RELATIONSHIPS
AMONG FISH POPULATIONS,
METALS CONCENTRATIONS, AND STREAM DISCHARGE
IN THE UPPER CLARK FORK RIVER
Glenn
R. Phillips
Department of Fish, Wildlife and Parks
Capitol Station
Helena, MT 59601
Abstract--Concentrations
of total recoverable copper, iron, and zinc in water were measured weekly
between early April and mid-July 1984 in various segments of the Clark Fork
River drainage upstream of Milltown Dam; stream discharge measurements were
also recorded. Fourteen locations were sampled including eight in the mainstem
and six in tributaries. All three metals were sometimes present in the Clark
Fork River at concentrations that exceeded criteria for protection of aquatic
life. Exceedances occurred at all mainstem stations, although conditions
appeared to be least favorable for aquatic life between Deer Lodge and the
confluence with Rock Creek. Of the metals measured, copper was present at the
highest concentrations relative to its toxicity and is probably the most
limiting. Copper concentrations in gills of brown trout collected after a fish
kill indicated lethal exposure to copper, confirming the biological
significance of copper to the river. Alternatively, cadmium concentrations
found in gills were well below lethal thresholds, suggesting that cadmium is
relatively less significant than copper in the Clark Fork.
Water entering the Clark Fork from the Little Blackfoot
River and Rock Creek is low in metals. Consequently, metals concentrations in
the Clark Fork are measurably lower downstream of these tributaries.
Fish population data for various segments of the river
(although limited) correlate well with water quality; i.e., lower fish numbers
correspond to more severe metals conditions. Direct flow of untreated Silver
Bow Creek water into the Clark Fork River resulted in an extremely high peak
metals concentration that was observable at all of the mainstem stations
sampled. During the bypass. the highest metals concentrations occurred at Warm
Springs. A more prolonged period of elevated metals concentrations occurred in
the reach of river between Deer Lodge and Rock Creek and is apparently owing
to erosion into tailings deposited in the flood plain. This prolonged exposure
appears to damage fish populations.
Limited measurement of pH in various portions of the
drainage indicate that pH is higher in the Warm Springs vicinity than in
downstream reaches. If this observation is characteristic of the rest of the
year, the implication is that metals are less soluble and probably less toxic
near Warm Springs than downstream.
INTRODUCTION
Beginning in the
mid-1800's and lasting for nearly a century, Silver Bo~ Creek and the upper
Clark Fork River received large quantities of metals- contaminated wastes
originating from the Butte-Anaconda complex of mining, ore processing, and
smelting industries. Prior to 1975, pollution was so severe that most of the
upper drainage was incapable of supporting fish life. Even today, excessive
metals deposits prevent establishment of a viable fishery in Silver Bow Creek.
Moreover, metals in the Clark Fork River continue to sup- press fish
production in much of that stream.
First attempts to treat the problem occurred in 1954 with
the construction of settling ponds at Warm Springs and initiation of a liming
operation upstream of the ponds. Metals complexed by hydroxides and carbonates
precipitated in the ponds, thereby lowering metals concentrations downstream.
Initial attempts at treatment were not entirely successful because the
quantity of metals entering the ponds was greater than could be effectively
removed. Significant improvements in the Clark Fork River fishery were not
realized until after installation of updated waste treatment systems at
Anaconda's Butte facility lowered the metals load in Silver Bow Creek
(mid-1970's). Unfortunately, tailings that were historically dumped in the
river are present in the flood plain downstream of Warm Springs. Consequently,
large quantities of metals continue to enter portions of the river,
particularly during periods of high river discharge.
The objectives of this study were to: (I) quantify copper
loading from various segments of the drainage, (2) determine the effects of
tributary dilution on copper concentrations in the mainstem, (3) evaluate the
relative severity and significance of the copper exposure experienced by
organisms living in different reaches of the river, and (4) observe
relationships between metals concentrations and river discharge. We also made
inferences concerning the biological significance of copper and cadmium in the
Clark Fork drainage using information on metals in gills of fish that were
victims of a fish kill.
MATERIALS
AND METHODS
During
spring 1984, weekly water samples and discharge measurements were taken from
14 locations in the Clark Fork River drainage upstream of Milltown Dam.
Sampling sites included eight mainstem stations and six in tributaries (fig.1). Discharge was either measured in the field using a Price current meter or
was obtained from U.S. Geological Survey (USGS) or National Weather Service (NWS)
gauging records. The USGS maintains continual discharge recording stations in
Warm Springs Creek, Rock Creek. Flint Creek, the Little Blackfoot River, and
the Clark Fork River at Deer Lodge, Gold Creek, and Clinton. The NWS collects
daily discharge measurements at Drummond.
Grab surface water samples were collected in polyethylene
bottles and acidified in the field (to pH 0.2) with concentrated, distilled,
reagent grade nitric acid. Samples were refrigerated at the time of collection
and finally analyzed for total recoverable copper, iron, and zinc by the
Chemistry Laboratory Bureau of the Montana Department of Health and
Environmental Sciences. Flameless atomic absorption spectrophotometry was
employed.
Quality control measures included duplicate analyses and
spikes on at least 10% of the samples and periodic measurement of EPA
certified standard solutions. Analytical results were not accepted unless
percent recovery of standard solutions and spikes were within the following
ranges: iron, 76-125%; copper, 88-121%; and zinc, 89-118%.
RESULTS
AND DISCUSSION
Metals
Concentrations Relative to Criteria
Highest concentrations of copper, iron, and zinc were measured on May 15,
1984, at the onset of runoff. During this event all three metals exceeded
criteria for protection of aquatic life; of these, copper occurred at the
highest concentrations relative to its toxicity {table 1) and was sometimes
present in excess of 50 times the criterion. By comparison, the highest
measured concentration of zinc and iron was twofold and sevenfold the
criteria, respectively.
In
addition to the metals that we measured, cadmium has been identified at
relatively high concentrations in tailings present in the Clark Fork drainage.
Unfortunately, cadmium is more difficult to evaluate than copper, iron, or
zinc because the criterion to protect aquatic life from cadmium is below the
detection limit attainable by most laboratories. For example, the cadmium
detection limit attainable by the Department of Health and Environmental
Sciences Laboratory is over 150 times higher (table
1) than the chronic criterion
for cadmium (l). Hence, cadmium may be present at concentrations that
have biological significance even when results of water analyses routinely
show that cadmium is present below the detection limit.
Because of the analytical difficulties of measuring cadmium
in water, we attempted to evaluate the importance of cadmium in the Clark Fork
River by measuring cadmium in gill tissue of brown trout; copper was similarly
evaluated. The fish kill that occurred in the Warm Springs vicinity on August
1, 1984, presumably provided a "worst case situation" for such an
evaluation. Immediately following that event cadmium concentrations in gill
tissue (dry tissue basis) of dead or dying brown trout ranged from 0.4 to 1.5
ug Cd/g (table 2). Comparatively, Mount and Stephan (5) found that bluegill (Lepomi8
macrochiru8) killed by acute exposure to cadmium always contained greater
than 150 ug Cd/g in gill, or nearly two orders of magnitude greater than we
measured in brown trout from the Clark Fork. In contrast, copper
concentrations in gill tissue from brown trout killed in the Clark Fork River
ranged from 14 to 40 ug Cu/g--indicative of lethal exposure to copper (1).
These results underscore the biological significance of copper in the Clark
Fork River system, but imply that cadmium may be relatively unimportant.
Eaton
(2) and Muska and Weber (6) have shown synergistic interactions between
some-metals, particularly copper and zinc. It is likely that some metal
interactions occur in the Clark Fork River. Such interaction, however, may not
be important from a practical standpoint because patterns of exposure are
similar for all metals. Indeed, all of the metals in the Clark Fork drainage
originate from virtually the same source (erosion of tailings deposits into
the river). Consequently, reclamation efforts to eliminate any one metal will
effectively control all of them.
State agencies have speculated for some time as to why
brown trout are able to exist in the upper Clark Fork River while rainbow
trout (Salmo gairdneri) are not. This is particularly perplexing since
rainbow trout are common in some tributaries; e.g., Rock Creek. The absence of
rainbow trout from the majority of the upper river may be due to their greater
sensitivity to copper than brown trout. McKim and others (4) found that 37 ug Cu/L
resulted in complete mortality of rainbow trout embryos while 555 ug Cu/L
was required to produce the same effect on brown trout. Copper
concentration exceeding 37 ug/L commonly occurs in the Clark Fork River;
however, we have not measured copper concentrations above 555 ug Cu/L during
any of our sampling.
Metals Concentrations and Loading vs. River Discharge
All three metals that were measured followed similar concentration patterns in
the river; hence, trends for anyone metal are representative of the others.
Since copper appears to be of the greatest biological significance,
concentration trends for the various segments of the drainage will be
discussed only for copper. Concentration and loading trends for zinc and iron
are included in appendices A and B.
HEADWATERS
Water quality in the Clark Fork River near Warm Springs is dependent on the
quality of water originating from the three headwater tributaries that combine
to form the source (fig. 1). These include Warm Springs Creek, the discharge
from Anaconda Minerals Company treatment pond number two (Silver Bow Creek
water that has been limed and diverted through artificial settling ponds), and
the Mill-Willow bypass (a channel that routes the combined flows of Mill and
Willow Creeks around the Warm Springs Ponds). During periods of extremely high
stream discharge, unlimed Silver Bow Creek water may spill into the
Mill-Willow bypass channel and subsequently into the Clark Fork River.
Spring runoff commenced approximately May 15. 1984, and
ended near the middle of July (fig. 2). Prior to runoff, the Pond-2 discharge
accounted for approximately 50% of total discharge to the upper river (fig.
2). All three of the headwaters contributed nearly equal discharges during the
majority of runoff; however. Warm Springs Creek predominated near the tail end
of runoff contributing about half of the total discharge of the Clark Fork
River at Warm Springs during the last 2 weeks in June (fig.
2).
Copper concentrations in Warm Springs Creek and in the
Mill-Willow bypass were highest during runoff (fig. 2). Copper in Warm Springs
Creek remained elevated throughout runoff but was highest at the beginning,
suggesting that copper salts may accumulate on the surface of the flood plain
during periods of reduced flow and then enter the stream as runoff commences.
An extremely high copper peak occurred in the Mill-Willow bypass at the onset
of runoff (0.7 ug Cu/L was measured). This peak was the result of
untreated Silver Bow Creek water spilling into the Mill-Willow channel.
Comparatively, highest copper concentrations in water leaving Pond-2 occurred
before runoff, when discharge from Pond-2 was highest relative to Warm Springs
Creek and the Mill- Willow bypass. Loading calculations indicate that before
runoff, Pond-2 is a major source of copper to the upper Clark Fork River (fig.
3).
Reasons for the higher amounts of copper leaving the ponds during the inter
have not been validated. Possible explanations include: (1) ineffective liming
resulting from ice cover and subsequent channeling of Silver Bow creek in the
vicinity of the liming station, (2) inefficient settling of metal precipitates
in the ponds because of reduced settling capacity caused by ice over and
channeling of water through the ponds during winter, (3) development of
reducing conditions near the bottom of the ponds in winter resulting in
mobilization of metals from sediments, (4) sloughing of metals-rich algal
cells from the ponds during winter. Information being collected as part of the
remedial investigation for the Silver Bow Creek Superfund Site should help
explain this phenomenon.
MAINSTEM
Discharge
patterns in the mainstem Clark Fork River change significantly at the
confluence with the Little Blackfoot River and again at the confluence with
Rock Creek (figs. 4, 5). Both of these major tributaries result in an
approximate doubling of flow in the Clark Fork River.
Peak
copper concentration at all eight mainstem stations occurred concurrently with
the spilling of Silver Bow Creek into the Mill-Willow bypass channel on May
15, 1984, and with the onset of runoff (figs.
4, 5). Loading calculations for
copper during the bypass event indicate that about one-third of the copper
present in the river upstream of Rock Creek originated in Silver Bow Creek (fig.
6). Erosion of tailings into the mainstem of the river or possibly
ground water inputs must have accounted for the remainder.
Copper loading, for the most part, showed a progressive
increase in the downstream direction (fig.
6). As flow increased, copper
originating from erosion into the riverbanks accounted for an increasing
proportion of the total copper load. Several of the downstream stations,
primarily those between Deer Lodge and Rock Creek, experienced a number of
copper peaks that did not occur at Warm Springs (figs.
4. 5) resulting in a
more prolonged period of elevated copper concentrations.
Copper
concentrations were lower downstream of the Little Blackfoot River and Rock
Creek than above (figs. 7, 8), demonstrating the beneficial influence of
tributary dilution. The benefits of dilution were, however, less pronounced
downstream of the Little Blackfoot River than below Rock Creek, apparently
because the riverbanks downstream of the Little Blackfoot are contaminated
with tailings.
Metals
and the Fishery
The
fishery record for the upper Clark Fork River is not complete; however, it is
instructive to review the fishery data relative to what is known about water
quality in the various segments of the river. Fish population estimates taken
between 1979 and 1983 for eight distinct segments of river (fig.
9) were
summarized by Knudson (3). Greatest fish numbers occur in the reach near Warm
Springs where numbers of catchable brown trout (> 6 inches) consistently
exceed 1,000 fish per mile. Numbers decline rapidly-downstream, to less than
400 fish/mi in the Deer Lodge vicinity, and to only 50 fish/mi near Bearmouth.
Downstream from the confluence of the Clark Fork with Rock Creek, fish numbers
begin to increase--approaching nearly 500 fish/mi in the reach below Milltown
Dam where the Big Blackfoot River joins the Clark Fork.|
The
fact that the highest fish numbers are present nearest the source of
contamination is somewhat perplexing; however, there are at least two possible
explanations. First, the close proximity of this stream reach to the liming
station may offer an advantage. Measurements of pH taken at several locations
in the upper Clark Fork on August 8, 1984, showed that the pH at Warm Springs
was greater than that at several other stations lower in the drainage (Table
3). Metals solubility tends to decrease with increasing pH. Consequently, fish
in the upper reach may be exposed to lower concentrations of dissolved metal
forms. Although this represents only one measurement, the trend is worth
exploring further.
Perhaps
more importantly, the duration of exposure to elevated metals concentrations
is noticeably shorter near Warm Springs than in reaches further downstream (figs.
4, 5). Possibly, the riverbanks 1n the Warm Springs vicinity are
relatively stable and erosion of bank tailings is not a serious problem there.
The most prominent event in the Warm Springs vicinity is the bypassing of
Silver Bow Creek into the Mill-Willow channel. During 1984 this bypass
resulted in unusually high metal concentrations, but this occurred for less
than a week's time. The fishery data suggest that the more prolonged elevation
of metal concentrations that occurs further downstream is of much greater
biological significance to the river than the short period of high metal
concentrations that occurred during the bypass.
It is noteworthy that a measurable improvement in the
fishery occurs downstream of both major tributaries. This is particularly
evident at the confluence of Rock Creek where fish numbers rebound from less
than 50 fish/mi to nearly 400 fish/mi; reduced concentrations of metals
downstream of Rock Creek appear to account for this improvement.
Existing data are insufficient to evaluate whether the
small increase in fish numbers that occurs downstream of the Little Blackfoot
River is biologically significant. A gradual increase in metal concentrations
between Garrison and Drummond suggests that there are sources of metals in
that river reach. Low fish numbers and high metal concentrations downstream of
Drummond
Clearly,
our understanding of the factors controlling fish populations and other
biological parameters in upper Clark Fork River are limited by the complexity
of the problems that exist and by a limited data base. However, information
such as presented at this symposium continues to bring us nearer to the level
of understanding that will be required to make responsible reclamation and
management decisions.
ACKNOWLEDGMENTS
Special
recognition is extended to Kurt Hill who performed the bulk of the fieldwork
during the study. Kurt Hill, Brian Albro, and Sandra Dunkel all contributed to
completion of illustrations. A number of biologists from the Department of
Fish, Wildlife and Parks, including Don Peters, Jim Vashro, Dennis Workman,
and Wayne Hadley, collected fish population data. Finally, Dr. Loren Bahls of
the Water Quality Bureau provided valuable editorial suggestions.
LITERATURE
CITED
1.
Brungs, W.E.; Leonard, E.M.; and McKim, J.M. 1973. Acute and long-term
accumulation of copper by the brown bullhead, Iatalurus nebulosus. J.
Fish. Res. Board. Can. 30: 583-586.
2.
Eaton, J.G. 1973. Chronic toxicity of a copper, cadmium, and zinc mixture to
the fathead minnow (Pimephales promelas Rafinesque). Water Res. 7:
1723-1736.
3.
Knudson, K. 1984. A preliminary assessment of impacts to the trout fishery of
the upper Clark Fork River, Montana. Report to Montana Dept. Fish, Wildl. and
Parks. 30 p.
4.
McKim, J.M.; Eaton, J.G.; and Holcombe, G.W. 1978. Metal toxicity to embryos
and larvae of eight species of freshwater fish II: Copper. Bull. Environ.
Contam. Toxicol. 19: 603-616.
5.
Mount, D.I.; and Stephan, C.E. 1967. A method for detecting cadmium poisoning
in fish. J. Wildl. Manage. 31: 160-172.
6.
Muska, C.F.; and Weber, L.J. 1977. An approach for studying the effects of
mixtures of environmental toxicants on whole organisms performances. In: Tubb,
R. A. (editor). Recent advances in fish toxicology: a symposium.
EPA-6OO/3-77-O85. Office of Research and Development, Corvallis, OR. 71-87.
7.
U.S. Environmental Protection Agency. 1980. Ambient water quality Criteria and
Standards criteria for cadmium. EPA 440/5-80-025. Division, Washington, DC.
110 p.
8.
U.S. Environmental Protection Agency. 1980. Ambient water quality Criteria and
Standards criteria for copper. EPA 440/5-80-036. Division, Washington, DC. 74
p.
9.
U.S. Environmental Protection
Agency. 1980. Ambient water quality Criteria and Standards criteria for zinc.
EPA 440/5-80-079. Division, Washington, DC. 88 p.
10.
U.S. Environmental Protection Agency. 1976. Quality criteria for water.
Washington, DC. 256 p.
11. Van Meter. W. P. 1974. Heavy metal concentrations in fish tissue of the upper Clark Fork River. Report of Montana University Joint Water Resources Research Center, Bozeman, MT. 37 p.