Distribution
of Trace Metals in Fine-grained Bed Sediments and Benthic Insects
in the Clark Fork River, Montana
Ellen V. Axtmann *, Daniel J. Cain and Samuel N. Luoma
U. S. Geological Survey
345 Middlefield Road, M.S. 465
Menlo Park, CA 94025
Abstract: The downstream distribution of Cu, Cd, and Pb in fine-grained sediments and benthic insect larvae of the Clark Fork River, Montana is characterized. This river has been heavily con laminated as a result of past mining and smelling operations near its headwaters. Concentrations of all metals in bed sediments displayed a simple exponential downstream decrease through the upper 181 km of the river. The trend suggested metal contamination originated from source(s) in the headwaters, with physical dilution occurring downstream. Additional data suggested floodplain sediments also were contaminated by the original source(s). Secondary inputs from cutbanks in the floodplains may have extended the downstream influence of the contamination. The exponential model predicted that sediment contamination should extend at least 550 km downstream, a result that was verified with data from a separate, independent study. Metal contamination, as observed in all taxa of insect larvae collected from the upper Clark Fork. Concentrations in the insect larvae were highest in the upper 100 km of the river, but downstream trends were more complex than those of the sediments. Some differences in trends occurred among taxa and metals. Areas in the river of enhanced or reduced metal contamination also were apparent. Metal contamination, however, was still evident at 381 km, the most downstream station sampled. Metal concentrations in sediments and insects decreased at the confluences of uncontaminated tributaries, but the influence of tributaries on metal contamination in the Clark Fork River was localized, extending for only 1-2 kin below the confluences.
INTRODUCTION
Copper mining and smelling operations near the headwaters of the Clark Fork
River, Montana have resulted in large-scale contamination of both bed and
floodplain sediments of the river (Andrews. 1987; Rice and Ray. 1985; Johns and
Moore. 1985; Luoma et a1., 1989). It is estimated that over 100 million metric
tonnes of mine and smelter wastes were released via Silver Bow and Warm Springs
Creek to the Clark Fork River (Andrews, 1987) and that more than 2 million cubic
meters of contaminated sediment have been deposited in the floodplain of the
Clark Fork (Moore and Luoma, 1990). Tailings ponds constructed in the headwaters
in the mid-1950s greatly reduced the contaminant input to the river. Although
smelting operations ceased in the 1980s and mining activities are reduced,
contaminated sediments stored in the floodplain provide a continuing source of
metals to the river (Johns and Moore, 1985; Axtmann and Luoma, 1990; Moore and
Luoma, 1990a)
Reduced trout populations in the Clark Fork as well as
occasional fish kills during high-flow events suggest that metals in the Clark
Fork may have toxic effects (Phillips. 1987; Moore and Luoma. 1990a). The
purpose of this paper is to compare the downstream distribution of Cu, Cd, and
Pb in fine-grained bed sediments of the Clark Fork with metal concentrations in
benthic insects (which have direct contact with the sediments), and to discuss
metal availability to the benthic food web in light of these distributions.
Sediment
In order to characterize the large-scale downstream distribution of metals in
the sediments of Clark Fork River, fine-grained bed and floodplain sediments
were collected during low flow in August 1986 and 1987 along the upper 381 km of
river (all distances measured as kilometers downstream from the confluence of
Warm Springs Creek) (Fig.1). Fine-grained bed sediments were collected in 1986
and 1987 from 17 sites at an average interval of 11 km throughout the upper 181
km of the river, and in 1986 at three sites in the lower 200 km of the river
(Fig.1). Bed sediments were also collected from five major tributaries within
25 km of their confluence with the Clark Fork River (Fig.1).
In 1989, intensive sampling of bed
sediment around the mouths of two tributaries, Rock Creek and Flint Creek, was
conducted to investigate small-scale effects of tributaries on the distribution
of metals in the Clark Fork. Rock Creek was chosen because it had extremely low
sediment metal concentrations and might have a diluting effect on metal
concentrations in the Clark Fork. Flint Creek sediments were enriched in Pb
relative to Clark Fork sediments at the confluence of the rivers, and data from
l986-1987 suggested there might be a Pb signal from Flint Creek in the Clark
Fork sediments (Axtmann and Luoma, 1990). Triplicate samples were collected in
the Clark Fork at each of three sites within approximately 1.5 km both above and
below the confluences of the tributaries. Triplicate samples were also collected
in the mouths of both tributaries and at sites on the Clark Fork up to 11 km
above and below the confluences.
Benthic
Insects
The larvae of six taxa of benthic insects from two orders (Trichoptera and
Plecoptera) were collected from the Clark Fork River and five of its
major tributaries. Some taxa were rare or absent at some stations, and some taxa
were preferentially collected for specific studies; thus not all taxa were
represented in all samples. Data from two taxa, the filter-feeding Hydropsyche
spp. (order Trichoptera) and the predaceous stonefly Claassenia sabulosa (order
Plecoptera) are emphasized in this paper because they were present at most
stations and represented the variety of bioaccumulation patterns observed.
Insect larvae were collected with kick nets from riffle areas
at the same stations that sediments were collected in August 1986 and 1989 (Fig.1). Insects were sorted on site by taxon, then held in plastic bags filled with
ambient river water in an ice cooler for 4-6 hours to allow the insects time to
clear their digestive tracts. The water in the bags was then drained and the
insects were frozen. Samples were thawed in the laboratory and thoroughly rinsed
with distilled water to remove particulates from exterior surfaces.
Identifications of specimens were verified and individuals from the same
taxonomic group were sorted by size to examine possible size-related differences
in metal concentrations. The body length of each Claassenia sabulosa was
measured from the head to the last abdominal segment. Different sizes of Hydropsyche
spp. were qualitatively separated by eye. Individuals of the same taxon and
of similar size were composited into samples to attain a minimum total dry
weight of 50 mg. Samples were dried at 800 C, weighed, and then digested by hot
16N HNO3 reflux. The acid was evaporated after the sample solution turned clear
then residue was reconstituted in 25% HCI. Metals were determined by AAS on
samples collected in 1986 and by ICAPES on samples collected in 1989.
Concentrations of Cu and Cd in biological standards (NBS standard reference
material 577a, bovine liver) prepared and analyzed by the same method were
within the range of certified values. The reliability of the method could not be
verified for Pb because concentrations in NBS reference material were below
analytical detection limits.
The large-scale distribution of metal contamination in
insects through the 381 km study reach was evaluated by aggregating data
collected in 1986 from upstream (0-60 km), mid-river (106 -164 km) and
downstream (191 -381 km) reaches. Effects of tributaries on small-scale
variation in insect metal concentrations in the Clark Fork were analyzed with
data collected from the same sites as sediments around the mouths of Hint Creek
and Rock Creek in 1989.
Data collected in August 1986 were analyzed for correlations
between insect and sediment metal concentrations. Data were log transformed for
the correlation analysis. Statistical significance of a correlation
(product-moment correlation coefficient, r, was set at p < 0.05).
Studies of the contribution of the
content of the gut to whole body metal burdens of insects in the Clark Fork also
were initiated in 1989. Pleronarcys californica, a detritus-feeding
stonefly, was employed in this study because its digestive tract could be easily
removed. Individuals of this species were collected from the Clark Fork at 168
km, from Flint Creek and from Rock Creek. Samples were collected and prepared as
described above. Specimens were secured on a paraffin surface the under a
microscope (magnification 8-40 x). A dorsal, longitudinal incision was made to
open the body cavity and expose the gut. The gut Wall was incised and the
contents were carefully removed. The gut content and the remains of the body
were analyzed separately.
Metal concentrations in tributary sediments generally reflected the history of
metal extraction in the watershed (Axtmann and Luoma,1990). Sediments collected
in Flint Creek, which has an extensive history of mining in its watershed, were
enriched in both Pb and Cd (although Cd concentrations in Flint Creek were not
as of enriched as in the Clark Fork). Less historic mining activity has occurred
in Rock Creek and the Bitterroot River basins than in other tributaries. These
systems had the lowest metal concentrations in sediments. Low concentrations
also were observed at the mouth of the Blackfoot River, more than 150 km
downstream from the inputs of small-scale mining activities (Moore et al.,
1990). The mean of the concentrations from the three latter systems was taken as
an operational reference concentration for each metal (indicative of pre-mining
metal concentrations) with which to evaluate metal enrichment in the Clark Fork
(Fig.2).
In contrast, analysis of closely spaced bed sediment
samples around the mouth of Flint Creek and Rock Creek in 1989 indicate that
tributary input has only a minor and localized effect on Clark Fork sediment
metal concentrations. The 1989 study showed that Cu concentrations were low in
the sediments of both Rock Creek a Flint Creek, relative to Clark Fork sediment
metal concentrations. When Cu concentrations from samples collected in 1989 are
plotted with the regression line and 95% confidence intervals for the combined
1986 and 1987 data set, it is clear that Cu concentrations are locally depressed
below the confluence of both Rock Creek (Fig.
3a) and Flint Creek, but that they
quickly return to their pre-tributary concentrations. Lead concentrations, on
the other hand, were higher in Flint Creek sediments than in the Clark Fork
sediments, but much lower in Rock Creek. Lead concentrations show a small, and
again localized, increase below Flint Creek (Fig.
3b), and a small, but local
decrease below Rock Creek. These results further suggest that erosion of
contaminated sediments from cutbanks below the confluence of tributaries
counteracts the diluting or enhancing effect of sediments from the tributaries.
Benthic Insects
Enriched metal concentrations were observed in all six taxa of benthic insects
collected from the upper Clark Fork River, compared to concentrations in animals
from tributaries. However, trends in metal contamination of benthic insects in
the river were complicated by differences among taxa and metals, and by the
scarcity of some taxa, especially in the most contaminated reaches of the river.
Hydropsychid caddisfly larvae are the most widely dispersed of the benthos
targeted for collection in this study. Metal contamination of caddis fly larvae
in the Clark Fork River was evident for 381 km downstream when concentrations in
specimens from the Clark Fork were compared with animals from reference
tributaries (Table 1). When stations were aggregated by river reach, metal
concentrations in the caddis fly, Hydropsyche spp., were highest in the
uppermost Clark Fork, and lowest downstream. The elevated concentrations of Cu,
Cd, and Pb in the reach between 191 and 381 km downstream of the Warm Springs
tailings ponds were significantly higher than in the reference tributaries (p
< 0.05; ANOVA, except where an individual value from a reference site was
compared to the aggregated numbers from the Clark Fork by t-test- Sokal and
Rolf, 1969, p. 168-169). Metal concentrations in Hydropsyche spp.
correlated positively with the exponential decrease observed in the metal
concentrations of fine bed sediments (coefficients of determination, r2 = 0.74,
0.150, 0.54 for Cu, Cd, and Pb, respectively; p < 0.05). One notable
difference in trends occurred for Cd between 20 and 100 km below the Warm
Springs Ponds where (concentrations in Hydropsyche spp. increased despite
significant decreases in sediment Cd concentrations (compare Fig. 2 and
Fig. 4).
Trends in metal concentrations of the predaceous stonefly, C. sabulosa, were
obscured partly because this species was either rare or absent at stations in
the upper 60 km, the most contaminated reach of the river. Below 60 km,
downstream trends in metal concentrations in this species were complex (Fig.
4),
although concentrations were higher than found in reference tributaries. Cadmium
in C. sabulosa correlated significantly with sediment Cd concentrations
(r2 = 0.48; p < 0.05), largely as a result of the difference between the
average Cd concentrations of insects collected above 180 km and below 200 km.
Correlations for Cu and Pb were insignificant Copper concentrations in C. sabulosa
in the Clark Fork River varied over a narrow range of 44 to 75 ug g-l, but
these concentrations were 1.4- 2 times greater than in specimens of this species
collected from tributaries. Lead concentrations, which were generally low in C. sabulosa,
decreased sharply below the most upriver station where the species was
found, however no consistent trend was evident further downstream.
Concentrations of some metals were lower at stations immediately below Rock
Creek (160 km; Cu and Pb) and the Blackfoot River (181 km; Cu and Cd) than at
stations located further downstream.
Although contamination in the undigested
content of the gut was important in Pteronarcys, the results also
indicated that the body (minus the gut) contained substantial quantities of
metals. Thus whole body metal concentrations do not simply reflect gut
contamination. More important, whole body metal concentrations and metal
concentrations of the body without the gut both generally reflected the relative
differences in metal contamination among the three sites studied, so negative or
positive influences of the gut content did not change interpretations of metal
distributions in the river. Gut contamination in other taxa in the Clark Fork
has not been determined, thus it is not known how typical the results for P.
californica are. Smock (1983b) analyzed the metal contents in the gut
material of 40 taxa divided into five different feeding categories. Although
detritivores were not singled out as a separate category, the results from his
study showed that the proportions of metals in the gut contents of insects that
ingest detritus and plant material were similar to filter feeders. Hydropsyche
spp. was included with several other taxa in this later group. In contrast,
predators had significantly lower proportions of metal in their gut content.
Extensive metal contamination resulting from historic discharges of mining and
smelting wastes was evident in both fine-grained bed sediments and in benthic
insects of the Clark Fork River. Elevated sediment metal concentrations extended
at least 550 km downstream from the original source of contamination in
headwaters, and contamination of biota extended at least 380 km downstream.
Data from both bed sediment and benthic insect larvae
indicate that metal contamination was greatest in the most upstream reaches of
the river and decreased downstream. Although bed sediment metal concentrations
follow a distinct exponential decline with distance away from the source, trends
in insect metal concentration were more complex. Areas of enhanced and reduced
metal concentrations occurred on spatial scales ranging from less than 5 km to
10's of km were evident.
The small body of available evidence suggests that the whole
body trace metal concentrations of aquatic insects reflect in at least a
relative sense levels of biologically available metals. The complex trends in
the Clark Fork River, suggest that the bioavailability of metals is spatially
heterogeneous and element-specific. Local conditions may create islands of
reduced or enhanced metal bioavailability, especially where there are sharp
physical and/or chemical boundaries such as at the confluences of tributaries.
In other cases the causes of the differences in bioavailability are less clear.
Cadmium, for example, appeared to exhibit greater bioavailability to caddisfly
larvae in the upper 100 km of the river than indicated by the sediments, in the
absence of any recognized chemical or physical change in the river.
Simple relations between environmental
contaminant concentrations and animal contaminant concentrations are rarely
observed in nature because of the complex interaction of geochemical and
biological factors (Luoma, 1989). Extractions of sediments with strong acids may
not be sensitive to geochemical conditions that affect metal bioavailability (Luoma
and Bryan, 1978; 1982; Langston, 1986). Differences in food selection may
contribute to differences in whole body metal concentrations among
species, with concentrations being higher in species that ingest sediment and
detritus than in predators (Smock, 1983b). The difference in Pb concentrations
between Hydropsyche spp., a filter feeder, and Claassenia sabulosa,
a predator, (Fig. 4) is consistent with Smock's (1983b) conclusions. The absence
of C. sabulosa from the most contaminated reach of the river may also
have weakened statistical
correlations between sediment and metal concentrations in this species.
The effect of tributary input on both
benthic insect larvae and sediment metal concentrations was localized,
suggesting tributaries are also a source of variability in trends. Inputs of
tributaries with relatively low sediment metal concentrations appeared to be
responsible for localized decreases in metal burdens in insect larvae,
presumably because exposure was reduced at the confluence of the tributaries
either by physical mixing (Clark Fork sediments and/or water were diluted with
relatively pristine material from the tributaries) or by a physicochemically
induced decrease in the biological availability of metals. Flint Creek had an
analogous, but opposite effect on Pb concentrations, by being an additional
source of Pb-enriched sediments to the Clark Fork. Drift of insects from the
tributaries into the main stem of the river could also have influenced metal
concentrations in insect samples collected below the confluences.
Concentrations of metals in both benthic insects and bed
sediments returned to their pre-tributary concentrations within 1 or 2 km below
both Rock Creek and Flint Creek. Two possible explanations for such a localized
effect are that tributary sediment loads are small compared to the Clark Fork
and/or input of contaminated material from cutbanks below the tributary
confluence quickly overshadow any diluting or enhancing effect of the
tributaries.
Phillips ( 1985) reported enhanced fish populations directly
below the confluence of Rock Creek. This study suggests that areas below the
mouths of uncontaminated tributaries may provide small, localized refuges from
severe metal contamination for biota of the Clark Fork River. How biological
communities respond to such patches of reduced metal exposure merits study.
Andrews,
E.D. 1987. Longitudinal dispersion of trace metals in the Clark Fork River
Montana.
In:Averett, R.C. and McKnight, D.M., eds., The chemical quality of water and the
hydrologic cycle. Lewis Publishers, Ann Arbor, MI. 1-13.
Axtmann,
E.V.; and Luoma, S.N. 1990. Large-sca1e distribution of metal contamination in
the fine-grained sediments of the Clark Fork River Montana. Applied Geochem. (in
press).
Bradley,
S.B.; and Lewin, J. 1982. Transport of heavy metals on suspended sediments under
high flow conditions in a mineralized region of Wales Environ. Pollut. 4:
257-267.
Brook. EJ.; and Moore. J.N. 1988. Particle-size and chemical control of As, Cd, Cu, Fe, Mn, Ni, Pb, and Zn in bed sediments from the Clark Fork River, Montana (USA). Sci. Total Environ. 76: 247-266. too,
Cain, DJ .; Fend, S. V .; and Carter, J .L. 1989. Temporal and spatial variability of arsenic in benthic insects from Whitewood Creek, South Dakota, Inc. Mallard, G.E. and Ragone, S.E., eds., U.S. Geological Survey Toxic Substances Hydrology Program. Water-Resources Investigations Report 88-4220.257-268.
Elwood,
J.W.; Hildebrandt S.G.; and Beauchamp, JJ. 1976. Contribution of gut contents to
the concentration and body burden of elements in Tipula spp. from a
spring-fed stream. J. Fish. Res. Board Can. 33: 1930-1938.
Hare,
L.; Campbell, P.G.C.; Tessier, A.; and Belzile, N. 1989. Gut sediments in a
burrowing mayfly (ephemeroptera, Hexagenia limbata): their contribution
to animal trace element burdens, their removal, and the efficacy of a correction
for their presence. Can. J. Fish. Aquat. SCI. 46: 451-456.
Johns,
C. and Moore, J .N .1985. Metals in the bottom sediments of the lower Clark Fork
River reservoirs. In:Carlson, C.E. and Bahls, L.L., eds., Proceeding-Clark Fork
Symposium, Montana Academy of Sciences, Montana College of Mineral Sciences and
Technology. 74-88.
Krantzberg,
G.; and Stokes, P.M. 1988. The importance of surface absorption and pH in metal
accumulation in chironomids. Environ. Toxicol. Chem. 7: 653-670.
Langston,
W. J .1986. Metals in sediments and benthic organisms in the Mersey Estuary.
Estuarine Coast. Shelf Sci. 23: 239-261.
Leenaers,
H. 1989. Downstream changes of total and partitioned metal concentrations in the
flood deposits of the River Geul (The Netherlands). Geojournal19: 37-43.
Luoma.
S.N. 1989. Can we determine the biological availability of sediment-bound trace
elements? Hydrobiologia 176/177: 379-396.
Luoma,
S .N. ; Axtmann, E. V .; and Cain, D J .1989. Fate of mine wastes in the Clark
Fork River, Montana, USA. In: Metals and metalloids in the hydrosphere; Impact
through mining and industry; and prevention in tropical environments.
International Hydrological Programme, UNESCO, Bangkok. 63-75.
Luoma
S.N.; and Bryan, G.W. 1978. Factors controlling the availability of
sediment-bound Pb to the estuarine bivalve Scrobicularia plana. Journal
Mar. Biol. Assoc. U. K. 58: 793-802.
Luoma,
S.N.; and Bryan, G.W. 1981. A statistical assessment of the form of trace metals
in oxidized estuarine sediments employing chemical extractants. Sci. Total
Environ. 17: 165-196.
Luoma,
S.N.; and Bryan, G. W .1982. A statistical study of environmental factors
controlling concentrations of heavy metals in the burrowing bivalve Scrobicularia
plana and the polychaete Nereis diversicolor. Estuarine Coast. and
Shelf Sci. 15: 95-108.
Moore,
J.N.; and Luoma, S.N. 1990a. Hazardous wastes from large-scale metal extraction.
Environ. Sci. and Tech. 24: 1278-1285.
Moore,
J.N.; and Luoma, S.N. 1990b. Hazardous wastes from large-scale metal extraction:
The Clark Fork Waste Complex. (this volume).
Moore,
J.N.; Luoma, S.N.; and Peter, D: 1990. Downstream effects of mine effluents on
an intermontane riparian system. Can. J. Fish. Aquat. Sci press).
Phillips,
G.R. 1985. Relationship among fish populations, metal concentrations and stream
discharge in the upper Clark Fork River .In: Calson, C.E. and Bahls, L.L.
(editors). Proceedings-Clark Fork Symposium. Montana Academy of Sciences,
Montana College of Mineral Sciences and Technology 57- 73. 22
Rice,
P.M.; and Ray, GJ. 1985. Heavy metals in flood plain deposits along the upper
Clark Fork River. In: Carlson, C.E. and Bahls, L.L. (editors) Proceedings-Clark
Fork Symposium. Montana Academy of Science, Montana College of Mineral Sciences
and Technology. 26-45.
Smock, L.A. 1983a. Relationships between metal concentrations and organism size in aquatic insects. Freshwater Biol. 13: 313-321.
Smock,
L.A. 1983b. The influence of feeding habits on whole-body metal concentrations
in aquatic insects. Freshwater Biol. 13: 301-311.
Sokal, R.R.; and Rohlf,F.J.1969. Biometry. W. H. Freeman and Co., San Francisco 776pp.