Large amounts of heavy metals are deposited and precipitated in the polluted plume
zone near the Dawu River mouth. The concentrations of copper, zinc, iron, manganese, and
aluminum in sediments increase dramatically, and transport downwards. The results from the
model show that a large amount of suspended matter is transported downstream and deposited in
the river bed and estuary. The persistent effect on aquatic biota should be studied in future work.
(3) The combination of the river transport model and the equilibrium model
(MINTEQA2) can describe the water quality in more detail and are more useful to predict
water quality.
ACKNOWLEDGMENT
This work is part of a project supported by UNESCO CERP project and the Chinese
National Science Foundation. Many thanks to Dr. W. Salomon (Institute of Soil and Fertility,
RA Haren, Holland) and Dr. R.C. Russo (Environmental Research Laboratory, U.S.
Environmental Protection Agency, Athens, Georgia) for support of the river model and
MINTEQA2 programs, and acknowledgments to Dr. Jennie Liu and Prof. Tang.
REFERENCES
Allison, J.D., D.S. Brown, and K.J. Novo-Gradac. 1991. MINTEQA2^>RODEFA2, A
geochemieal assessment model for environmental systems: Version 3.0 User's Manual.
U.S. Environmental Protection Agency, Athens GA. Publication No. EPA/600/3-91/021.
Baker, J.P., and C.L. Schofield. 1982. Water, Air, and Soil Pollution. 18: 289.
Moore, J.W., and S. Ramamoorthy. 1984. Heavy Metals in the Natural Water. Springer
Verlag. New York. 268 pp.
Ramsey, D.L., and D.G. Branoon. 1988. Water, Air, and Soil Pollution. 39: 1-14.
Salomon, W., and J. Bril. 1990. River model computer program for transport of heavy metals
in suspended sediment. Institute of Soil and Fertility, Ra Haren (Gr) Holland.
Xun Bao Guo. 1986. The first symposium of environmental protection for mining processes.
10,11: 81-84. Research Center for Eco-Environmental Sciences. 1989. Assessment and
investigation of Lo An River water quality. (Unpublished), 5 pp.
176
-------
WATER ENVIRONMENT MANAGEMENT IN CHINA
by
Zhu, Xing-Xiang1
Since 1982, China has strengthened the control and management of water pollution. In
April 1984, the Fifth Session of the Sixth National People's Congress promulgated the "Water
Pollution Control Law", and the Environmental Committee of the State Council passed "The
Regulations of Water Pollution Control Technology Policy." From the beginning of last year,
water resources protection plans have been drafted for seven big water systems (basins). In the
management structure, the water resources protection bureaus of offices have been set up for all
seven water systems. Over these organizations, it thus gained some experience in managing
rivers according to the water systems, and enhanced the management of monitoring and
planning work. Nevertheless, as there are still many problems remaining unsolved in water
pollution control management, technology, and finance, water pollution in some places is very
serious at present.
BRIEF INTRODUCTION ON DEVELOPMENT AND USAGE
OF WATER RESOURCES
China's total water resources quantity ranks sixth in the world. The total quantity of water
resources is 2,800 billion CM. and the average personal occupation per year is about 2,700 CM.
In China, the main supply of water resources is rain. Owing to the fact that water
resources are not evenly dispersed both in areas and seasons, the contradiction between supply
and requirement become sharper. China has about half of her land located within dry or semi-
dry areas. In these areas, the precipitation is less than 400 mm. In addition, the south part of
China has more underground water than the north. The underground water in the area south to
Yellow River covers about 62% of the country's total supply. Rainfalls in most of the areas are
very much influenced by monsoon, and more than 60% of the precipitation is concentrated in
3 or 4 months every year and mostly in the form of storms.
To solve the problems, some measures should be taken, such as making facilities to let
south water flow to the north to mitigate water problems.
In recent 20 years, along with city development and raising of the people, the quantity of
water normally consumed doubled in all cities, and the amount increases by 4% every year. The
main problems in water resource usage are that agricultural water consumption is too big (88%
of the total consumption), and industrial water consumption per unit is too high. For instance,
when making 1 ton of steel in China, most steel mills need to use 70 to 100 CM of water.
JWater Environment Management Division, Department of Pollution Management, National
Environmental Protection Agency, PRC.
177
-------
In addition, management structure should be reformed; the responsibilities of water supply, use
management, and flood control belong to separate departments. These departments are
overlapped, doing things their own way, which results in very low efficiency.
WATER POLLUTION AND ITS EFFECTS
The whole country's wastewater is 36.8 billion CM per year, of which 75% are industry
wastewater and 25% are domestic discharges; 80% of the wastewaters are discharged directly
into water bodies, thus contaminating 90% of the water bodies around cities. For instance,
Shanghai's daily effluent quantity is 5 million CM, of which 3.2 million CM is discharged into the
Huangpu River, the drinking water source of Shanghai City,
WATER ENVIRONMENT MANAGEMENT
In China, the law authorizes that the environmental protection department of the State
Council is an organization practicing unified supervision and management of water pollution
control. It's main responsibilities are organizing the work of making a national water pollution
control plan; getting up enforcement and practicing supervision on national water pollution
control policies, guiding principles, laws and regulation; coordinating the work of departments of
the State Council and guiding the National Environmental Protection Agency (EPA) on the
levels of provinces, autonomous regions, and municipalities directly under the central govern-
ment. The EPAs of local governments are responsible for supervision and management of
water pollution control in their jurisdictions.
MAKING NATIONAL AND LOCAL STANDARDS
The environmental protection department of the State Council takes care of making
national standards on environmental water quality and on discharge of pollutants, while the
Peoples' Government of the provinces, autonomous regions, and municipalities directly under
the central government have the right to make local standards which are not stipulated by the
national standards. The local standards, relating to neighboring provinces, will be coordinated
and adjusted by the environmental protection department of the State Council so as to avoid
shifting responsibilities onto each other, and to promote the establishment of local standards.
Local government may set local discharge standards. When enforcing laws, if there are local
standards, the local ones prevail.
DESIGNATING WATER RESOURCES PROTECTION AREAS
The designating of water resources protection areas is organized by EPA, in consultation
with the department concerned and approved by country government or above. Transregional
areas should be approved by all governments concerned. At present the emphasis of water
protection is to protect drinking water sources.
178
-------
WASTEWATER DISCHARGE APPLICATION AND REGISTRATION
In order to successfully control water pollution, China will gradually practice the system of
discharge license on the basis of application and registration system, to let the work step from
qualitative management to quantitative control so as to set targets for the management In
China, the first one to practice the system is Shanghai City. There are 108 enterprises which
have obtained licenses in the city.
In July 1987, the state EPA decided to adopt the application and registration system for
the whole country and, if the condition allows, to practice the licensing system in some place so
as to get experience to promote the whole country. Currently, there are about 50 cities in which
this has been practiced.
IMPOSE DISCHARGE FEES
The enterprises which discharge pollutants to waterbodies should pay discharge fees
according to the regulations of the country. If the discharge exceeds the national or local
standards, the enterprises should follow the regulations to pay the discharge fees for exceeding
standards and be responsible for treatment.
To impose discharge fees is an effective measure of pollution control by legal and
economic means. Before imposing fees, the economic loss caused by discharging pollutants is
unreasonably shifted onto every one in the society, while the polluter may discharge without any
legal restrictions and responsibilities.
The funds for treatment and control of pollution are provided by the government Some
discharging units even think that pollution control is a problem only for EPA to consider. After
practicing the system, the situation changed greatly.
COMPREHENSIVE CONTROL AND TREATMENT OF CITY WASTEWATER
The drainage facilities in China are badly insufficient. Drainpipes cover only 30-40% in
cities, and the wastewater treatment rate is less than 3%. In order to comprehensively treat city
wastewater, to treat industrial wastewater together with domestic wastewater, and to change the
situation of treating wastewater separately and on small scale, the regulation stipulates that so
long as the wastewater effluents by enterprises meet the standards of municipal drainage system
receiving water quality, the wastewater could be discharged and will be treated in the city
wastewater treatment plant. If it exceeds the standard, the water should be pretreated before
discharge.
MEASURES ON WATER POLLUTION CONTROL
(1) We must pay emphasis to controlling new pollution sources and reducing discharge of
existing sources. Newly built enterprises should evaluate environmental impacts and carry out
179
-------
"three at the same time" actions (the same designing, the same construction, the same used for
one project). This system is practiced comparatively well in big and medium construction
projects at the rate of 90%.
(2) The management should be carried out in accordance with the law and on the basis of
science.
Now the very important problems: Investigate and solve basic problems to serve manage-
ment
(1) Industry discharge target and standards. It should be reasonable in controlling the
discharge from industries. The amount permitted to discharge should be decided according to
the level of the production process. Less advanced production processes should be limited, so
that more advanced production processes can be adopted, and pollution can be reduced.
(2) Problems about the effects of aquatic pollutants on human health and aquatic life still
need more investigation.
(3) The inter-relationship among discharges from different cities in the same basins has
become a urgent problem in China. If it cannot be solved successfully, the effectiveness of city
pollution control will be blind.
180
-------
WATER QUALITY PROJECTIONS FOR LAKE BOSTEN,
XINJIANG, PEOPLE'S REPUBLIC OF CHINA
by
Zhang, Guo-An1, Zhu, Dong-Wei2, Steve C. McCutcheon3
Pei, Xin-Guo4, Zhong, Xin-Cai4
BACKGROUND
Lake Bosten watershed in northwestern China covers an area of 149,200 km2, and is
divided into an upper and a lower basin by the Kuluk Mountains. The latitude is 40°25'-43°19'
north; the longitude is 83°-91° east The Yanqi Basin is generally known as the upper basin,
while the Kurle plain is generally referred to as the lower basin.
The Yanqi Basin spreads out from the south slope of the Tianshan Mountains with an_
area of 43,884 Km2 and an elevation of 1048 to 1205 m above sea level. The latitude is 41°25'-
43° 19' north; the longitude is 83°00'-88°20' east. Dominated by a continental climate, the
basin is very dry with light rainfall and a high evaporation rate. The average annual
temperature is about 8°C. The pan evaporation rate is about 1800 to 2500 mm, while the
precipitation is as low as 50 to 80 mm. Lake Bosten is at the lowest elevation of the Yanqi
Basin. It is the terminal point of surface water and groundwater of the Basin.
The Peacock River starts at a pumping station on Lake Bosten, flowing for more than 780
km through Tiemenguan Canyon, over the Kurle Plain, terminating at Lake Lob Nor. The
Kurle Plain is situated on the northern fringe of the Tarim Basin and at the southern foot of the
Tianshan Mountains. It is strongly influenced by the Takliffiakan Desert, the largest desert in
China and the second largest desert in the world. The annual average temperature is about
10.7° C, pan evaporation is 2668 mm, and precipitation is 61.2 mm.
Lake Bosten was called "west sea" in ancient times. It is far from any ocean. Lake
Bosten is composed of a large lake area, a small lake area, and a wetland area. The large lake
area is about 900 km2. Its volume is about 8200 million m3. The water surface is about 1045
meters above the mean datum (sea level) in recent years. The length of the large lake is about
60 km east to west and the width is about 25 km in the south-north direction. The average
depth is about 8 meters, while the maximum depth is 16 meters. The small lake area is about
360 km2, of which 45 km2 is free water surface, 280 km2 is reed area, 39 km2 is salt affected
Institute of Environmental Protection, Xinjiang, PRC.
2Department of Environmental Sciences, Nanjing University, Nanjing, PRC.
Environmental Research Laboratory, U.S. Environmental Protection Agency, Athens, Georgia,
USA.
"Xinjiang Academy of Agricultural Science, Xinjiang, PRC.
181
-------
lowlands. The small lake was separated from the large lake by a dike in 1982. The 130 km2
wetland area is at the northwest corner of Lake Bosten (Figure 1).
Lake Bosten functions as a freshwater modulate for irrigation and drainage in the Basin
area. A great variety of fish, such as bighead, blunt-snout bream, crab, and shrimp live in its
waters. The economical importance of the lake is also reflected in its production of 250 k-ton of
reeds and 2000 tons of fish every year. Lake Bosten is the summer resort of people from the
Kurle area, since the temperature here in summer is much lower than that of the urban area.
Lake Bosten is the most valuable resource in the region. It provides water for domestic
use, industry, irrigation, as well as receiving waste waters. It plays a vital role in every aspect of
the region.
PROBLEMS
In the last 30 years, serious changes have occurred in the lake as a result of human
activities, such as farming and possibly climate changes. There are three main problems in the
lake: (1) Water level has been dropping continuously; (2) Water quality is seriously degraded;
(3) Aquatic ecology is largely changed.
Yellow pitch
Yanqi
• Heshuo
Wet Lands
Hongshaliang
Peacock River
Small Lake Area
Figure 1. Lake Bosten, adjacent wetlands, reed production area, Kaudu River,
Peacock River, and Jiefang 1 Canal connecting the Kaidu and Peacock
Rivers. Heshuo is to the north of LakeJBosten.
182
-------
Lake Bosten is the terminal point of the Kaidu River and other small rivers in the Yanqi
Basin and is the source of the Peacock River. Since 1949, there has been a large agricultural
development in the Yanqi Basin. Because large amounts of surface water were consumed in the
upper watershed, the fresh water supply decreased (Figure 2). Since 1961, the Jiefang 1 Canal
was constructed to increase the water supply. It diverted part of the flows from the Kaidu River
into the Peacock River. The Kaidu River is divided into west and east branches. The west
branch goes into the small lake area, while the east branch goes into the large lake area. In
1958 a control gate was built on the Kaidu River to increase the water supply in the Peacock
River, and therefore greatly reduce the water flowing to the large lake.
1049
1048-
J 1047
~o
JB
UJ
1046-
Q)
"o
1045-
1044
Figure 2. Lake Bosten water levels from 1955 to 1989.
The water quality of Lake Bosten is serious degraded by pollutants from the Yanqi Basin
since the 1960s. In 1958 the total dissolved solids (TDS) was about 385 ppm; in 1975 the
concentration was about 1450 ppm; in 1987, it was about 1860 ppm. It was a fresh water lake
30 years ago, but it is now becoming a slightly saline water lake (Figure 3). The contaminants
consist of mineral pollutants, principally of sulfate, chloride and other dissolved solids, and
moderate concentrations of the nutrients, nitrogen and phosphorus. Because of this pollution,
Lake Bosten and its outflow, Peacock River, are of generally poor water quality.
The dropping water level reduced water surface area and the reeds area significantly.
Reeds and wetlands covered an area of 837.6 k-mu in 1959; 743.8 k-mu in 1981; and about 600
k-mu in recent years. (The mu is the traditional Chinese unit of land area, equal to 0.1647 acres
or 666.5 m2.) In the last 30 years, the total area has decreased about 230-k-mu. In 1965, the
yield of reeds was about 400 k-ton a year; it was only 316 k-ton in 1981; in recent years, it is less
that 220 k-ton.
183
-------
2000-
1800-
1600-
^ 1400-
§" 1200-
w 1000-
_o
•a 800-
"ci
H 600H
400
200
„.!..
4....I
i i
t i
'"7—f"
i I
S 1"
1
1.
i
i i
Q-\ "-I i i i I i i i i I I i I I I i M i i
58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88
Time in Years
Figure 3, Total dissolved solids in Lake Bosten from 1958 until 1989.
SALT ACCUMULATION AND CONTROL IN UPPER WATERSHED
The Tianshan Mountains are the source of the Kaidu River and all other small rivers. The
Kaidu catchment covers 27,000 km2. The average annual salt carried into the Yanqi Basin is
1115.7 k-ton. '
The Yanqi Basin is located in the southern part of Xinjiang. High mountains bar the
approach of monsoons and wet winds from the Pacific and Indian Oceans, and only a smattering
of the cold winds from the Arctic and Atlantic comes through mountain glens to reach the
Basin. In this area, the potential evapotranspiration is about 1,000 mm, which is about 15 times
that of rainfall. The moisture deficit, which is the difference between the potential
evapotranspiration and rainfall, is the active factor in salt accumulation, first in ground water
and second in the soil itself.
Since the Han Dynasty, people began to take water from the rivers for irrigation in the
Yanqi Basin, but for a long time, the agricultural production was small. Since 1949, farming has
been developed significantly. By the end of 1989, the irrigation area reached over 1543 k-mu,
an increase of 11 times compared with the amount before 1949 (Figure 4). As a result of the
increased irrigation, ground water tables in the region were elevated and salinity in the soil had
also increased greatly, according to investigations in the last several years.
184
-------
20
70
Year
i—i—i—i—i—i—i—i—n—i—n—i i i I—r1
73 76 79 82 85 88
-»— Ground Water Depfh -*- Irrigafion Wafer
Figure 4.
Groundwater depth and irrigation water consumption in the Yanqi
Basin from 1949 until 1989.
By 1962, irrigators in the Yanqi Basin began the construction of canals to remove saline
drainage water associated with irrigation with waters from Kaidu River and other small rivers.
All of the drainage from the Yanqi Basin terminate at Lake Bosten. As the drainage system
became more complete, the amount of salt return flows increased rapidly. In 1975, the return
flow was 0.465 billion m3, and 621 k-tons of salt. In 1981, the return flow was 0.265 billion
m3,and 637 k-tons of salt. In recent years, farmland has not been increased and the drainage
canals are blocked by sedimentation. Therefore, the return flow and salt going into the lake
have decreased.
WATER RESOURCE DYNAMIC PROGRAMMING ANALYSIS
For the planning of economic development hi the watershed area, the method of "dynamic
planning" is adopted for the investigation of the utilization of the water resources. The
evaluations have been conducted for the following three proposals (scenarios):
(1) With the proposed water demand as the objective, which is based on the runoff,
evaporation, and precitation rates calculated for the years 1991-2000 using time
series analysis.
(2) With the yearly increment of 10 cm of the lake water level as the objective, which
is based on the runoff, evaporation, and precipitation rates calculated for the years
1991-2000 using time series analysis.
185
-------
(3) With the current water level unchanged as the objective, which is based on the
mean value of runoff data of many years.
The results of the calculations on the above three proposals are briefly explained as follows
(see Figures 5 to 7):
(1) With the use of ground water and the inflows of surface water increased, the
agricultural drainage would also be reduced as the result of the lower ground water
level. Therefore, the objective can be achieved as proposed. The water level of
the lake increases/decreases with the increase/decrease of the runoff to the lake.
With the hydrological conditions unchanged till year 2000, the salinity would
decrease slightly.. "
(2) In this proposal, although the relatively high water level is maintained and the
salinity is decreased, the adjustment of the runoff of the lake could not function to
a satisfactory level. The outflow of the Peacock River fluctuates to the degree that
the water demand of the agricultural production downstream could not be met.
(3) In this proposal, the relatively high wafer level and low salinity are maintained, with
the average runoff of many years as the basis. The water demand downstream
would also be ensured. Based on the current plan, this is a relatively ideal and
realistic plan.
CD
>
0)
1046.0-
1045.0-
1044.0-H
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
Time in Years (1983—2000) "
scenario—1
1 scenario—2
scenario-3
Figure 5. Projections of water levels in Lake Bosten until year 2000
based on three dynamic programming scenarios.
186
-------
E 30.0
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
Time in Years (1983—2000)
scenario—1
• scenario-2
scenario—3
Figure 6. Projections of outflows in the Peacock River.
2.0
0.0
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
Time in Years (1983—2000)
• scenario—1
• scenario—2
scenario—3
Figure 7.
Projection of salinity levels at the pumping station delivering
water to the Peacock River from Lake Bosten.
187
-------
LAKE BOSTEN MODELING
MODELING APPROACH
According to the general salinity gradient pattern, Lake Bosten was divided into 5
segments, as shown in Figure 8. In segments 4 and 5, which constitute the major part of the
lake, there is no consistent concentration gradient. Hie smaller segments 1 and 2 were to
represent the concentration gradient from the Kaidu River mouth and to the Peacock outflow.
2 - Segment
2 -Channel
Figure 8. Model segmentation for Lake Bosten.
The modeling study was designed to conduct a water balance and salinity modeling.
DYNHYD5 was chosen for water balance and EUTRO4 was used to simulate the salinity
distribution (Ambrose et al. 1988). DYNHYD5 accounts for the factors influencing water
balance: inflow, outflow, rainfall, and evaporation (DYNHYD5 was specially modified to be able
to simulate evaporation). Information on adjective flows were then provided to WASP4.
Dispersion coefficients were added and adjusted in WASP4 to describe the mixing effects caused
by circulations in the lake.
Data gathered from 1987 to 1989 were used to calibrate and validate seasonally changing
dispersion coefficients. Long-term simulation from 1958 to 1982 was also made to validate the
model's ability to do long-term predictions.
The data base used for calibrating DYNHYD5 was collected in 1987. The measurements
include monthly inflow from the Kaidu River and agricultural runoff canals, outflow through the
pumping station, water elevation, pan evaporation, and rainfall.
The simulation results of salinity with calibrated dispersion coefficients are shown in
Figures 9 through 14. From the above results, it can be concluded that the models were
validated for both water balance and salinity simulation.
188
-------
12 18 24
Time in Month (1987-1989)
30
36
• Measured —*— Simulated
D
Figure 9. Simulation of seasonal water levels in Lake Bosten.
2500.0
2000.0
CD
M
T3
1500.0 -S /-
o
1000.0 -—/—-\-;
500.0
0.0 "TI TTiTrrrr -n rri 11 i I 11 i i 11 11 rl 11111 i m-hrr rrrrrr n-mrn-rhTrn
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Time in Month (1987—1989)
simulation
measured — minimum — maximum
imum
Figure 10. Simulation of seasonal salinity in Lake Bosten
near the outflow to the Peacock River.
189
-------
2500.0 cr-
^2000.0
\
CD
«•£• 1500.0
w
T)
"a
•+-•
o
1000.0
500.0
1
T 1
1
in
1
TTTTTT
~rr
0.0 i nrrrrir TTTVITT
0.0 5.0 10,0 15.0 20.0 25.0 30.0 35.0
Time in Month (1987 —1989)
1 simulation * measured — minimum — maximum
Figure 11. Simulation of seasonal salinity in the middle of Lake Bosten.
1050.0
1049.0-
"-' 1048.0
c
o
•I—
o
.1 1047^
01
"5
1046.0-
1045.0
0 3 6 9 12 15 18 21 24
Time in Years (1958—1982)
• simulation
• measurements
Figure 12. Simulation of the long-term water balance in Lake Bosten.
190
-------
2500.0
o.o-
niitiiiiraiiiraiimiiinniiniiiiiiiiiiiliiniiimiitiiiiiiiuiiiiiiiiiiramjiirtnni
69 12 15 18 21 24
Time in Years (1958—1982)
1 simulation
measurement
D
Figure 13. Simulation of long-term salinity in the middle of Lake Bosten.
2500.0
2000.0-
—N
C?
^ 1500.0-
61
TJ
o
H-
1000.0-
500.0-
0.0
0
iwiiiTihrniiinrnuniiMt'iniiwraiinnisi
6 9 12 15 18 21 24
Time in Years (1958—1982)
simulation
measurement
Figure 14. Simulation of long-term salinity in Lake Bosten near the outflow
to the Peacock River.
191
-------
FUTURE PROJECTIONS
There are many questions of concern about the future of Lake Bosten under different
management scenarios. The government has made tremendous effort in the region's
Development Plan to achieve a good management of the region's water resources. According to
the population increase, and agricultural and industrial development, a certain amount of water
supply must be guaranteed both from upstream Kaidu River and the Peacock River. Ground
water use will be increased rapidly to ease the pressure on surface water resources. Diversion
from Kaidu River will be reduced slightly through the increase in ground water use. Water
coming into the lake through the Kaidu River will increase to compensate for the increasing
water needs from the Peacock River. To assess the consequences of the ongoing governmental
plan for Lake Bosten, two scenarios were simulated for the next 10 years with the calibrated
model.
(1) Prediction driven by long-term averaged climate data
According to the governmental plan, the agricultural area in the Yanqi Basin will increase
to meet the needs of increasing population. The increasing agricultural activity will also elevate
the amount of salt coming into Lake Bosten.
The simulated water levels and salinities are shown in Figures 15 and 16. It can be seen
that the water level and salinity in Lake Bosten is increasing gradually for the 10 years under
this situation. ,
1047.0
1046.5- !-
1046.0-1—
Ql
k,
a>
1045.5
1045.0 liiiinirnittnniimrtmiTmmmim
TnTniniiniiMiiiTrnrriiTrhiiiiiniiinTmiTiTTi
456789
0123
Time in Yeras (1991—2000)
10
Figure 15. Projection of water surface of Lake Bosten with improved
management of water and averaged climate data.
192
-------
2500.0-r
0.0
0
Minimi ilnmim Mini mimiimiiii.iiulimiiiimiMii UNI iiimniiiiufr
_ mil rrrrrmTTlTVi rnn
0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9,0
Time in Years (1991—2000)
Figure 16. Projected salinity distribution in Lake Bosten. The upper three curves
represent segments 3, 4, and 5f The middle curve represents segment 2.
The lowest curve represents salinity near the outflow to the Peacock River.
(2) Climate data obtained through time series analysis
Projections of changes in climate and river flows must be related to past trends, frequency
of variance about a trend, and random factors because long-term average values may not be
able to represent the future situation very well. Therefore, a time series analysis was made for
Kaidu River flow, evaporation, and precipitation. Historical records from 1952 to 1986 were
used to obtain the three components (trend, frequency, and random) of the time series (Rich
1973). Records for 1987 to 1990 were used as a validation of the derived time series.
Figure 17 is the simulated lake water surface level. The water level has relatively large
oscillations through the years and decreases on the average. Figure 18 shows the simulated
salinity in Lake Bosten. The salinity goes up and down in correspondence to the oscillations in
the water level, but the changes are much smaller than that of the water elevation. On the
average, the salinity is higher than in the case where the long-term average data are used.
193
-------
1.046.0n
1045.5-
JH 10,45,0
If 1044.5-
» 1044.0
o
5:
1043.5;
TQ43.Q'uimimninnrnTnihiiiiiMiiiiiriMiiiiiiiuMrnTTTTiiiiiMUMiirmrriTmirnriirfniiTTTiiTiniiiiiiilillltl
2 3 4 5 6 7 8
Time in Years (1991—2000)
Kgure 171 Projection of water surface and storage in Lake Bosten
using a time series analysis.
10,
2500.0-f—
r m m rn IrnrrrnnrtTriTriifrinrrrnrnTnniiMnTTrrJnirrri
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0
Time in Years (1 99 1 —2000)
Figure 18. Projected salinity distribution in Lake Bosten using the time series analysis.
The upper three curves represent segments 3, 4, and 5. The middle curve
represents segment 2. The lowest curve represents salinity aear the outflow
to the Peacock River.
194
-------
CONCLUSIONS
1. DYNHYD5 and WASP4 were successfully used in modeling Lake Bosten. They were
efficient in simulating long-term water balance and salinity changes in large arid region
lakes. These models can be easily expanded in the future to simulate other water quality
problems such as eutrophication.
2. Evaporation from the surface of Lake Bosten was correctly estimated using the pan
evaporation data available. Important factors in the water balance in Lake Bosten were
analyzed. It was found that Kaidu River inflow, evaporation, pumping station outflow, and
agricultural runoff are the controlling factors of Lake Bosten's water balance. Ground
water inflow and outward percolation do not have decisive roles in the overall water
balance.
3. Agricultural runoff is the major reason for the elevated salinity in Lake Bosten. Salinity in
Lake Bosten responds to outside changes slowly. It took a long period to elevate the
salinity. It will also take a long time to reduce it.
4. The future of Lake Bosten depends on the management method. Drastic increase in -
ground water use is necessary to maintain water level in Lake Bosten and meet the water
resource supply to the region. Salinity in the lake can be altered greatly through lake
management.
5. Lake Bosten is the middle loop of an inland watershed. Water and salts come through the
Kaidu River, Lake Bosten and Peacock River, disappear and accumulate in the desert.
Human activities destroyed the natural balance of water and salts. More surface water was
used than the system could naturally support, therefore the water level was decreased. The
high input, low drainage irrigation practice accumulated salts in the large area of farmland.
Removal of this salt caused the imbalance of salt in Lake Bosten. As the salts come with
the river flow and cannot disappear, its balance must be taken into consideration in future
water resource management.
ACKNOWLEDGMENTS
The authors would like to express their gratitude to Mr. Zhang Chunghua, Zang Yuxiang,
Liang Sicui, senior engineers of the P.R.C. National Environmental Protection Agency, and Dr.
Rosemarie C. Russo, Director, Athens Environmental Research Laboratory, U.S. Environmental
Protection Agency, who have visited Xinjiang many times and helped to organize to solve the
problems with Lake Bosten. Mr. Robert Ambrose of the U.S. EPA made, valuable proposals on
analysis and monitoring of the water quality of the lake.
The completion of this project would not be possible without the valuable suggestions from
the following friends: Mrs. Li Xixian, Han Xiankun, Chen Jingsheng, Fu Guowei, Wu Shenyan,
and Jin Xiancan, professors and senior engineers from China.
During the modeling work at the Environmental Research Laboratory, Athens, the authors
had many valuable and beneficial discussions with Mr. Robert Ambrose, and with Dr. P. F.
Wang, Mr. Tim Wool, and Dr. Mansour Zakikhani, all of the AScI Corporation, and Professor
Wu-Seng Lung of the University of Virginia. The help from Mr. Ben Bedell is also appreciated.
The measurement of field data is the result of the group effort of many colleagues in Xinjiang.
Their contributions are gratefully acknowledged here.
195
-------
This project was conducted under the US-PRC Agreement to Cooperate in the Field of
Environmental Protection, Annex 3. Dr. Gao Zhizhang, Institute of Environmental Protection,
Urumchi, and Dr. Rosemarie C. Russo, U.S. EPA, Athens, Georgia are PRC and U.S. Project
Leaders.
REFERENCES
Ambrose, R. B., T.A. Wool, J.P. Connolly, and R.W. Schanz, January 1988. WASP4, A
Hydrodynamic and Water Quality Model-Model Theory, User's Manual, and Programmer's
Guide, EPA/600/3-87/039.
Rich, L. G., 1973. Environmental Systems Engineering. McGraw-Hill, New York.
196
-------
EFFECT OF CORRELATED INPUTS ON WATER QUALITY
MODEL UNCERTAINTY
by
Linfield C. Brown1
INTRODUCTION
The role of quantitative uncertainty analysis is assuming increasing importance in water
quality measurement. Decisions affecting appropriate use of stream assimilative capacity and
waste load allocations are increasingly subject to challenge because of questions concerning the
quality of the data and/or the validity of model assumptions used to make forecasts. Method-
ologies for assessing the uncertainties in model simulations of pollutant behavior are useful in
addressing these issues. Although uncertainty analysis has been discussed in recent water quality
modeling literature (Beck 1987, Brown and Barnwell 1987), much of the work reported has
assumed that all input variables and parameters are independent. The significance of
correlation among the inputs on errors in model forecasts has always been presumed to be
important, but only recently has it begun to be closely examined.
This paper is a continuation of the work of Song and Brown (1990) to evaluate the
importance of correlated model input variables and parameters in the study of model output
uncertainty. Two uncertainty analysis techniques are used and compared; first order error
analysis (FOEA) and monte carlo simulation (MCS). The model used for demonstrating the
influence of input variable correlation is an enhanced version of the basic Streeter-Phelps
dissolved oxygen equation. This model was selected because it is well documented in the
literature and because it forms the basis of many of the water quality models currently in use.
Furthermore, the relationships between model inputs and output state variables are well
understood.
DESCRIPTION OF PROTOTYPE MODEL AND PRIOR WORK
The model used in this study'is a modified form of the Streeter-Phelps model (Song and
Brown 1990) that includes the effect of reaeration, gross photosynthesis, algal respiration,
sediment oxygen demand, and a distinction between biochemical oxygen demand (BOD) from
nitrogenous and carbonaceous material (NBOD and CBOD). Distributed sources of CBOD and
NBOD are not considered. The solution for the dissolved oxygen (DO) deficit is given in
Equation 1 (Thomann 1972).
Dt =
+
(1)
Department of Civil Engineering, Tufts University, Medford, Massachusetts, USA.
197
-------
where: D, is the predicted dissolved oxygen (DO) deficit concentration (mg/L) at time t, D0 is
the initial DO deficit concentration (mg/L), k, is the reaeration coefficient (per day), t is travel
time (day), LO is the initial carbonaceous BOD (CBOD) concentration (mg/L), N0 is the initial
nitrogenous BOD (NBOD) concentration (mg/L), 1^ is the CBOD decay coefficient (per day), km
is the NBOD decay coefficient (per day), P is the photosynthetic oxygen production rate (mg/L-
day), R is the algal respiration rate (mg/L-day), and Sb is the benthic oxygen demand rate
(mg/L-day).
Burges and Lettenmaier (1975) and Chadderton et al. (1982) have performed first order
error analysis on the DO deficit from Streeter-Phelps models, without accounting for the effects
of correlation among the input parameters. Song and Brown (1990) showed that the effect of
correlation among the inputs in Equation 1 could increase the variance of the forecast DO
deficit by as much as 20-40%, 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 nonlinearities on the validity of the first order analysis were small compared to the
correlation effects.
UNCERTAINTY ANALYSIS TECHNIQUES
In general, uncertainty in water quality models results from both the use of incorrect
models (lack of fit) and the errors associated with measuring model variables and model
parameters (Burges and Lettenmaier 1975). The so-called measured uncertainties include
variation in replicate laboratory or in situ measurement; spatial variation at a given sampling
location, either across the width of the river or through the water column; temporal variation
over the sampling period; and longitudinal variation over a specified length of river (Notini
1987). In this study, attention is focused on the uncertainty associated with measured inputs,
using both first order error analysis and monte carlo simulation techniques. In first order error
analysis, uncertainties in model input parameters 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 (Benjamin
and Cornell 1970, and NCASI1985), as follows:
E(Y) = f(X)
(2)
V(Y)
v(Xj)
(af/8X.)(3f/axj)cov(xi,xj)
(3)
where: E(Y) and V(Y) denote the expected value and the variance, respectively, of the output
variable; V(Xj) is the variance of the input variable Xg X denotes the mean vector of the input
parameters; and cov(XbXj) represents the covariance between Xi and Xj. All derivatives are
evaluated at X and are the model sensitivities to the corresponding inputs. These model
198
-------
sensitivities for Equation 1 have been computed and are reported by Song and Brown (1990).
The first term in Equation 3 represents the contributions to output variance from the variance
of each input variable acting independently. The second term in Equation 3 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
simulation output values can be analyzed statistically to compute an output error that represents
the combined effects of all model input uncertainties. Several references describe methods for
determining an adequate number of simulations to provide reliable information on model output
frequency distributions. NCASI (1985) used 200, Song and Brown (1990) used 1000, and Notini
(1987) and Burges and Lettenmaier (1975) used 2000. Considering the reliability of the output
results and the computation time requirements, 1000 simulations were chosen for monte carlo
analysis in this study. The multivariate normal distribution was used for all simulations.
EXPERIMENTAL DESIGN
INPUT DATA SET
The input data set is similar to the set of hypothetical stream conditions used by
Chadderton et al. (1982) and Song and Brown (1990). The value of D0 for stream conditions 1,
5, 9, and 13 has been changed from 0.0 to 0.5 mg/L to avoid difficulties in specifying input errors
as a percentage of the mean value, and the values of P, R, and St range from 0.5 to 2.0, 0.3 to
1.2 and 0.1 to 0.4 mg/L-day, respectively (see Table 1). Simulations were performed under 16
different sets of input conditions that are sub-divided into four groups, defined by the self-
purification ratio, f=ka/(kc+km). Because the value of the self-purification ratio is strongly
influenced by the velocity, Chadderton et al. (1982) describe the four f ratios as representing
general groups of streams varying from sluggish (Group 1) to swift (Group 4). Within each
group, four different combinations of CBOD loads, NBOD loads, and initial dissolved oxygen
deficits are used to represent a variety of water quality conditions.
MODEL INPUT UNCERTAINTY
In this study, uncertainty components in the predicted dissolved oxygen deficit (Eq.l) are
considered to involve nine inputs; D0, !»,, N0, km, k,., k« P, R, and S,,. Chadderton et al. (1982)
have shown that the travel time, t, does not significantly influence output uncertainty; thus its
variability was not considered. As indicated in Equation 3, the output uncertainty can be
partitioned into two parts; the variance of each model input and the covariance (correlation)
among pairs of model inputs. The variance of each model input was set by specifying the value
of its coefficient of variation (standard deviation divided by the mean). Two cases were
considered; small error (5%) and large error (25%). Covariance among pairs of model inputs
was expressed by specifying a value for the correlation coefficient.
199
-------
Table 1. Input data set for uncertainty simulations.
Stream KC
Condition —
K.
per day
K,,
Lo
No . Do
mg/L
P
R
So
mg/L-day
te
day
Group 1: f=0.83 sluggish
1 0.08
2 0.08
3 0.08
4 0.08
0.10
0.10
0.10
0.10
0.04
0.04
0.04
0.04
12.0
11.2
10.0
8.00
6.00
5.60
5.00
4.00
050
1.00
2.00
3.00
ZOO
ZOO
ZOO
ZOO
1.20
1.20
1.20
1.20
0.40
0.40
0.40
0.40
7.86
7.04
5.21
1.70
Group 2: f=133 low velocity
5 050
6 050
7 050
8 050
9 1.67
10 1.67
11 1.67
12 1.67
13 333
14 333
15 333
16 333
1.00
1.00
1.00
1.00
5.00
5.00
5.00
5.00
10.0
10.0
10.0
10.0
0.25
0.25
0.25
OJ25
Group
0.83
0.83
0.83
0.83
1.67
1.67
1.67
1.67
16.0
15.2
14.0
12.0
8.00
7.60
7.00
6.00
050
1.00
2.00
3.00
150
150
150
150
0.90
0.90
0.90
0.90
030
030
030
030
138
132
1.18
0.96
3: f=ZOO moderate velocity
20.8
20.0
18.4
16.4
Group 4:
20.8
20.0
18.4
16.4
10.4
10.0
9.20
830
f=ZOO swift
10.4
10.0
9.20
8.20
050
1.00
2.00
3.00
050
1.00
ZOO
3.00
1.00
1.00
1.00
1.00
050
050
050
050
0.60
0.60
0.60
0.60
030
030
030
030
0.20
0.20
0.20
0.20
0.10
0.10
0.10
0.10
033
032
0.29
0.24
0.17
0.16
0.14
0.12
The correlation among the water quality model input variables and parameters has not
been well studied and attempts to locate appropriate field data on which to base quantitative
estimates were met with limited success. Thus, correlations were assigned on a qualitative basis
using two criteria; sign (positive and negative) and strength [strong (0.7-1.0), moderate (0.4-0.6),
weak (0.1-03), or negligible (0.0)]. For example, consider the correlation between ke and l^.
Assigning a correlation coefficient value of 0.6 indicates a moderate positive correlation between
the BOD load and its biochemical oxidation coefficient.
200
-------
Two different correlation matrices were constructed for this study. The first (Table 2)
was designed to represent the input correlation structure for the case of applying the model in a
forecast mode. The second (Table 3) was designed to represent a calibration mode. The
primary differences between these two cases is that in the forecast mode, values of model inputs
are generally assumed, or fixed by regulation (policy); while in the calibration mode, many of the
model inputs are measured. The correlation relationships for these two cases are quite
different.
Table 2. Assumed correlation coefficients among model inputs
(forecast mode).
1 *.
k.
kn
I*
No
Do
p
R
s>
1.0
0.0
0.4
0.6
03
0.0
0.0
0.0
05
k.
1.0
0.0
0.0
0.0
0.0
-05
-0.4
-0.4
k.
1.0
03
0.6
0.0
0.2
0.2
0.2
Lo
1.0
0.8
0.2
03
03
0.4
N0 | Do | P | R | ^
1.0
0.1 1.0
05 0.0 1.0
05 0.0 0.7 1.0
0.4 0.1 0.2 0.4 1.0
Table 3. Assumed correlation coefficients among model inputs
(calibration mode).
k. I k» U I
No
Do
R
k.
kn
No
Do
p
R
S,
1.0
0.0
-03
-0.5
0.2
0.0
03
03
0.0
1.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
0.2 1.0
-05 -0.6
0.0 0.0
-0.1 -0.4
-0.1 -0.4
0.0 0.0
1.0
0.0
0.2
0.2
0.0
1.0
0.0
0.0
0.0
1.0
0.7
-03
1.0
-03
1.0
201
-------
For the forecast mode correlation matrix (Table 2), 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 k<. and L0, k<. and N0> were modeled as
moderately positive. Strong correlations were assumed between L,, and N0, and between algal P
and R. Weak to moderate correlations were assumed among the loads and benthal demand
(e.g., LO and Sb), and between load and rate coefficient (e.g., kc and N0). The correlations
between D0 and all other inputs was assumed to be negligible to weak. The correlations
between k» and algal P and R, were assumed moderate to strong, and negative.
For the calibration mode correlation matrix (Table 3), 17 pairs of inputs were assumed to
be correlated. The major differences from the correlations, in Table 2, are in assessing whether
measurement of these input variables can be achieved independently of each other. For
example, the value of k,, 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 L,,, cannot be
measured independently from the value of the NBOD load, N0, nor independently from rate
coefficients, l^and k, (assuming a standard long-term BOD test is performed to obtain values of
these inputs). Furthermore, because these input values are normally obtained by taking
differences (i.e., CBOD = Total BOD - NBOD) or by regression analysis (i.e., fitting first order
kinetic expressions to obtain kc and !«,), the sign of these correlation coefficients is usually
negative. Thus the calibration correlation matrix has many more negative entries than the
forecast correlation matrix.
Finally, it should be noted that these correlation coefficient assignments are subjective and
are certain to be refined with further study. However, in the absence of appropriate field data,
the two sets of values used in this study (Tables 2 and 3) are believed to be a reasonable
starting point for assessing the range of effects of input parameter correlation on model output
uncertainty.
RESULTS
MODEL SIMULATION OUTPUT
A computer program was written to perform the model simulations and manage the results.
The output of the simulations consisted of the mean values of dissolved oxygen deficit (DOD)
for the 16 stream conditions for 21 time intervals; the standard deviations of DOD for both
correlated and independent inputs; and, for the first order error analysis, the percentage
contributions to the total uncertainty in the predicted DOD from the input variables and
parameters acting independently and in correlated pairs. Most of the tables and the figures
shown in this paper are based on the output data from stream condition number 5 of the input
data set (see Table 1). The results from this simulation condition, in general, are representative
of the results from all 16 of input conditions. Important differences and trends among the four
groups of simulation conditions are discussed in the text.
Figure 1 shows the DOD mean and standard deviation with both independent and
correlated inputs, that results from a first order error analysis. The simulation is for stream
condition number 5 using the forecast correlation matrix in Table 2 and value of 25% error in
202
-------
the input parameters. The addition of correlation to the model inputs results in a standard
deviation of the DOD (Sc) that is larger than that from independent inputs (Si) for travel times
as long as 2.5tc (where tc is the time to the DO sag point). The largest difference between the
standard deviations (Sc and Sj) is about 40% and occurs just prior to the sag point For travel
times longer than t/tc = 2.5, the effect of input correlation is to reduce the DO deficit standard
deviation.
&
8
D DOO
+ DOD - 81
0 COD * II
DOD - Sc
X BOB + «e
TIB* - t/tc
Figure 1. DOD standard deviation for stream condition 5 with correlated and independent
inputs - (FOEA, 25% input error, forecast correlation matrix). :
These findings are similar to those of Song and Brown (1990), who showed that the largest
difference between Sc and S{ was about 33%, and that input correlations reduced DO deficit
standard deviations for travel times longer than t/tc = 3.0. The larger correlation effect prior to
the sag point (40% vs 33%) and the earlier occurrence of the reduction in output variance
(at t/tc = 2.5 rather than 3.0) in this study are attributed to the inclusion of correlations between
P, R, Sb and k,, correlations that were not included by Song and Brown (1990).
EFFECT OF MAGNITUDE OF INPUT ERROR
The importance of correlation among input parameters and variables is strongly influenced
by the overall magnitude of the error in those inputs. Using first order error analysis, the
standard deviation of the DO deficit was computed for both input error levels, using both inde-
pendent and correlated inputs. The results are shown in Figure 2, which displays the simulated
DO deficit standard deviation as a function of travel time. At the 5% input error level, the DO
deficit standard deviations from correlated inputs is hardly distinguishable from those with
correlated inputs. At the 25% input error level, however, the differences are readily apparent.
Figure 2 also highlights the travel times at which the inclusion of input correlation increases and
then decreases the DO deficit standard deviation relative to that with independent inputs.
203
-------
D SI (SSI
+ «e (5*1
(25*>
A SC (2SW
Ti«* - t/tcr
Figure 2. Effect of input error on DOD standard deviation using forecast
correlation matrix - (FOEA, stream condition 5).
The implication to be drawn from Figure 2 is that the larger the error associated with
model inputs, the larger will be the effect of correlation among those inputs on the simulated
DO deficit standard deviation. Similar effects of input error can also be shown when a monte
carlo analysis of these simulations is performed.
The type of uncertainty analysis methodology; i.e., first order error analysis or monte carlo
simulation, is also strongly affected by the magnitude of the input errors. Using the simulation
results with correlated inputs, the standard deviation of the DO deficit was computed for both
input error levels (5% and 25%), for first order and monte carlo methodologies. The results are
shown in Figure 3, which plots the simulated DO deficit standard deviations as a function of
travel time. As was observed in Figure 2, the standard deviations are essentially indistinguish-
able at the 5% input error level. However, at the 25% error level the DO deficit standard
deviations from the monte carlo analysis are larger than those from the first order analysis. The
differences are small prior to the sag point, and appreciable (W to 30%) beyond the sag point.
These differences are attributed to the fact that the monte carlo analysis correctly accounts for
model nonlinearities, which the first order technique ignores. The implications from Figure 3
are twofold. The first is that a first order analysis will always under estimate the DO deficit
standard deviation because of model nonlinearities. The second is consistent with the
observations from Figure 2, namely that the nonlinearity effect is enhanced when the magnitude
of the input errors is large. Thus the disparity between these two uncertainty analysis
techniques becomes greater as the magnitude of the input errors becomes larger. Similar effects
of uncertainty analysis technique can be shown for the case with independent inputs.
204
-------
FOE* (5%)
HCS.(5%)
(25*)
A HCS (25%)
Tl»« - t/tc
Figure 3. Effect of input error on DOD standard deviation for FOEA and MCS
methodologies - (forecast correlation matrix, stream condition 5).
COMPONENTS OF VARIANCE ANALYSIS
The first order error analysis procedure allows the output variance to be partitioned
into contributions from each input variable (Equation 3). The data in Figure 4 show the percent
contribution from the input parameters and correlation to the uncertainty (variance) of the
simulated DO deficit. The simulation results are for stream condition 5, an input error of 25%,
and the input correlation matrix in Table 2. Of the nine input parameters, only five are shown.'
The components from kn, N0, R and Sb are not presented because their magnitude is small
Inspection of Figure 4 shows that for this stream condition the major variance
components prior to the sag point are from k,., L,,, and correlation; while after the sag point, k»
and correlation effects dominate the system. The effect of D0 is felt only at very short travel
times and the contribution from P, is substantial only at long travel times. These results
reinforce those reported by Song and Brown (1990). One difference is that in this study the
effect of Lo is larger, because of its larger input error (25% vs 10%). Also the correlation
component is about 20% larger prior to the sag point, and decreases rapidly as the effect of P
begins to increase (t/tc > 2.0). This effect is attributed to including more terms in the input
correlation matrix (25 vs 11).
First order error analysis also allows a further partitioning of the correlation portion
of the DO deficit variance. The pairs of inputs having the largest contribution for this stream
condition are reported in Table 4. Those pairs of inputs that contain either L,,, It,., P, or k, are
205
-------
Ti»« - t/te
Figure 4. Components of DOD variance for stream condition 5
{FOEA, 25% input error, forecast correlation matrix)
Table 4. Percent contribution to DOD variance from correlation among selected model
inputs*.
Time
t/tc
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
3.0
4.0
Variance Contribution from Correlated
(kJO
0.0
8.7
8.6
8.4
8.1
7.7
73
6.7
6.1
53
43
-7.0
0.9
(L«kc)
0.0
52.1
52.0
50.9
49.5
47.7
45.6
43.2
40.2
363
303
-46.2
-25.5
(NcW
0.0
19.4
21.8
24.4
27.6
315
36.7
43.9
543
71.0
1023
94.6
255
(PX)
0.0
-2.1
-4.2
-6.7
-10.1
-14.6
-20.7
-295
-42.8
-64.8
-107.1
-1725
-77.9
(P,N0)
0.0
-5.7
-6.6
-7.7
-9.1
-10.9
-13.4
-16.8
-21.9
-30.3
-46.2
-56.7
-21.2
Input Pair (%)
(RW
0.0
-2.4
-3.0
-3.6
-4.4
-55
-7.1
-93
-12.8
-18.6
-29.9
-473
-24.7
(R,P)
0.0
1.0
2.0
3.2
4.8
7.0
9.9
14.2
205
31.1
51.4
8Z8
37.4
(S^D.)
0.0
-1.6
-2.0
-2.6
-33
-43
-5.6
-8.0
-11.6
-17.8
-303
-64J
-47.2
•Stream Condition No. 5, Forecast Mode Correlation Matrix, 25% Input Error
206
-------
the major contributors. Negative values mean that the effect of the input pair is to reduce the
DO deficit variance. In Figure 4 it was shown that these inputs are also the important ones
when acting individually (independently). Therefore, in order to adequately model the effects of
input correlation, it appears important that both the variance and covariance among ke, 1L P
and k» be accurately estimated. '
In general, the results in Table 4 show that a pair of input parameters having a high
correlation coefficient (i.e., k, and N0) will not necessarily have a large percentage contribution
to the correlation component of the output variance (Equation 3). The magnitude of their
contribution is determined by the product of three factors: their joint sensitivities on the DO
deficit, their input variances, and their correlation coefficient The effect of a large value of any
one of these factors can be attenuated by small values of the other.
The self-purification ratio and the magnitude of the input variables and parameters
also influence the relative importance of the inputs in the uncertainty analysis. Table 5 contains
the average percentage contributions to the DO deficit variance for the stream conditions in
three of the four groups of streams used in these simulations. The results from Group 4 are not
shown, because they are nearly identical to the results from Group 3. The average results from
Groups 2 and 3 are similar to one another, but clearly different from Group 1. For Group 1
stream conditions (sluggish, low velocity) the travel times are long and thus the zero order algal
and sediment processes dominate the first order processes of CBOD and NBOD oxidation. The
effects of P, R, Sb and their interaction through input correlations are the important components
of DO deficit variance. At all travel times the effect of input parameter correlation is to
decrease (by as much as one-half) the DO deficit variance. This effect is directly attributable to
the strong positive correlation between P and R, and the opposite sign of the sensitivity of the
DO deficit to these variables.
For the stream conditions in Groups 2 and 3 (increasing stream velocities) the effects
of P, R and Sb are negligible compared to contributions from k,, L,,, and k.. The effect of
correlation among input parameters is to increase the DO deficit variance, with the greatest
effect occurring prior to the DO sag point. As travel time increases, the effect of input
parameter correlation decreases to negligible levels for Group 3, and to a large negative value
for Group 2. The percentage contributions to the total uncertainty in the output DO deficit
from correlation is about 30-50% for these groups, over travel times from t/tc = 0 to 2.0.
The results shown in Table 5 more fully explain the findings of Song and Brown (1990) who
by not including the correlations among P, R, Sb and k, concluded that the effect of correlation
among input parameters appeared to become increasingly important as stream velocity
increases. This study convincingly shows that velocity is not necessarily the important factor.
What is important are the sensitivities of DO deficit to the dominant processes affecting DO
deficit. For stream conditions when P, R, and Sb are the major processes (Group 1) inclusion of
the correlation among these inputs is essential to obtain a clear assessment of their effect on
DOD variance. Likewise, for stream conditions (Groups 2 and 3) when first order processes
(e-g-> ka, k,., L,,) dominate, inclusion of their covariance structure is important.
207
-------
Table 5. Average percent contribution to DOD variance from each stream group*.
Time
t/IC
Variance
k*
k.
kn
LO
Contribution from Input
No
Do
P
Parameter (%)
R
s,,
Correlation
Group 1: f=0,83 Sluggish
05
1.0
2.0
4.0
7.4
5.7
2.7
1.0
0.8
15
23
1.0
0.6
0.6
05
0.2
9.8
10.0
84
3.7
0,7
0.8
0.8
05
20.0
10.2
3.4
0.8
68.2
93.9
13Z1
164.6
245
33,8
475
593
2.7
3.8
53
6.6
-34.7
-60.4
-102.4
-137.7
Group 2: f=133 Low Velocity
05
1.0
2.0
4.0
15.6
10.7
2.0
8.6
6.9
17.4
433
74.7
1.4
1.4
1.2
0.2
223
24.2
23.8
17.8
Group 3
05
1.0
2.0
4.0
16.4
115
2.9
53 .
8.9
203
43.4
70.6
13
1.4
1.1
03
21.9
22.6
21.4
17.7
1.6
2.1
3.2
6.5
: f=2.0
1.6
1.9
2.6
5.3
4.9
1.1
0.2
0,0
1-4
2.2
5,0
31.1
05
0.8
1.8
11.2
0.1
0.1
0.2
1.2
45.4
40.0
193
-51.2
Moderate Velocity
3.4
0.6
0.1
0.0
0.0
0.0
0.1
0,4
0.0
0,0
0.0
0.2
0,0
0.0
0.0
0.0
46.4
41.7
28,4
0.4
•Forecast Correlation Maxtrix (Table 2), 25% Input Error
ALTERNATE CORRELATION MATRIX
Two input correlation matrices were described in the Experimental Design section of this
paper. The results presented thus far have been for uncertainty analysis using the forecast
correlation matrix (Table 2). Uncertainty analysis using the alternate correlation matrix
(calibration mode - Table 3) reveals a somewhat different picture of the effect of input
parameter correlation on model output variance. Using the calibration correlation matrix
(Table 3), the DOD standard deviation was computed for both input error levels (5% and 25%),
using both independent and correlated inputs. The results from the first order error analysis are
plotted as a function of travel time in Figure 5, and represent stream condition number 5.
Comparison of the results in Figure 5 with those of Figure 2 confirms the previous finding
that input parameter correlation becomes more important as model input errors become larger.
The simulated standard deviation increases dramatically with the magnitude of the input error,
and the effect of input parameter correlation is substantially larger at the high level of input
error than at the low (5%) level.
There are also important differences between Figures 2 and 5. For the calibration matrix
(Figure 5), the effect of input parameter correlation, for all travel times (up to t/tc = 4.0), is to
reduce the DOD standard deviation, rather than to increase it as was found for the forecast
correlation matrix (Figure 2). The primary reason for this reversal can be traced to the fact that
208
-------
the important input correlations in Table 3 (calibration mode) have opposite signs from those in
Table 2 (forecast mode). That these negative correlations tend to lower output variance should
not be surprising. Overestimation of one input (say L»,) is often offset by underestimation of a
correlated input (say k,.), thus stabilizing error propagation in the model. This finding is
consistent with the reality that, except in rare cases, water quality in riverine systems is normally
less variable than predicted by a mathematical uncertainty analysis using an assumption of
independence among errors in model inputs.
a si (5%>
+ Sc (5%)
O 81 (25%)
A *c (25%)
•1
I
§
3
•H
Tl»» - t/tc
Figure 5. Effect of input error on DOD standard deviation using calibration
correlation matrix - (FOEA, stream condition 5).
The results for the average contribution to the DOD variance for each of the four stream
groups is presented in Table 6. When compared to the results from the forecast matrix in Table
5, it is apparent that the relative magnitude of the contributions from each input have changed.
For Group 1 streams, the effect of correlation is still to reduce the DOD variance, but its
influence is less variable with travel time. For Groups 2 and 3, the strong negative effect of
correlation is countered by larger positive contributions from k,., ka. and L,, (which are the
dominant processes affecting DOD for these stream conditions).
Figure 6 summarizes the comparison between the forecast and calibration mode correlation
matrices and their effect on DOD uncertainty for stream condition 5. The ratio of Sc to Sj is
plotted for both correlation structures as a function of travel time. Prior to the sag point,
correlation among inputs may either increase or decrease the DOD standard deviation by as
much as 25 to 40%. The effect of correlation diminishes at travel times beyond t/tc = 2.0, but
tends to yield a net reduction in the DOD standard deviation.
2139
-------
Table 6. Average percent contribution to DOD variance from each stream group*
Time
t/tc
Variance Contribution from Input Parameter (%)
V
k.
kn
I*
No
D0
P
R
s,,
Correlation
Group 1: f=0.83 Sluggish
05
1.0
2.0
4.0
11.0
73
3.0
1.1
1.1
1.9
2.4
1.1
0.9
0.8
05
0.2
14.7
12.8
85
3.6
1.1
1.0
0.9
05
23.8
12.7
4.2
0.8
101.2
119.2
136,8
149.2
36.4
43.2
493
53.7
4.0
4.8
5.5
6.0
-94.2
-1045
-110.9
-116.2
Group 2: f=133 Low Velocity
05
1.0
2.0
4.0
585
30.7
35
6.1
245
50.2
693
575
5.2
4.2
1.9
0.1
83.9
69.9
38.2
13.9
6.2
6.2
5.1
4.9
143
3.1
03
0.0
5.0
6.2
8.1
23.2
1.8
2.2
2.9
8.4
0.2
0.2
0.3
0.9
-99.6
-72JS
-29.7
-15.1
Group 3: f=2.0 Moderate Velocity
05
1.0
2.0
4.0
61.6
33.1
53
S3.
3Z2
58.8
78.1
73.9
5.1
3.9
2.0
03
82.7
65.4
38.7
18.7
5.9
55
4.7
5.4
10.4
0.6
0.1
0.0
0.1
0.1
0.2
,0.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-97.9
-68.6
-29.2
-4.2.
•Calibration Correlation Matrix (Table 3), 25% Input Error
ror»c«*t Correlation*
•4- Calibration Correlation*
T±m» - t/tc
Figure 6. Effect of correlation matrix on DOD standard deviation
(FOEA, 25% input error, stream condition 5).
210
-------
SUMMARY AND CONCLUSIONS
1. Correlation among input parameters is important in the study of output uncertainty in
Streeter-Phelps type models. Typical forecast mode correlations tend to increase (25-40%),
while calibration mode correlations tend to decrease (20-35%), DO deficit standard deviations
when compared to those using independent inputs. With the exception of streams dominated by
algal and sediment processes, the effect of input parameter correlation is most important in the
vicinity of the DO sag point
2. Correlation among model inputs assumes increasing importance as the magnitude of the
error in model inputs increases. Those processes that dominate the overall DO balance will also
dominate the variance-covariance relationship among model inputs and outputs.
3. First order error analysis consistently underestimates output uncertainty when compared
to monte carlo simulation. This effect is attributed to model nonlinearity. The difference of the
output standard deviations from these two methods is about 10% in the vicinity of the DO sag
point and may be as large as 30% at travel times approaching 4t<.. Although FOEA ignores
model nonlinearity, this methodology is useful because it provides the percentage contribution to
the total uncertainty from every input parameter that has an uncertainty components.
4. Uncertainty analysis is an important and useful tool in assessing the precision of water
quality model predictions. Reliable quantitative information on the variances and covariances
among model inputs is essential for these techniques to be used with confidence. Methods for
rationally determining appropriate input parameter correlation coefficients and variances require
further study.
ACKNOWLEDGEMENT
The author wishes to thank Dr. M. Bruce Beck and Mr. Thomas O. Barnwell, Jr., for their
advice in reviewing the selection of suitable values for the input parameter correlation matrices.
REFERENCES
Beck, M.B. 1987. Water quality modeling: A review of the analysis of uncertainty. Water
Resources Research, 23(8):1393-1442.
Benjamin, J.R., and C. A. Cornell. 1970. Probability, Statistics and Decision for Civil Engineers,
McGraw-Hill, New York, NY.
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,
May 1987, USEPA, Environmental Research Laboratory, Athens, GA 30613.
Surges, S.J., and D.P. Lettenmaier. 1975. Probabilistic Methods in Stream Quality
Management, Water Resources Bulletin, AWRA, 11(1): 115-130.
Chadderton, R.A, AC. Miller, and A.J. McDonnell. 1982. Uncertainty Analysis of Dissolved
Oxygen Model, Journal of the Environmental Engineering Division, ASCE. 108(EE5):1003-
1013.
211
-------
NGASL 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.
Notini, B.R. 1987. Uncertainty Analysis In Water Quality Modeling, MS Thesis, Department of
Civil Engineering, Tufts University, Medford, MA
Song, Q., and L.C. Brown. 1990. DO Model Uncertainty With Correlated Inputs, accepted for
publication, Journal of the Environmental Engineering Division, ASCE December 1990.
Thomann, R.V. 1972. Systems Analysis and Water, Quality Management, Environmental
Research and Applications, New York.
APPENDIX: NOTATION
The following symbols are used in this paper:
cov(*) = covariance
D0 = the initial dissolved oxygen deficit concentration (mg/L)
DI = predicted dissolved oxygen deficit concentration (mg/L)
E(*) = expected value
f = stream purification ratio
ij = subscripts
k, = reaeration coefficient (per day)
ke = CBOD decay coefficient (per day)
km = NBOD decay coefficient (per day)
LO = initial carbonaceous BOD (CBOD) concentration (mg/L)
n = limit on summation
N0 = initial nitrogenous BOD (NBOD) concentration (mg/L)
P = photosynthetic oxygen production rate (mg/L-day)
R = algal respiration rate (mg/L-day)
Si, = benthic oxygen demand rate (mg/L-day)
Sc = standard deviation of DO deficit from correlated parameter inputs (mg/L)
Si = standard deviation of DO deficit from independent parameter inputs (mg/L)
t =s travel time (day)
tc = travel time to DO sag point (day)
V(*) = variance
X = input variable
Y = output variable
212
-------
WATER QUALITY MODELING FOR A TIDAL HARBOR:
APPLICATION TO ZHENJIANG HARBOR, P.R.C.
by
Zhu, Dong-Wei1, Steve C. McCutcheon2,
Zuo, Yu-Hui1, Yue, Zhen-Hua3, Tim A. Wool4
ABSTRACT
The port facilities of Zhenjiang are located in a freshwater tidally-affected harbor on the
south bank of the lower Yangtze River in the People's Republic of China (PRC). The harbor is
on the inside of a very active river bend. Due to the intense sedimentation in recent decades,
the harbor is now separated from the Yangtze River by newly created islands. To control the
sedimentation, the original river channel that was dredged to maintain a shipping canal
connecting the eastern end of the harbor to the Yangtze River was blocked. A new canal was
opened at the western end of the harbor. However, not enough attention was paid to the water
quality problems in relocating the harbor entrance. Water quality was soon unacceptable, and
pollution control became necessary. Plans for reducing pollutant loads and increasing water
exchange were proposed. The hydrodynamic model DYNHYD5 and water quality model
WASP4 were employed to investigate transport and transformation of pollutants in the harbor
under proposed control scenarios. The results indicated that the proposed waste load reduction,
or opening a gate at the closed^end of the harbor, would improve the water quality significantly.
If the two measures could be implemented together, water quality may be expected to reach the
water quality standard of the PRC.
INTRODUCTION
Zhenjiang Harbor is an inland harbor on the south bank of the lower Yangtze River, the
largest river in China. Zhenjiang Harbor is several hundred kilometers from the river mouth.
The harbor is under tidal influence but it is not affected by salinity intrusion. The water in
Yangtze River is extremely turbid and has a deep yellowish brown color from the heavy
sediment loads. The river has been changing its channel significantly in the vicinity of
'Department of Environmental Science, Nanjing University, Nanjing, PRC.
2Center for Exposure Assessment Modeling, Environmental Research Laboratory, Athens,
Georgia, USA
Kfentral Environmental Monitoring Laboratory of Zhenjiang City, PRC.
4AScI Inc., Athens, Georgia, USA.
213
-------
Hgure 1. Map of Zhenjiang Harbor off the Yangtze River showing the old river
channel at the eastern end of the harbor blocked by a dike to control
sedimentation and a new ship canal at the western end.
Zhenjiang Harbor. Within several decades, Jinshan Hill and Jiaoshan Hill (Figure 1)
have changed relative position from the middle of the main river to the present inland location
south of the present river bank. Only an old river channel at the eastern end connected the
harbor to the main Yangtze River in the late 1980*8. This channel required continuous
dredging. To reduce the heavy economic burden of dredging, studies were conducted and a
reconstruction plan proposed and adopted. A new ship canal was opened in November, 1986 at
the western end of the harbor where the river bed is more stable and sedimentation is not as
intense. The original ship canal at the eastern end was blocked after about 2 years.
Zhenjiang is an important industrial city in Southern Jiangsu Province. Paper, food, and
chemical industries pollute the harbor. After the original navigation canal was blocked,
extensive anoxia occurred which affects the aquatic life in the harbor.
Pollution control proposals for Zhenjiang Harbor have been made and are under
evaluation. Mathematical modeling was chosen to aid in planning. To support modeling
studies, hydraulic and water quality monitoring in the harbor were conducted by Zhenjiang
Environmental Protection Agency and Nanjing University. This paper summarizes the modeling
study conducted at the US EPA Environmental Research Laboratory at Athens, Georgia with
Nanjing University.
214
-------
ZHENJIANG HARBOR MODELING SETUP
CHARACTERISTICS OF ZHENJIANG HARBOR
Zhenjiang Harbor is about 10 kilometers long from east to west and as wide as several
kilometers in the wet season (summer). It is influenced by the semi-diurnal tide of the Yangtze
River. Typically, the flooding time of one tidal cycle is about one-third that of the ebbing time.
That is about 3.5 hours flooding and 9 hours ebbing time. A larger tidal range occurs in the
winter and smaller tidal ranges occur in the summer. The average tidal amplitude is 1 m. The
maximum difference is as large as 5 between summer (high) and winter (low). This causes the
water surface area in the harbor to be much larger in the summer than in the winter.
After the original navigation channel was blocked, the harbor became a big pocket with a
narrow neck connecting the Yangtze River. Water exchange was poor at the eastern end away
from the new channel. Water exchange depends not only on the tidal range, but also on the
average tidal height, which controls the water surface area in the harbor. Water exchange with
the Yangtze River is more intensive in the wet season despite the relatively small tidal range.
MODELING REPRESENTATION
In this investigation, near shore mixing of industrial and municipal wastes into the harbor
was not as important as understanding the general water quality conditions related to the
widespread anoxia. Also important was that the harbor is shallow and unstratified. As a result,
a simpler one-dimensional representation using 12 model segments or control volumes was
selected to represent the harbor. As will be seen later in the data presented, this adequately
represented the conditions observed and allowed exploration of all important water quality
management plans.
The water quality kinetics of the harbor were also simplified to the extent possible. Due
to the high turbidity and low levels of phosphorus observed, it was possible to exclude
phytoplankton modeling. As a result, the anoxia model (dissolved oxygen mass balance) was
reduced to four important basic processes: carbonaceous biochemical oxygen demand (CBOD,
decay of organic matter), oxidation of ammonia in sewage and steel production wastes, sediment
oxygen demand (SOD), and reaeration. Mixing of higher dissolved oxygen waters from the
Yangtze was taken into account in simulating the hydrodynamics and transport of the harbor.
MODEL CALIBRATION AND VALIDATION
HYDRODYNAMIC MODEL DYNHYD5
Transport and contaminant mixing was simulated for six occasions for which data were
available for model calibration and validation. There are fours data sets collected in 1987 when
the harbor was open to the Yangtze at both the eastern and western ends, and two sets
collected in 1989 when the old river channel at the eastern end was blocked. The data sets
cover all three important seasons — wet, moderate, and dry.
The model employed was a standard US EPA model, DYNHYD5 that is widely
distributed and publicly available (Ambrose et al. 1988). However, the Yangtze is a very
dynamic river, difficult to simulate because of the long noted abrupt changes due to tidal bores
215
-------
(steep wave moving up river with the incoming tide). As a result, many hydrodynamic models
are subject to some numerical instability that occurs in these simulations, and thus the
simulations of DYNHYD5 were checked with a more stable model developed by the first author
and called NETHYD (Zhu 1990). The tidal conditions were difficult to represent completely,
but after checking with a different numerical approach, and investigating effects of hydraulics on
water quality, it was found that transport and mixing could be adequately simulated.
Tidal conditions in the Yangtze and Zhenjiang do indicate residual tidal bore effects but a
tidal bore did not occur this far inland. Water elevations rise steeply during flood tide of
approximate 3.5 hours and fall very slowly over the remainder of the 12.5 hour tidal cycle during
which ebb tide occurs. It is this rapid rise and slow fall in water surface that can not be fully
handled by explicit numerical schemes like that employed in DYNHYD5. Implicit mathematical
schemes do better for these conditions, so the NETHYD model was used to check the worst
simulations from the DYNHYD5 model.
A time step of 20 seconds was selected to represent the dynamic tidal conditions as best
as possible and to achieve as much numerical stability as possible. The initial water surface was
selected as a flat projection of the initial boundary condition where the harbor joined the river,
and initial velocities in the harbor were selected as zero. Thus an hour of simulation time was
required for the simulations to catch up with actual conditions. This is shown in Figure 2 where
simulated flows begin at zero and actual flows are approximately 400 to 500 m3 s'1.
1987-11-5
upper end
LEGEND
maosured
6.0 9.0
time, hours
3.0 5.0 9.0
time, hours
Figure 2. Comparison of DYNHYD5 simulations of flows (m3 s'1) to
measurements made November 5, 1987. Flows at the upper end were
measured near the entrance of the ship canal on the Yangtze River
and the lower measurements were made near the dike across the old
river channel constructed in 1989.
216
-------
The DYNHYD5 model was calibrated under the most sensitive conditions when both
channels to the harbor were open. Under these conditions the Manning roughness coefficient,
the main calibration parameter, was selected to be 0.03, a value very much typical of these types
of river flow conditions in cohesive muddy banks. When the eastern channel was blocked, the
simulations were not sensitive to the Manning coefficient, and thus its value derived under
critical conditions was used in all simulations since the channel bottom roughness did not
change.
Typical results obtained with DYNHYN5 (shown in Figure 2) are neither the best nor the
worst match with measured data but some effect of numerical oscillations are evident. Figure 3
represents a comparison with the NETHYD model results and shows that oscillations are
dampened but not completely, and that although the DYNHYD model is slightly less stable it
simulates the correct trends in water flow. Further investigation of the water quality parameters
and consideration of the large volume of water involved indicates that the typical divergence of
model simulations, sometimes worsened by a small numerical instability, only affects final results
to a limited degree when the simulated flow and water surface elevations (not shown) have the
right trends and approximately correct magnitudes. Thus, hydrodynamic transport was
adequately simulated. ,
Comparison of DYNHY05 and
NETHYO with
t987-11-5 Data
time, hours
Figure 3. Comparison of simulations from an explicit numerical scheme
(DYNHYD5) and an implicit scheme (NETHYD) using the
November 5, 1987 boundary conditions.
217
-------
DISPERSIVE TRANSPORT WITH THE EUROT4 MODEL
The water quality model subroutine EUTRO4 of the WASP4 model was linked with the
DYNHYD5 model to use the hydrodynamic simulations of advective transport in the harbor.
The linkage was based on running the hydrodynamics models through 90 time steps for each 30
minute time step for the water quality simulation. Water quality conditions change much more
slowly than hydrodynamics and thus a savings in computational effort if realized. However,
calculations done every 30 minutes are still quite intensive and they reasonably approximate
dynamic conditions. As a result of using a short time step (30 minutes), some care was
necessary to select an equation for the dispersion coefficient so that dispersion could be
estimated for real time changes versus tidally averaged changes normally used, ^e therefore
selected the dispersion equations by Tracor (1971) cited by Bowie et al. (1985) for longitudinal
dispersion in unstratified waters:
E =100nUm«R5/6
where,
E
n
U.
R
= longitudinal dispersion coefficient, ft?/sec
= Manning's roughness coefficient
= maximum tidal velocity, ft/sec
= hydraulic radius, ft
The estimated dispersion coefficients for the Zhenjiang Harbor model were on the order
of 2.0 m2/sec, a reasonable value. An initial value of 2.0 m2/sec was chosen for calibration, and
water quality was found to be insensitive to the dispersion coefficient Therefore, 2.0 m2/sec was
used thereafter in all calibrations and simulations.
WATER QUALITY SIMULATIONS
Four of six groups of independent data were used in the calibration and validation of
EUTRO4. These data sets were measured, three in wet season and one in dry season, with
temperatures varying from as low as 5°C up to almost 30?C. The tidal boundary conditions,
measured for one-day periods were repeated to reduce the influence of the initial water quality
conditions and to expand the simulations. Therefore,, it was not realistic to compare the intra-
tfdal predictions with the measured values. Tidally-averaged concentrations simulated by
EUTRO4 were used to calibrate and validate the model, • ' ,
The calibrated CBOD decay rate is:
kd = (0.16) LOS*1"-20) —
where T is temperature. The typical result is given in Figure 4.
For the nitrogen cycle, the sequence of transformation of organic nitrogen is given by:
Organic N fa Ammonia N *» NO2-N' + NO3-N
where, k,, = mineralization rate of organic nitrogen, day1, and km = nitrification rate of
ammonia, day1.
218
-------
EUTR04 Calibration With
Data Measured in 1989-2-28
LEGEND
o = simulated
a = a.m. measured
* = p.m. measured
2.0 3.0 4.0 5.0 6.0 7.0
segment no.
Figure 4. Calibration results comparing simulations with CBOD measured
February 28, 1989 during the morning and evening.
The calibrated rate constants were km = 0.008 x l.OS^-20^ k, = 0.07 x 1.08. Results
are shown in Figure 5. The same rates were used in all simulations.
The reaeration process was treated with the approach included in EUTRO4. It includes
the calculation of friction-induced hydraulic reaeration and wind-induced reaeration. After
comparing these two terms, the larger one was chosen and used.
For the hydraulic reaeration coefficient, EUTRO4 uses the Cbvar method (Ambrose et
al., 1988). For the wind-induced reaeration, EUTRO4 adopts the method by O'Connor (1983).
In most segments, the calculated reaeration coefficients k, were between 0.15 to 0.20 day1. In
some segments, however, k, was occasionally as low as 0.12 day1. Under the monthly averaged
wind speeds, wind was almost always the dominant force of reaeration.
Bowie et al. (1985) reports the common ranges of sediment oxygen demand (SOD). As
the bottom of Zhenjiang Harbor is sandy, an initial value of 0.7 (g O2/m2-day) was used to begin
the calibration and then adopted for the calibrated model. Typical results of dissolved oxygen
simulations are presented in Figure 6.
DESIGN CONDITION
Typical tidal boundary conditions were selected as the design condition, and three
different conditions were considered (Table 1).
After analyzing the results of all three conditions, Winter-1 condition was chosen as the
design condition and used in the sensitivity analysis and scenario simulations.
219
-------
1.0 2.0
EUTR04 Calibration With
Data Measured in 1989-2-2S
LEGEND
o =: simulated.
A = a.m. measured
*"= p.m. measured
4.0 5.0
segment no.
8.0 9.0
Figure 5. Calibration results comparing simulations with ammonia measured
February 28, 1989 during the morning and evening.
EUTRO+ Calibration With
Data Measured fn 1989-2-28
LEGEND
a = simulated
. ^ = a.m. measured
» = p.m. measured
segment no.
Figure 6. Calibration results comparing simulations with dissolved oxygen
measure February 28, 1989 during the morning and evening.
220
-------
Table 1. Design conditions for Zhenjiang Harbor
Mean Water Surface Elevation (m)
Tidal Range (m)
Water Temperature (°C)
Air Temperature (°C)
Wind (m/sec)
Winter-1
1.5
1.0
8.0
5.0
3.4
Winter-2
1.6
0.8
8.0
5.0
3.4
Summer
4.4
0.8
23.0
25.0
3.4
SENSITIVITY ANALYSIS
Sensitivity of water quality to the dispersion process was analyzed by using four different
values of longitudinal dispersion coefficient: 0.0, 2.0, 4.0, and 8.0 m2/sec. The tidally averaged
concentrations were not sensitive to these changes.
The first order decay rate of CBOD was from 0.16 day1 to values of 0.10 and 0.20 day1.
The resulting CBOD and dissolved oxygen concentrations are presented in Figure 7.
Sensitivity lo Kd
1.0 3.0 5.0 7.0 9.0 1.0 3.0 . 5.0 7.0 9.0
Figure 7. Sensitivity of dissolved oxygen and CBOD simulations to changes in the
decay coefficient.
221
-------
The calibrated nitrification rate coefficient km was 0.07 day1.. There were no clear
differences in NH3-N concentrations under mineralization rates of 6.004, 0.008, and 0.012 day1.
Finally, dissolved oxygen is somewhat sensitive to changes in SOD of 0.0, 0.7, and 1.4 g CVm2-day.
EVALUATION OF POLLUTION CONTROL SCENARIOS
POLLUTANT TREATMENT OR DIVERSION UNDER CURRENT COMMTIONS
1. 50% and 80% Total Load Reduction
In order to examine the effects of waste water treatment all the pollutant loads from
industries and domestic sewage were reduced by 50% and 80%, respectively. The results were
compared with the current loading situation in Figure 8, indicating that the water quality in the
harbor can be greatly improved by reducing current loads.
Effects of Pollutant
Reduction
1.0 3.0 5.0 7.0 9,
segment no.
1.0 S.O 5.» 7.0 S.O
segment no.
1.0 3.0 9.0 7.0
segment no.
Figure 8. Effect of reducing current pollution loads for all sources.
2. Collect Pollutants from Four Major Plants
There are several plants that contribute a large percentage of pollutant loads to Zhenjiang
Harbor. The treatment or diversion of these effluents may also result in a big reduction in
pollutant concentrations and increase in dissolved oxygen. Since these sources Could be managed
easier than collecting domestic wastewaters, specific load reductions were evaluated. Therefore,
the effects of collecting all pollutants from the following four plants, Titanium Dioxide Plant,
Dadong Paper Mill, Coke Chemicals Plant, and Printing and Dyeing Mill, were evaluated. The
results shown in Figure 9 demonstrate that CBQD, dissolved oxygen, and ammonia concentrations
were all improved considerably.
222
-------
Eff.ct. of Collecting
Pollutants from « Plants
D»
|
LEGEND
o = full load
i = 4 plant cut
1.0 3.0 5.0 7.0 9.0 1.0 3.0 5.0 7.0 9.0
segment no. segment no.
LEGEND
o = full load
* = •* plant cut
.'"*—*"A--A
1.0 3.0 S.O 7.0
segment no.
Figure 9. Effect of diverting or treating loads from the four major industrial plants
discharging into Zhenjiang Harbor.
EFFECTS OF GATE AT CLOSED END
1. Estimating Outflow Through Gate
A daily average outflow of 27 m3/sec was given in the proposal to open a gate through the
closed end of Zhenjiang Harbor. However, no design data were available to incorporate the effect
of tides on gate discharge. The operational method proposed was to close the gate during flooding
tide, and open when the tide begins to go down, so that water keeps going in one direction. The
following analysis was made: assume zero flow during flooding time and a constant outflow during
ebbing period. In order to encompass all tide variable discharge situations, four different outflows
were considered: 10, 20, 30, and 40 m3/sec during the ebbing period.
2. Effects Under Current Situation
The effects of the tide gate operation under the current pollutant loads were analyzed. The
results are presented in Figure 10. An apparent improvement can be achieved if the flood tide
average outflow can reach 30 m3/sec or higher, for both pollutant and dissolved oxygen
concentrations.
3. Results of 50% and 80% Load Reduction
How much improvement can be achieved through a combined measure of opening the gate
and reducing loads was also addressed. To evaluate this scenario, 30 m3/sec open gate outflow was
used with both the 50% and 80% load reduction. The results are shown in Figure 11 in
223
-------
Effects of Row
Through th« Got*
|
LEGEND
o=0 cms
a = 10 cms
+ =20 cms
x = 30 cms
o = 40 cms
i.o
segi
5.0 7.0
nent no.
3,0 5.0 7.0
segment no.
3.0 5.0 7.0'
segment no.
Figure 10. Effects of tide gate flow rates during flood tide on water quality in
Zhenjiang Harbor.
comparison with the current situation. The water quality was further improved through this
combined measure.4. Percent Reduction Needed to Reach 4 mg/1 CBOD
Load reduction to bring the maximum daily average CBOD concentration down to 4 mg/1
(water quality would meet the 3rd class of the Surface Water Quality Standard of the PRC),
assuming the gate was in operation, was also investigated- From Figure 11, it can be extrapolated
that about 65% to 70% of CBOD load reduction would reach this value under the 30 m3/see gate
outflow condition.
TOXICITY OF AMMONIA AND PHENOL
The concentration of ammonia is rather high in the harbor. This may cause a toxicity
problem due to un-ionized ammonia (NH3). The relative concentration of un-ionized ammonia is
affected by pH and temperature. The method of calculating un-ionized ammonia is described in
Bowie et al. (1985). In Zhenjiang Harbor, the pH is usually around 7.5 and temperature from 5 to
25°C. Under these conditions, the un-ionized ammonia can be as high as 1%. The typical high
ammonia concentration in winter is 2 to 3 mg/1 in Zhenjiang Harbor, this gives 0.02 to 0.03 mg/1
un-ionized ammonia.
The total and un-ionized ammonia in the U.S. Water Quality Criteria (USEPA, 1986) are
15.7 and 0.092 mg/1 for acute toxicity; 3.9 and 0.022 mg/1 for chronic toxicity. Therefore, the
ammonia toxicity is not a serious problem in Zhenjiang Harbor.
224
-------
Effects of toad Reduction
and Outflow
a
Oo
82
LEGEND
o = no cut. no gate
a = SO0; cut. gate
* = 80" cut. gate
I
3.0 5.0 7.0 9.0
segment no.
LEGEND
0 = no cut, no gate
i = 50% cut, gate
* = 80% cut, gate
1.0 3.0 5.0 7.0 9.0
segment no.
1.0 3.0 5.0 7.0 9.0
segment no.
Figure 11. Combined ^effect of tide gate releasing 30 m V on average during flood
tide and 50% and 80% load reductions.
'ati°m k ****$*»* Harbor ">*»' unfavorable conditions are lower
75" f°r ^^ t03ddty' 3S *« b? the U'
1986). No further analysis was made.
1)
2)
3)
CONCLUSIONS
A reduction of 50% or 80% in pollutant loads, through pollutant treatment or diversion
would be expected to reduce the concentrations of CBOD and ammonia by JmpIraWe
amounts and increase DO levels bv about 2 and 4 ma/T. rfor 50% and 80% reduction,
lower
Opening a gate on the dike blocking the lower e5d of the harbor has the potential of
improvmg the water quality significantly. If coabined with pollutant collection, the water
quality would be expected to meet or exceed certain water quality standard (3rd class of
the
225
-------
ACKNOWLEDGMENT
Mr. Robert B. Ambrose Jr., -Manager of the EPA Center for Exposure Assessment
Modeling provided technical guidance to the modeling work. Valuable comments and advice were
also given by Dr. James L. Martin of AScI Corporation at ERL-Athens and Dr. loannis Tsiros of
Hydraulics Laboratory, Agricultural University of Athens, Greece,, during the report review. The
authors also thank Mr. Kit McCormick and Mr. Ralph Power of Computer Sciences Corporation
at ERL-Athens for .graphics support during the modeling study.
REFERENCES
Ambrose, R.B., T.A. Wool, J.P. Connolly, and R.W. Schanz. 1988. WASP4, A Hydrodynamic amd
Water Quality Model-Model Theory, User's Manual, and Programmer* Guide, EPA/600/3-
87/039.
Bowie, GJ., W.B. Mills, et al. 1985. Rates, Constants, and Kinetics Formulations in Surface
Water Quality Modeling, EPA/600/3-85/040,
O'Connor, DJ. 1983. Wind Effects On Gas-Liquid Transfer Coefficients, ASCE 3. of
Environmental Engineering, 109(3), June, 1983.
Tracer, Inc. 1971. Estuary Modeling: An Assessment For Water Quality, Office of US
Environmental Protection Agency, Project 16070DZV.
USEPA. 1986. Quality Criteria For Water 1986, Office of Regulations and Standards,
Washington, DC. EPA 440/5-86-001.
Zhu DW., etal. 1990. NETHYD-A Hydrodynamic Program For One Dimensional Channels or
' liter-Connected Channel Networks, USER'S MANUAL. Unpublished report, US EPA
Center for Exposure Assessment, Athens, GA.
226
-------
WATER RESOURCE MANAGEMENT STRATEGIES FOR
RESTORING AND MAINTAINING AQUATIC LIFE USES
by
Thomas Willingham1 and Allen Medine2
INTRODUCTION
^^^^^^^^^^^^f^^'
^^^^^V^f^^^^^^: ^«-*->- • **»
<° "
Arkfln A C°mprfei?sive Water Qua% Assessment Methodology has been initiated in the
Arkansas River basin to permit the development of a long-ternTmonitoring program establish
g ^ ^Tf58 °f DeCeSSaiy remedial actions' and aUow *e dSopmlnt of a
numeric model for water quantity and quality management. This integrated
rpSOU-Ce mana*eTnt ^ nCeded t0 mana*e the mo^oring ^ restorStofactivitie
fra , n f T8 ,aPProacuhes to river basi" management have often resulted in a
fragmented understanding of the identified environmental problems.
It must be recognized that the Arkansas River is a very complex resource This
complexity is due to the natural flows, inter- and intra-basin water transfers, contaminant
loadings of various types and magnitudes, and the difficulty in perceiving and quantifying the
response of the system to management activities. The major goal of the methodolo^to
tte^*nt-qUaTatl7 I6™8 &?t™th aCCCptable unce*ain*, ^e results of manipSation of
the interacting physical-chemical-biological processes on the overall quality of the resource and
to ensure that these manipulations of the system, whether physical, chemL, or bi^aT'
sPtSs oAhten gnat1-TK ?IS Papef deSCrfbeS this meth°dology, ^ particular thermit al
steps of the process, which has been implemented for the Arkansas River-Initiative.
'U-S. Environmental Protection Agency, Denver, Colorado, USA
2Water Science, Boulder, Colorado, USA.
227
-------
ARKANSAS RIVER INITIATIVE
Water quality of the upper Arkansas River basin has been impacted for well over a
century due to post-settlement activities, notably mining, milling and smelting, urbanization,
farming, and other contemporary activities. Water quality impacts due to mining and processing
activities have been especially acute in a number of portions of the basin, including areas such
as the Leadville mining district, the Cripple Creek district, Chalk Creek and others. Toxurmetal
releases from mine drainage (acidic and non-acidic), tailings ponds, smelter waste, contaminated
soils, natural weathering, and municipal wastewater facilities contribute to the severe
degradation of water quality and resource use. Non-point (diffuse) sources of toxic and
conventional pollutants contribute to degradation of water quality on a local scale but the
primary threat to aquatic life, within the major tributaries and mainstem in the upper basin
above Salida, remains toxic metal contamination.
In order to provide guidance for the range of technical activities required for successful
restoration of the basin aquatic life uses, a multi-phase, multi-agency program has been
developed which spans the 10- to 20-year period estimated for restoration. This program, the
Arkansas River Initiative, contains the following elements:
I. Policy, Planning & Project Definition
n. Management & Technical Problem Definition
HI. Implementation of Acute Toxicity Reduction
IV. Chronic Toxicity Reduction & Goal Re-Evaluation
V. Assessment of Optimal Biological Conditions
VI. Final Restoration & Long-Term Management
The first element involved the preparation of necessary initiative documents which outline
the overall program and serve as the blueprint for establishing the long-term funding to support
the project. Additionally, due to the extensive nature of the project, the estimated funding
needs and the public/private interests in the basin, the establishment of multi-agency workgroups
was essential to ensure a cooperative basis for restoration activities. The major agencies
involved in the project include the U.S. Environmental Protection Agency, the U.S. Bureau of
Reclamation, the U.S. Geological Survey, and various State of Colorado agencies including the
Department of Health and the Division of Wildlife. To support the Arkansas River Initiative
and guide the technical activities, a comprehensive Water Quality Assessment Methodology
(WQAM) has been initiated in the Arkansas River basin to address water quality
and resource use impairment from the Pueblo Reservoir to the headwaters (Figure 1), defined
as the upper Arkansas River basin.
WATER QUALITY ASSESSMENT METHODOLOGY
The WQAM involves a progression of data collection, analysis, interpretation, decision-
making, and restoration activities to gradually direct the project to the attainable goals as
defined throughout the process. This process has been developed and formalized through the
interaction of many professionals in the water quality field, including environmental chemists,
biologists, environmental engineers, water resource engineers, hydrologists, ecologists and many
others. Only through a well-executed, multi-disciplinary program of cooperation among the
228
-------
.MOO
Figure 1. Arkansas River basin map showing water quality monitoring locations for
historic data. Inset diagrams the Upper Basin Management Unit and the
Leadville, Colorado area.
229
-------
various interested parties will the goal of successful, timely and cost-effective restoration of the
upper Arkansas River basin become reality. The WQAM provides the "blueprint" for this
cooperation, while the conceptual model, as described herein, provides the focus for subsequent
technical activities.
The WQAM, as established for the Upper Arkansas River Basin Project (Figure 2), is
essentially a generic process which is well-suited for any water quality, water quantity, and
resource restoration project. It develops site-specific attributes by continually focusing on
activities within the basin, on data needs based on interpretative complexity, on establishment
and re-evaluation of goals for .restoration, and on the very-real economic limits to restoration
expenditures. The process can be as simple or as complicated as dictated by the magnitude of
the environmental disturbance and/or as desired by the project directors.
WATER QUALITY ASSESSMENT METHODOLOGY
STEP 1
STEP 2
STEP 3
STEP 4
STEP 5
STEP
STEP 7
STEP 8
DEFINE EHUIROUIEHTilL SYSTEM
GENERAL STATEMENT Of .GOALS
| DATA COMPILATION |
ASSESS POIENTIALW ATTAINABLE
OR UNDISTURBED CONDITIONS
RE-EUflUWTE 6QALS
LINK CONTAMINANT MfNAMI.CS
10 RECEPTOR EXPOSURE
AND RESOURCE USE CONSTRAINTS
RESOURCE RESTORATION
ASSESSMENT AM) CONTROL
PROCESS IMPLEMENTATION
GOALS ATTAINED
MAINTENANCE MONITORING
Figure 2. Water Quality Assessment Methodology (WQAM) Process
showing the 8 major steps.
230
-------
STEP1: DEFINE ENVIRONMENTAL SYSTEM AND GOALS
h««'n °nn PrOCeSS im°^a the identification °f current and beneficial uses within the
basin and the overall statement of goals. The major activities for this step include the following:
1. Identify Current/Future Resource Use
2. Isolate Management Units
3. Identify Non-Attainment of Use
4. Develop Overall Project Workplan
Resource use in the upper Arkansas River basin includes a combination of aquatic life
irrigation, and water supply uses with the potential for allocation conflicts. Water resources in
this region are crucial to the economic well-being of the area and also to the state of Colorado
Water resource management projects developed to address the needs of agricultural uses u
tnal uses, and municipal uses testify to the importance of this resource. Recreational uses
^ of aquatic «* are
or no Jt * SntidPatf J? the resource use from a wa^r consumption standpoint (consumptive
or non-consumptive) will increase and it is certainly possible that the competition for the
resource and the need for legal, equitable, and timely allocation of the resource may intensify.
It is anticipated that as improvements in water quality are realized from control on point and/or
diffuse contaminant loadings, recreational use will increase as well. Some of these competing
uses (eg water quantity/diversion and water quality and fisheries protection) may lead to
difficult decision-making for the resource manager. By understanding these resource needs and
the effects of resource consumption on resource quality, the manager is in a better position to
regulate a complex resource such as the upper Arkansas River basin. It is fully intended as
part of the site-specific application of the WQAM, that these competing needs are considered in
the overall basin restoration strategy. History has shown that resource allocation decisions are
altemateS " eC°n°micaIly mot™ated ™thout consideration of the resulting impacts to
This initial step also involved the identification of tentative management sub-basins to
divide the system into smaller segments for easier management This sub-division has been
based on chemical, physical and biological attributes of the resource and is somewhat qualitative
ine upper Arkansas River basin management units includes the following:
1. Upper Basin above Lake Fork
2. Lake Fork to Salida
3. Salida to Canon City
4. Canon City to Pueblo Reservoir
5. Pueblo Reservoir.
The first management unit comprises the upper headwaters around Leadville which are
severely degraded from the presence of toxic metals, including zinc, cadmium, lead, copper, and
others. Water quality has been shown to be acutely toxic to aquatic life. By far the most signi-
ficant contaminant loading occurs from the California Gulch, Leadville Mine Drainage Tunnel
231
-------
and the Tennessee Creek (St. Kevin Gulch) areas. As previously mentioned, diffuse or non-
point loadings are also significant in certain portions of this first unit.
The inflows from Lake Fork and Lake Creek generally tend to improve water quality
through dilutional processes and are largely responsible for defining the segment from Lake
Fork to Salida as the second management unit. Metals and suspended sediment loads (and
benthic impacts) contribute to chronic toxicity in portions of the second unit. The segment from
Salida to the Pueblo Reservoir has been divided into two units, Salida to Canon City (third unit)
and Canon City to Pueblo Reservoir (fourth unit), due to gradient and geologic variations in the
segment Below Canon City, the river is erosive in nature and the highly soluble calcium and
manganese minerals result in major changes in major ion chemistry.
Finally, the Pueblo Reservoir has been designated the last management unit due to the
differences in the behavior of contaminants and conventional pollutants in reservoir
environments as opposed to riverine environments. In addition, the biological and chemical
conditions which develop in Pueblo Reservoir on an annual basis are much different than in the
upstream flowing segments.
Problems have been identified in each area that are generally related to toxic metal
discharges, sediment loading, and other more conventional pollutants. The extent to which
metals are impacting water quality is reflected in the approximately 300 river miles (483 km) of
the basin and tributaries which exceed aquatic life protection standards for zinc, cadmium,
copper, and lead, and the fact that 50-100 miles (81-161 km) of the river exceed agricultural and
water supply standards (primarily manganese, lead, and zinc).
From this initial evaluation, the previously discussed 20-year program for monitoring and
restoration of the basin has been developed and will be used as a guide in the early project
stages. As more information becomes available through continued monitoring and interpreta-
tion, this plan will be modified to reflect changes in the goals for resource restoration or
progress toward the desired environmental quality.
STEP 2: DATA COMPILATION AND CONCEPTUAL MODEL
Data Compilation and Review
The conceptual model process involved the compilation of existing information on the
physical, chemical, and biological attributes of the system, initial interpretation of the informa-
tion, and modification of the project goals as necessary. The goal of this step was to support the
interpretation used to develop'the Conceptual Model of the Upper Arkansas River Basin. The
conceptual model permits the delineation of individual sub-basin problems, specific contaminant
concerns, long-range contaminant transport and provides required input into a focused,
scientifically-based monitoring program to be developed as part of Step Three of the WQAM
process.
This step includes a detailed review of existing data for water quality, the hydrolpgic
system, the resource usage, and the current response of biological populations. The critical flow
constraints are identified along with estimates for water quality needs based on a tentative,
use-attainability analysis. Using mass balance, loading, and mass transport analyses in conjunc-
232
-------
ton with the available toxicity data will lead to a determination of the critical pollutants
Management limits are identified for acceptable general water quality, for the "bioavailable"
forms of critical pollutants, for the acceptable range in water quantity, and for water transfer
timing based on the consideration of use: aquatic life, drinking, agriculture, industry and
recreation. J
For example, with respect to metal toxicity in the upper basin, it is intended that these
initial research and monitoring activities will develop the site-specific factors that affect the
available forms and the subsequent biological response which could be anticipated from
restoration. The conceptual model would then be utilized to develop a monitoring strategy to
promote site-specific contaminant regulation based on the biological and chemical monitorine
and integrate dominant mechanisms which regulate the metal forms. In this manner the data
collection activities will be based on our current understanding of contaminant dynamics
biological impacts, and mass transfers in conjunction with water quantity constraints and'other
use constraints in the Arkansas River basin.
The Arkansas River basin from the headwaters to the Pueblo Reservoir has received
considerable study and routine monitoring for a variety of issues. The data collection activities
include monitoring chemical, physical, and/or biological conditions in tributaries and mainstem
stations. Some of these studies have addressed the following issues:
Metal pollution problems in the upper basin,
General water resources appraisals,
Water quality relationships to streamflow,
Impoundment water quality effects at Pueblo Reservoir,
Statistical analysis of streamflow characteristics,
Development of dissolved solids accounting, models,
Tributary or local contamination problems, and
Chemical transformations and dynamics.
results of each of these studies have provided an understanding of specific portions
ot the basin or specific aspects of the hydrologic, chemical, or biological systems, there has been
no direct integration of this information or development of a basin-wide strategy for understand-
ing the complex nature of contaminant source, transformations, transport, and impact The
Arkansas River Initiative has been developed to begin this process and the conceptual model as
described in this document provided a mechanism for guiding future data collection, interpreta-
tive needs and selection of target sub-basins for more immediate restoration action.
Results of this analysis confirm that the most important metals with respect to the aquatic
resources include zinc, cadmium, lead, and copper. Manganese is important from an agricultur-
al supply perspective and in certain tributaries, while iron is important for two major reasons-
habitat destruction due to bulk chemical precipitation and chemical attenuation of toxic metals
through sorption. The transformation of metals in the upper basin from initially dissolved metal
to particulate metal and subsequent physical transport properties and erosional processes during
runoff indicates that the river shows a series of transitional zones which increase the complexity
of the interpretations. Very precise mass balances will be necessary to observe and document
additional loadings in the lower basin which are not related to instream resuspension processes.
233
-------
Overall there are dramatic changes in the major ion chemistry from the headwaters to
the Pueblo Reservoir. In the upper basin, the chemistry is dominated by iron, manganese, and
sulfate originating from the major mining areas. In the lower basin the water is a calcium
bicarbonate type which reflects intermittent high concentrations of iron, manganese, and zinc
from upstream sources. Mean conductivity at Malta (AR52) is 190^mhos/cm and this increases
slightly to about 240 jumbos at Canon City. Between Canon City and Portland, the .conductivi-
ties increase abruptly due to the change from a granitic to a sedimentary geologic environment.
In this transition, the river decreases in elevation from 9600 feet (above mean sea level) to
approximately 4600 feet at Pueblo. The river profile and slope reveals a steep gradient in the
upper basin and the lower gradient below Canon City. The river slope is important not only for
stream velocity considerations but also in establishing depositional environments which attect
contaminant transport on an annual basis.
Dissolved anions (chloride and sulfate) show gradually increasing concentrations except
for the higher sulfate loading in the upper basin mining districts. Calcium, magnesium, and
sodium concentrations are relatively constant from the East Fork to Canon City and averaged 33
mg/L, 8.6 mg/L, and 6.9 mg/L, respectively. Between Canon City and Portland, the concentra-
tions of these major ions doubled, once again illustrating the transition zone which exists below
Canon City. Total and dissolved concentrations of calcium, magnesium, and sodium are
essentially equal at all sample locations.
It was observed that the mass loading of iron into the system in the upper basin coupled
with additional sediment influx from erosional processes along the mainstem results in a
significant opportunity for sorption of pollutants. The available data indicate that there are
significant decreases of both dissolved and particulate metals in the Arkansas River in the mid
and lower portion of the basin (lower management units). This will be further discussed in •
subsequent sections.
Biological Conditions
There is no doubt that the aquatic resources of the Arkansas River system have been
severely degraded due to metal contamination. Acute toxicity is found in the upper basin
management unit and in a number of tributaries. Chronic toxicity has also resulted in decreased
diversity, limited reproduction, and altered populations. Studies of benthic populations have
shown significant impacts below point sources with varying degrees of recovery based on metal
transformation and dilution processes.
Studies of the brown trout (Salmo trutta) population of the upper Arkansas River by the
Colorado Division of Wildlife beginning in 1981 have indicated that rapid growth is followed by
poor survival beyond year three. The major contaminants responsible for the aquatic resource
impacts include zinc and cadmium, and to a lesser degree lead and copper. Detailed investiga-
tions must be directed at biological phenomena in the river system including the fish food
supply, metal stress, forage fish populations, macroinvertebrate growth, penphyton, and the role
of wetlands in attenuating metals.
:234
-------
It is generally agreed that the aquatic resources of the river have been degraded due to
acute and chrome toxicity in the upper basin above Salida. Other tributaries with high metal
concentrations show similar degradation. .Based on the observed chemistry, the degradation is
due primarily to zinc and cadmium concentrations with localized areas affected by lead coooer
other metals, low pH, and bulk precipitation of iron. ' '
A toxicity profile from the East Fork to Salida, conducted by the U. S. Environmental
Protection Agency in September 1987, has documented the observed acute toxicity in the area
above Lake Creek and shown chronic toxicity as far as Chalk Creek, roughly 80 kilometers
Toxicity is most likely due to the independent and/or combined effects of zinc and cadmium.
For that same data set, the relation of the toxicity of the mainstem stations to Cerio-
daphnia is strongly related to the estimate of free ion activity {Zn2*}, as calculated by
MINTEQA2. This response indicated that the dissolved chemistry and, in particular the free
metal can be used as a controlling, variable to relate water quality as well as metal transforma-
tions (sorption, complexation, precipitation) to anticipated biological response. This relation can
be used to develop a site-specific approach to establishing acceptable water quality.
Conceptual Model for Contaminant Behavior in the Basin
While the conceptual model is based on the analysis of existing data, it is not the detailed
analysis which begins at the fourth step in the application of the WQAM process, as shown in
figure 2. Rather, the analysis conducted for the conceptual model provides an integrated
interpretation of our current understanding of the nature and extent of water qualitf problems
aquatic resource impacts, contaminant source areas, contaminant transformations and contami-
nant transport. The development of the conceptual model also leads to recommendations
concerning additional data needs concerning future restoration and the formal identification of
Data Quality Objectives (DQOs).
Upper Basin above Lake Fork. Contaminant loadings have been identified in this area
from a variety of studies and monitoring programs. Mass balance analysis of the loadings has
identified the relative contributions of major sources to the overall metal load measured at the
downstream location, designated as AR51 (Figure 3); 70% of the zinc loading from California
Oulch, 17% from the Leadville Mine Drainage Tunnel, and 7% from Tennessee Creek (St
Kevin Gulch). It must be emphasized that, upon entry of point sources or non-point sources
into the river, there are immediate transformation processes which begin to remove metals from
solution, primarily through sorption to iron oxyhydroxides and natural sediments. Under these
scenarios, the total loading to the system is difficult to define.
Contaminant loadings for zinc were used to illustrate the overall metal contamination
situation in the upper management unit. While cadmium is the other major target contaminant
detection limits and the sensitivity of the analytical data prevent the use of cadmium in
constructing mass balances as accurate as zinc balances. Since zinc is attenuated to a lesser
degree than cadmium, lead, or copper, its use in assessing the degree to which metals are lost
from the water column will be conservative, e.g., other metals will be lost to a greater degree.
235
-------
UPPER BASIN SOURCES OF CONTAMINANTS
TENNESSEE CREEK
~ 7%
LAKE FORK
~ 6%
PERCENT OF TOTAL LOADING AT AR51
ZINC LOAD ~ 170 kg/day (1988-1989)
Figure 3 Summary of zinc mass loading in the upper basin illustrating the impor-
tance of California Gulch (Station CG01 in Figure 1) in regulating chemi-
cal conditions in the Arkansas River. The East Fork and mine drainage
from the Leadville Mine Drainage Tunnel (Station EF01 and LD01 in
Figure 1) are the second-most significant source of toxic metals.
California Gulch. There is no doubt that the single most significant and damaging metal
loading source is due to the California Gulch tributary. Anthropogenic metal loadings in the
California Gulch basin are associated with direct mine discharge (i.e. Yak Tunnel), diffuse
loadings from tailings waste rock and smelter wastes, the alluvial groundwater system, and, to a
lesser degree, natural contaminant loadings. While the point loads may be important under the
current loading situation, following initial point source reductions, diffuse loadings will most
likely become important with respect to the residual loading. Diffuse loadings are also more
important during dynamic, precipitation runoff events.
It should be pointed out that there is most likely a widespread areal surficial soils and
floodplain contamination from previous smelter emissions and blowing tailings. This contamina-
tion may result in significant surface runoff loadings during these dynamic flow events. This
diffuse loading may be more significant on the steeper slopes (south bank of California Gulch)
in the vicinity of the Arkansas Valley smelter and other smelters. Aerial photography reveals
laree expanses of unvegetated slopes characteristic of smelter emission impacts. Data on sods
and erosional transport of contaminants should be thoroughly reviewed to confirm or deny this
potential loading. Other areas of significant disturbance are also evident in the area primarily to
236
-------
the south of California Gulch but also on the north bank in the lower basin around Stringtown
and at the confluence with the Arkansas River.
Chemical Transformations. Studies by researchers from the U.S. Geological Survey in the
upper basin are currently underway to describe the chemical transformations which regulate the
transport and distribution of metals. It is generally accepted that the role of iron oxyhydroxides
are very important and probably the dominant process regulating the dissolved concentrations of
zinc, cadmium, lead, and copper. Studies are currently underway to illustrate the significance of
photoreduction of iron and the effect on metal transport and transformations in St. Kevin Gulch,
a tributary of Tennessee Creek.
Chemical precipitation of iron and aluminum is very rapid in both tributary and mainstem
waters and will contribute to the sorption properties of suspended and bed sediments. Metal
concentrations in the bed sediment region support this process. Of particular importance will be
the change in the sediment sorption characteristics as restoration activities are implemented
within the basin and reduce the total mass loading of iron, aluminum, and manganese. Contin-
ued study of the role of particulates, photoreduction, and chemical variables (pH, organics,
competing ions) will be necessary to understand fully the contaminant behavior under a variety
of restoration scenarios.
Contaminant Fate. It appears that wetland areas in the mainstem and tributaries are an
important repository for metals in addition to the mainstem and tributary sediments region
during depositional flow periods. While the wetlands may be perceived as a more permanent
fate for contaminants, the bed sediment regions are more temporary and become sources of
instream contaminants during changes in the hydrologic conditions. Study of the dynamics of
sedimentation and resuspension (scour) along with establishment of the critical scour velocities
will be important in describing contaminant transport in this portion of the basin and in other
parts of the Arkansas River.
Overall Conceptual Model. Based on the observed data concerning the sources of metals in
the Arkansas River basin, the likely transformation processes, metal transport and observed
impact, the conceptual model for the Arkansas River basin has been presented in Figure 4. The
upper basin is characterized by major metal source loadings and a complex set of transformation
processes. These chemical processes which occur in the upper basin include generation of acidic
waters, mixing of acid and neutral surface waters, neutralization, and chemical precipitation of
iron and aluminum. Toxic metals are attenuated through dilution and sorption to precipitated
iron, aluminum, natural sediments (clays, organics), and wetlands substrate. Metal toxicity is
directly related to the metal speciation which is affected by pH, major ions, organics, and
organism-specific factors. These metal processes work in concert to result in the dissolved metal
conditions which regulate the biological populations. Factors which affect the form of the metal
in solution and the overall metal toxicity are presented in Figure 5.
In the middle basin, the system is characterized as showing chronic toxicity, receiving
additional metal loads, and an overall trend for decreased metal concentrations. Sedimentation
and resuspension in relation to the hydrologic system appears to be largely responsible for,the
dynamic changes in total metal chemistry during runoff. Continued sorption of dissolved metals
will maintain much lower concentrations of metals than in the upper basin. Metal removal
237
-------
UPPER BASIN
METAL SOURCES
MIXING
NEUTRALIZATION
CHEMICAL PRECIPITATION
IRON. ALUMINUM
SORPTION OF METALS
Pb.Cu.Cd.Zn
SPECIATION
ACUTE TOXICITY
MIDDLE BASIN
EROSIONAL TRANSPORT
ADDITIONAL LOADS
SORPTION
CHRONIC TOXICITY
PUEBLO R.ESERVOIR
Overflow
Long-Tern
Burial of Sedinentf
Figure 4. Conceptual model for contaminate source, transformations, transport and
fate in the Arkansas River. The conceptual model will guide the subse-
quent activities for resource restoration in the basin.
METAL TRANSFORMATION PROCESSES REGULATE
THE FORM OF METAL IN ENVIRONMENTAL SYSTEMS
METAL TOXICITY REGULATION
SITE - SPECIFIC ENVIRONMENTAL FACTORS
CONTAMINANT FORM Inorganic, Organic
Soluble, Particulat*
Ion, Complex, Oxidized/Reduced
Colloidal, Precipitated, Sorbed
MULTIPLE TOXICANTS Joint Action, No Interaction,
Antagonism
ENVIRON. FACTORS Temperature, pH, Dissolved 02>
Light, Salinity
ORGANISM CONDITION Age, Size, Life Stage, Activity,
Sex, Starvation, Adaptation, and
Change in Life Cycle
Certain of these factors may have wide spatial &
temporal variation. The type of environment
is also Important in assessing these factors.
Figure 5. Site-specific environmental factors which regulate metal transformations
and aquatic toxicity in surface water environments.
238
-------
mechamsms^gure 6) indicate that, at the conditions in the mainstem (neutral pH, oxidizing
conditions) the major metal removal process would be sorption to iron arid aluminum oxides
orgamcs, and clay minerals. Chemical precipitation, while a dominant process for major ion '
removal (Fe, Al), is not likely operative for the metals, zinc, cadmium, Copper, and lead aL>t
for concentrated, relatively neutral mine discharges. Chemical transformation in concenSd
pomt sources may include a combination of these processes for major ions and zinc
SORPTION
^^=1^=^^^=
Fe-Ox ides
Al-Oxides
Organios
Clays
-10
Boundaries are illustrative
Figure 6. Mechanisms for the partitioning of dissolved metal to particulate
metal in surface water environments. Under typical ambient
conditions of pH and Eh in riverine environments, toxic metal
removal from the dissolved phase is dominated by sorption processes.
Other conditions found during treatment of wastewaters may favor
. production of metal precipitates such as hydroxides, carbonates,
or sulfides.
Finally, in the Pueblo Reservoir, the principal mechanism for maintaining good water
quality is the continual removal of the particulate metals to the bed region. Dissolved metal
concentrations are low so that water quality problems are not significant. Water quality data
and biological conditions indicate that the water quality appears protective for multiple uses
239
-------
It is apparent that the sediments of the reservoir are the sink for the metal loadings in the
upper basin and other tributaries of the river. Pre and post-impoundment sediment concentra-
tions are very similar. Based on the long-term loading pattern this would be expected as the
majority of metals entering the lower reaches of the Arkansas River and the reservoir are in the
particulate phase. The continuing metal attenuation by the sediments should be evaluated with
respect to changing chemical characteristics in the reservoir to determine if conditions could
exist which may release previously-bound metals. The role of sediments in removing metals and
the changes in the reservoir chemistry toward a more productive system as restoration activities
are implemented may result in alteration of the sediment-water column distribution of metals
and should be evaluated as part of the monitoring program.
Overall, the conceptual model will support data needs for subsequent phases of the
project, such as numerical modeling of individual tributaries (California Gulch, Chalk Creek,
Tennessee Creek, Fourmile Creek) for remedial technology evaluation (Steps 6-7) or for basin-
wide management modeling (Steps 6-8). For example, based on this detailed conceptual
understanding of the Arkansas River, the modeling objectives for the long-term numeric
modeling effort can be determined. The detailed review will evaluate the effects of various
management options on environmental processes, both from a direct observation of effects as
well as from potential or possible effects based on the relationships developed in Step Two.
Tentative selection of the modeling framework and the specific temporal and spatial scales to be
described by the numerical modeling will be made following this determination of appropriate
transformation processes for critical pollutants, transport mechanisms, physical processes and
exposure assessment needs to serve as direction to the monitoring program development.
Following this initial review of the data and the development of the conceptual model, the
project goals are reviewed and rephrased, as necessary, to account for unrealistic goals or
difficult restoration scenarios.
STEP 3: MONITORING PROGRAM DEVELOPMENT
Subsequent to Step Two, Step Three involves the development or modification of the
monitoring program to identify chemical, physical, and biological factors affecting the attainment
of use and/or protection of use including delineation of water quality, contaminant sources,
physical characteristics, hydrologic aspects of contaminant transport and fate, and, the biological
conditions and response. The monitoring program development also includes significant
attention to issues such as quality assurance/quality control, health/safety, data quality objectives,
sampling methodologies, biological methods, toxicity studies, and analytical methods. The data
analysis methods that will be used in subsequent steps in the methodology will be specified as
well: statistical, mass transport analysis, contaminant correlations, and data management and
validation methods. A monitoring program has been implemented for the California Gulch area
in response to the Superfund process. Further integration of this program with the overall
monitoring program for the basin will be completed in the near future.
240
-------
SUBSEQUENT STEPS IN WQAM PROCESS
-As the additional steps in the WQAM process have not been initiated, they will not be
described in detail; rather, they will be described qualitatively to present the continuing rationale
for a focused resource restoration program. Following the development and implementation of
the monitoring program, subsequent steps in the process are directed toward a detailed
understanding of the physical-chemical-biological phenomena which interact to determine the
overall resource quality.
STEP 4: DESCRIBE ENVIRONMENTAL QUALITY
While the conceptual model has resulted in an initial analysis of the extent and quality of
available data, it is these subsequent steps that provide the detailed analysis of chemical impacts
upon aquatic resources, contaminant source areas, the nature and impact of diffuse contaminant
loadings, the chemical transformations which regulate contaminant transport, and the overall
influence of the hydrologic system.
STEPS: ASSESS POTENTIALLY ATTAINABLE CONDITIONS
Continued steps in the process evaluate attainable conditions under reasonable contami-
nant control strategies and provide the link between resource use and necessary environmental
quality. As part of this step, it is the development of segment-specific quality goals for critical
contaminants that forms the basis for future resource restoration activity. These criteria are
developed using a combination of information from research, in-stream biomonitoring and
toxicity testing (source areas and stream profiles) to best protect or establish, acceptable use.
STEP 6: LINK CONTAMINANT DYNAMICS TO RECEPTOR EXPOSURE
The last steps of the process (Figure 2) address the development of the approach for
selecting and evaluating appropriate control strategies for restoring and maintaining the
established resource quality goal. The major aspect of Step 6 is the linkage of contaminant
dynamics to receptor exposure and resource use constraints. It is critical to establish relations of
source, transformations, and transport to resulting environmental quality which may support load
reductions or restoration decisions by management. It is emphasized that these relations may be
empirical or mechanistic depending upon the necessary complexity, and may employ numeric
modeling to support activities in certain parts of the basin. This step will permit the establish-
ment of treatment strategies or other loading controls based on total maximum daily loads
(TMDLs) to eliminate acute and chronic toxicity to the desired aquatic populations.
STEP 7: RESOURCE RESTORATION
The resource restoration phase, or the implementation of control strategies, is the actual
application of remedial technologies to control source loadings within the basin or other
activities to enhance resource use, such as habitat restoration, water quantity management (basin
diversions, reservoir releases, etc.), fish stocking programs, etc. Direct correlation of remedia-
tion ettorts with observed changes in water quality and restoration of resource use must be
made to permit the assessment of restoration effectiveness. For activity in the Arkansas River
basin, the assessment and control process should apply phased, iterative, and low-maintenance
technologies to the maximum degree possible. Source control in the basin for major metal
sources is likely to be required indefinitely.
241
-------
As part of this process, provisions should be developed for the reduction of the water
quality and biological monitoring to a maintenance-level plan upon meeting or approaching the
goals. This maintenance-level plan should include select chemical or biological variables
reflecting resource quality.
For example, sediments and macroinvertebrate metal levels tend to reflect long-term
water quality while individual sampling events for water quality provides chemistry data only for
the time of sampling. In addition, biological population data (periphyton, macroinvertebrates,
fish) in addition to sediment and select water quality data (target metals, general water quality)
should provide an overall indication of the health of the aquatic ecosystem. During Steps 4
and 5, these important variables should be selected for the Arkansas River and used in the long-
term monitoring program.
STEPS: GOALS ATTAINED, MAINTENANCE MONITORING
Finally, upon implementation of the selected control strategies and necessary maintenance
programs, the response of the environment to those controls is documented with respect to goals
attainment. Assuming successful restoration, a maintenance monitoring program is developed
which includes select chemical or biological variables that best reflect the resource quality.
These variables are monitored to ensure continual achievement of resource quality and will
trigger detailed monitoring should any degradation be observed which may be indicative of
control process failure.
SUMMARY
Water quality problems in the Arkansas River basin have prompted the initiation of a
wide variety of research, monitoring, and interpretative activities\in the last 5 years. To
coordinate the various programs, a comprehensive water quality assessment methodology has
been initiated in the Arkansas River basin under the Arkansas River Initiative. Based on the
current understanding of water quality problems, the WQAM provides a scientifically-based
support to the development of a long-term monitoring program and establishing the sequencing
and effectiveness of necessary remedial actions to restore aquatic resource uses. While elements
of the process have been and continue to be applied in other systems (Animas River, Chalk
Creek, and Eagle River), this is the first direct application of the methodology.
The major goal of the methodology, as described in this paper, is to describe, in quantita-
tive terms and with acceptable uncertainty, the results of manipulation of the interacting
physical-chemical-biological processes on the overall quality of the resource. Additionally, a
secondary goal is to ensure that these manipulations of the system, whether physical, chemical or
biological, restore and preserve the designated uses. The process also provides the "blueprint"
for the technical tasks to ensure communication and cooperation among the various govern-
mental and private/public parties. The detailed assessment of the system on a frequent basis
provides the needed scientific and economic justification for funding studies in the basin or
expending federal, state or private funds for remedial action. Continued application of the
WQAM to the Arkansas River basin will provide guidance for the range of technical activities
required for successful restoration of the basin and minimize duplication or a fragmented
approach to individual water quality or quantity problems.
242
-------
PARTICIPANTS
F.L. Back
Alliance for a Living Ocean
44 Sunrise Drive
Montvale, New Jersey 07645, USA
William Benson
Department of Pharmacology
School of Pharmacy
University of Mississippi
University, Mississippi 38677, USA
Harold L. Bergman
Fish Physiology & Toxicology Lab
Department of Zoology & Physiology
University of Wyoming
Laramie, Wyoming, 82070 USA
Linfield C. Brown
Department of Civil Engineering
Tufts University
Medford, Massachusetts 02155, USA
Dan Castleberry
USFWS, NFCRC Dixon Field Station
6924 Tremont Road
Dixon, California 95620, USA
Joseph J. Cech, Jr.
Department of Animal Biology
Division of Wildlife & Fisheries Biology
University of California, Davis
Davis, California 95616, USA
Chai, Minjuan
Oceanography Department
Xiamen University
Xiamen, Fujian Province, PRC
Monica Choi
Department of Wildlife & Fisheries
Biology
University of California, Davis
Davis. California 95616. USA
David Cohen
Division of Water Quality
S.W.R.C.B.
901 P Street
Sacramento, California 95801, USA
Victor De Vlaming
Division of Water Quality
Water Sources Control Board
901 P Street
Sacramento, California 95801, USA
Diego Fridmann
Department of Clinical Pathology
University of California, Davis
Davis, California 95616, USA
Richard Hansen
2005 Nimbus Road
Rancho Cordova, California 95670, USA
Alan Heath
Department of Biology
Virginia Polytechnic Institute • "
Blacksburg, Virginia 24061, USA
Laura Inouye
Department of Environmental
Toxicology
University of California, Davis
Davis, California 95616, USA
L.A Kepmer
Sierra Foothill Laboratory
820 South Highway 49
Jackson, California 95685, USA
Laura J.M. Lazzelle
142 Marilyn Ave.
Stockton, California 95207, USA
Lin, Hao-Ren
Department of Biology
Zhongshan University
Guangzhou, Guangdong Province, PRC
243
-------
Hong Dn
Department of Zoology
University of British Columbia
Vancouver, B.C. V6T 2A9
Canada
Lin, Yuehuan
Research Center for Eco-Environmental
Sciences
Academia Sinica
P.O. Box 934
Beijing, PRC
Jennie Jingyi Liu
Research Center for Eco-Environmental
Science
Academia Sinica
P.O. Box 934
Beijing, PRC
Nicholas T. Loux
Environmental Protection Agency
Environmental Research Laboratory
960 Collegfc Station Road
Athens, Georgia 30605-2720, USA
Maryann McEnroe
Division of Natural Science
State University of New York
Purchase, New York 10572, USA
Allen Medine
Walsh & Associates, Inc.
4888 Pearl East Circle - Suite 108
Boulder, Colorado 80301, USA
Scott Ogle
Department LAWR
University of California, Davis
Davis, California 95616, USA
Jose Pereira
National Marine Fisheries Service
National Oceanographic and Atmospheric
Administration
212 Rogers Avenue
Milford, Connecticut 06460, USA
Jimmie Pigg
Oklahoma Department of Health
P.O. Box 53991
Oklahoma City, Oklahoma 73152, USA
Bob Pine
USDA Botany Department
University of California, Davis
Davis, California 95616, USA
David J. Randall
Department of Zoology
University of British Columbia
Vancouver, B.C. V6T 2A9
Canada
Del Rasmussen
California State Water Resources
Control Board
P.O. Box 100
Sacramento, California 95814, USA
David A. Richard
405 Leslie Rd.
Stockton, California 95207, USA
Rosemarie C. Russo
Environmental Protection Agency
Environmental Research Laboratory
960 College Station Road
Athens, Georgia 30605-2720, USA
MikeSaiki
U.S. Fish & Wildlife Service, National
Fisheries Contaminant Research Center
6924 Tremont Road
Dixon, California 95620, USA
Gwen Starrett
California State Water Resources
Control Board
P.O. Box 944213
Sacramento, California 94244, USA
Deborah JL Swackhamer
University of Minnesota
Environmental and Occupational Health
School of Public Health
Minneapolis, Minnesota 55455, USA
244
-------
Nelson A. Thomas
U.S. Environmental Protection Agency
Environmental Research Laboratory
6201 Congdon Boulevard
Duluth, Minnesota 55804, USA
Robert V. Thurston
Montana State University
Fisheries Bioassay Laboratory
Bozeman, Montana 59717, USA
Winona Victory
U.S. Environmental Protection Agency
Region IX - Office of Regional Administration
235 Mission Street
San Francisco, California 94103, USA
Thomas Willingham
Region VHI - Water Quality Standards Office
U.S. Environmental Protection Agency
One Denver Place - 999 18th Street
Denver, Colorado 80202, USA
Lee Ann Woodward
Division of Environmental Studies
University of California, Davis
Davis, California 95616, USA
Chieh Wu
U.S. Environmental Protection Agency
OEPER (RD-682)
401 M Street, SW
Washington, D.C 20460, USA
Xu, Li-Hong
Institute of Hydrobiology
Academia Sinica
Wuhan, Hubei Province, PRC 430072
Yuan, Chuan-ruh
Department of Zoology
Nanjing University
Nanjing, Jiangsu, PRC 210008
Zhang, Guo-an
Xinjiang Institute of Environmental
Protection
Urumchi, PRC
Zhang, Yong-Yuan
Institute of Hydrobiology
Academia Sinica
Wuhan, Hubei Province, PRC 430072
Zhu, Dongwei
Department of Environmental
Engineering
Georgia Institute of Technology
Atlanta, Georgia 30332, USA
Zhu, Xing-Xiang
National Environmental Protection
Agency
Water Environment Management Division
No. 115, Xizhimennei Naoxiapjie
Beijing, PRC
JoeZinkl
Department of Clinical Pathology
University of California, Davis
Davis, California 95616, USA
245
U.S. GOVERNMENT PRINTING OFFICE: 1993 - 750-002/80261 '
-------
------- |