response of benthic invertebrate assemblages to metal ... · 1994). some insect taxa accumulate...

598

J. N. Am. Benthol. Soc., 2003, 22(4):598–620q 2003 by The North American Benthological Society

Response of benthic invertebrate assemblages to metal exposure andbioaccumulation associated with hard-rock mining in

northwestern streams, USA

T. R. MARET1,3, D. J. CAIN2, D. E. MACCOY1, AND T. M. SHORT2

1US Geological Survey, 230 Collins Road, Boise, Idaho 83702 USA2US Geological Survey, 345 Middlefield Road, Menlo Park, California 94025 USA4

Abstract. Benthic macroinvertebrate assemblages, environmental variables, and associated minedensity were evaluated during the summer of 2000 at 18 reference and test sites in the Coeur d’Aleneand St. Regis River basins, northwestern USA as part of the US Geological Survey’s National Water-Quality Assessment Program. Concentrations of Cd, Pb, and Zn in water and (or) streambed sedimentat test sites in basins where production mine density was $0.2 mines/km2 (in a 500-m stream buffer)were significantly higher than concentrations at reference sites. Zn and Pb were identified as theprimary contaminants in water and streambed sediment, respectively. These metal concentrationsoften exceeded acute Ambient Water Quality Criteria for aquatic life and the National Oceanic andAtmospheric Administration Probable Effect Level for streambed sediment. Regression analysis iden-tified significant correlations between production mine density in each basin and Zn concentrationsin water and Pb in streambed sediment (r2 5 0.69 and 0.65, p , 0.01). Metal concentrations incaddisfly tissue, used to verify site-specific exposures of benthos, also were highest at sites down-stream from intensive mining. Benthic invertebrate taxa richness and densities were lower at sitesdownstream than upstream of areas of intensive hard-rock mining and associated metal enrichment.Benthic invertebrate metrics that were most effective in discriminating changes in assemblage struc-ture between reference and mining sites were total number of taxa, number of Ephemeroptera, Ple-coptera, and Trichoptera (EPT) taxa, and densities of total individuals, EPT individuals, and metal-sensitive Ephemeroptera individuals.

Key words: mining, metal exposure, macroinvertebrates, bioaccumulation, structural response.

Mining activities over the last century haveprofoundly altered water-quality, aquatic-bio-logical, and hydrologic conditions in the Coeurd’Alene (CDA) River basin (Ellis 1940, Hoilandet al. 1994, Horowitz et al. 1995, Woods andBeckwith 1997, Maret and Dutton 1999). Fromthe late 1800s to early 1980s, the CDA MiningDistrict, located in the CDA valley of northernIdaho, was among the USA’s leading producersof Pb, Ag, and Zn. These past mining activitieshave resulted in US Environmental ProtectionAgency (USEPA) Superfund investigations andremediations in the South Fork CDA River. Thesecond-largest USEPA Superfund site is locatedin the study area—the Bunker Hill Smelter nearKellogg, Idaho (Doppelt et al. 1993). In addition,the US Department of the Interior is currentlyconducting Natural Resource Damage Assess-ments and other remediation activities in theCDA River basin. Despite treatment and clean-

3 E-mail address: [email protected] Any use of trade, product, or firm names in this

publication is for descriptive purposes only and doesnot imply endorsement by the US Government.

up of mine wastes, Cd, Pb, and Zn concentra-tions continue to exceed the USEPA AmbientWater Quality Criteria (AWQC) and streambedsediment guidelines for the protection of aquaticlife (Farag et al. 1998, Brennan et al. 1999).

Metals can affect aquatic organisms as toxicsubstances in water and sediment, or as a toxi-cant in the food chain (Sorensen 1991, Rainbow1996). Metals associated with mining activitiesnear CDA River basin streams have reduced na-tive fish species richness and abundance (Maretand MacCoy 2002), reduced survival andgrowth of fish (Farag et al. 1999), and reduceddiversity and densities of benthic invertebratespecies (Hoiland and Rabe 1992, Hoiland et al.1994).

Some insect taxa accumulate metals in pro-portion to metal concentrations in their environ-ment and, therefore, are effective as biomonitors(Cain et al. 1992, Hare 1992). However, interpre-tation of bioaccumulation data is complicated bythe presence of both absorbed and adsorbedforms of metals, with the latter form not incor-porated into tissue but contributing to the over-

2003] 599HARD-ROCK MINING EFFECTS ON BENTHIC INVERTEBRATES

all body concentration. Metals associated withintracellular tissues or biochemical fractions areunambiguous indicators of exposure to biolog-ically available forms of metals. Analysis of met-als in cytosol of resident invertebrates is a sim-ple approach to assess site-specific metal bio-availability (Cain and Luoma 1998).

Metals in invertebrate tissues also represent aconcentrated source that may be toxic in the dietof fish (Woodward et al. 1994). Farag et al.(1999) concluded that ingestion of metal-con-taminated invertebrates may be the principalroute of metal exposure to fish in streams of theCDA River basin.

Responses of benthic invertebrate assemblag-es have been used to assess ecological impactsof metal contamination in streams (Kiffney andClements 1994, Poulton et al. 1995, Beltman etal. 1999, Deacon et al. 2001, Mebane 2001). Re-duction in the abundance, diversity, and com-position of functional-feeding groups of mac-roinvertebrates are related to elevated metalconcentrations associated with mining (Cle-ments et al. 2000, Mebane 2001). Taxa richnessand abundances of metal-sensitive taxa areamong the most reliable measures of metal ef-fects on invertebrate assemblages (Carlisle andClements 1999). Metal-sensitive taxa have beenidentified in studies of Rocky Mountain streamsof Colorado (Clements 1994, Clements et al.2000) and Montana (McGuire 2001). These stud-ies showed that specific taxa of Ephemeropteraare highly sensitive to metal contamination, andthat their abundances in mining-affectedstreams can be a reliable measure of metal im-pacts on benthic invertebrate assemblages.

Studies based on spatially extensive, synopticsampling of multiple reference and contaminat-ed sites are one way to characterize metal im-pacts on stream biota (Clements et al. 2000). Themost effective assessments of anthropogenicstressors account for the influence of naturalstressors (Hughes et al. 1986). Information fromthese types of studies provides a basis for iden-tifying attainable conditions that can be used toset remediation goals for stream habitats affect-ed by past mining activities.

Benthic invertebrate assemblages, selected en-vironmental variables, and associated basinmine density were evaluated at 18 stream sitesin the CDA and St. Regis River basins duringthe summer of 2000 as part of the US GeologicalSurvey (USGS) National Water-Quality Assess-

ment (NAWQA) Program. The objectives of ourpaper are to 1) describe the extent and severityof hard-rock mining impacts on benthic inver-tebrates; 2) evaluate correlations among metalconcentrations in surface water and streambedsediment, metal concentrations in benthic inver-tebrates, and changes in assemblage structure;and 3) characterize the biotic and abiotic con-ditions of reference streams in the study area tohelp resource managers formulate recoverygoals and implement effective remediation ac-tions. We test the hypothesis that streams af-fected by mining and associated metals supportfewer invertebrate taxa and lower abundancesbecause invertebrates are exposed to elevatedconcentrations of metals.

Study Area

The study area in northern Idaho and westernMontana (Fig. 1A) is entirely in the NorthernRockies ecoregion (Omernik and Gallant 1986).Land use for all study basins ranges from 85 to99% forested land, and the remainder consistsprimarily of rangeland. Agricultural and urbanland uses constitute ,1% of all study basins.The study area has a complex geologic historyof sedimentation, compressional deformation,igneous activity, and, most recently, extensionalblock faulting (Kendy and Tresch 1996). Rocktypes in the study area are almost entirely me-tasedimentary. Because of the extensively min-eralized rocks, metal concentrations can be ele-vated in streambed sediment in undisturbedstreams (Maret and Skinner 2000).

Streams in the study area are exclusivelycoldwater. The main source of surface andground water is snowmelt runoff from April toJuly. Streamflow conditions during the studyperiod were similar to or below long-term av-erage conditions. The mean August streamflowfor 2000 was ;84% of the long-term averagestreamflow for the North Fork CDA River nearEnaville and the South Fork CDA River nearPinehurst (Fig. 1A, sites 10 and 17). Streams inthe study area typically have high water clarity,coarse-grained substrates (cobble and boulders),high gradients (.0.5%), well-defined riffles andpools, and very sparse macrophyte growth.Mining and logging practices, channelization,and roads have affected riparian areas, reducedinstream habitat, and degraded stream and

600 [Volume 22T. R. MARET ET AL.

FIG. 1. A. Study area (shaded), sampling sites, and production mines (active and inactive) in the Coeur d’Aleneand St. Regis River basins. B.—South Fork Coeur d’Alene and St. Regis River basins showing the 500-m streambuffer used to quantify production mine density in the study basins. See Table 1 for site names.


TABLE 1. Sampling sites, production mine density in the basin and in a 500-m buffer (250 m from eachbank) upstream from each site, and site type, Coeur d’Alene and St. Regis River basins. R 5 reference site, T5 test site. Site numbers are shown in Fig. 1.

Sitenum-ber Site name

Production minedensity/km2 a

BasinBuffer

(500 m)Sitetype

12345

St. Regis River above Rainy Creek, MTSt. Regis River near Haugan, MTSt. Regis River near St. Regis, MTNorth Fork Coeur d’Alene River near Prichard, IDWest Fork Eagle Creek below Settlers Grove, ID

0.0400.0390.0250.0010.068

0.0000.0750.0580.0030.162

RRRRR

6789

10

East Fork Eagle Creek near Murray, IDUpper Prichard Creek near Murray, IDPrichard Creek at Prichard, IDBeaver Creek near Murray, IDNorth Fork Coeur d’Alene River near Enaville, ID

0.0760.2180.1540.1800.027

0.2000.3540.3150.2000.034

TTTTR

11121314

South Fork Coeur d’Alene R above Mullan, IDCanyon Creek near Burke, IDCanyon Creek at Woodland Park, IDSouth Fork Coeur d’Alene River at Silverton, ID

0.1760.1450.4200.314

0.0000.1971.0380.498

RRTT

15161718

East Fork Pine Creek above Nabob Creek near Pinehurst, IDPine Creek below Amy Gulch near Pinehurst, IDSouth Fork Coeur d’Alene River near Pinehurst, IDSt. Joe River at Red Ives Ranger Station, ID

0.1620.2180.2330.011

0.3960.4830.4120.032

TTTR

a Production mines (active and inactive) taken from US Bureau of Mines (1995)

groundwater quality (Woods and Beckwith1997, Beckwith 1998).

More than 250 active and inactive hard-rockmines of various sizes are located in the studyarea (Fig. 1A), and many of these mines are lo-cated near streams. Production mine densityranges from 0.001 to 0.42/km2 (Table 1). His-torically, metal extraction and processing wererelatively inefficient, yielding large volumes ofmetal-rich tailings that were deposited in andaround nearby streams. Mine tailings in thestudy area typically contain elevated concentra-tions of trace metals such as As, Cd, Cu, Pb, Hg,and Zn (Woods and Beckwith 1997). These tail-ings and mines continue to provide a source oftrace metals to streams, lakes, and reservoirs asstreams meander through and erode tailings de-posits and transport them downstream. Minetailings entering the South Fork CDA River havebeen transported and deposited along the riverchannel and floodplain into CDA Lake anddownstream into the Spokane River (Woods2000, Grosbois et al. 2001). There is concernthese elevated metals from upstream miningsources may be affecting aquatic life and bioac-

cumulating in fish tissue at levels exceeding hu-man consumption guidelines for the SpokaneRiver.

Methods

Site selection

Sampling sites in the CDA and adjacent St.Regis River basins were selected to representstreams affected by a range of mining activities,from minimally disturbed reference sites (littleor no prior mining activity in the basin) to high-ly disturbed test sites (affected by cumulativemining impacts in the basin). Criteria suggestedby Hughes et al. (1986) were used to select ref-erence sites: 1) examination of existing data, 2)consultation with local land management agen-cies familiar with streams in the area, and 3)reconnaissance of candidate sites. Referencesites represented regional conditions (site 18),conditions upstream from major mining activi-ties (sites 4, 5, 11, and 12), and conditions insimilar (paired) basins (reference sites 1, 2, and3 in the St. Regis River basin, reference site 11,


and test sites 14 and 17 in the South Fork CDARiver basin). To reduce the effects of stream size,each pair of sites in the South Fork CDA and St.Regis River basins were approximately the samedistance (river km) from the mouth of the St.Regis or the South Fork CDA River. Both basinsare similar in size, and site pairs are similar inelevation and stream geomorphology (see Rei-ser 1999). The location of test sites was based onthe location of production mines identified inthe US Bureau of Mines (1995) Minerals Avail-ability System.

Eighteen sites were selected for sampling (Ta-ble 1). A site consisted of a representative reachgenerally containing repeating geomorphicchannel units (riffles, runs, and pools). Becausethe length of stream sampled was a function ofstream width, reach lengths ranged from 150 to500 m (Fitzpatrick et al. 1998). The stream sitesrepresented 2nd- through 5th-order streams(Strahler 1957). All stream sites were wadeableexcept sites 10 and 17.

Environmental variables

Environmental variables were evaluated foreach site (Tables 1, 2) and consisted of basin andreach characteristics, physical habitat, and waterand streambed sediment physicochemistry.

Geographic characterization. Basin area andproduction mine density were determined us-ing Arc/Info, a geographic information system(GIS). Several sources were used to constructthe geographic data layers. Basin boundarieswere constructed using hydrography and hy-drologic unit boundary data layers (USGS 1975).Because most of the mining activities in thisarea occurred near streams, a proximity analy-sis also was conducted on mine locations rela-tive to streams in each basin. A 500-m bufferwidth (250 m from each bank) was selected forthe entire drainage network (at 1:100,000 scale)upstream from each site (Fig. 1B) after extensiveanalysis and comparison of different bufferwidths (including none). This buffer width bestcharacterized mining activities directly affectinglocal instream conditions. The buffer includedonly ;35% of the total study basin area but in-cluded 59% (158 of 269) of all production mines.Site elevations were derived from USGS 1:24,000-scale Digital Raster Graphic maps. Infor-mation on production estimates of individualmines was unavailable.

Physical habitat. Instream habitat variableswere measured following procedures outlinedby Fitzpatrick et al. (1998). Stream gradient wasdetermined onsite for each reach by using a rodand level. Discharge at the time of sampling wasmeasured onsite using a Marsh–McBirney flowmeter or was determined from streamflow-gauging records. Stream width was measuredwith a laser rangefinder at 6 locations through-out the reach. At each riffle location where in-vertebrates were sampled, estimates of depth, %open canopy, velocity, and % embeddednesswere measured or estimated. Velocities weremeasured at 0.6 of the stream depth. Percentopen canopy was determined by measuring leftand right canopy angle using a clinometer. Per-cent embeddedness was visually estimated tothe nearest 10%. Dominant substrate size wasvisually categorized into size classes (i.e., silt,sand, gravel, cobble, and boulder). For data anal-ysis, a mean value was calculated for each hab-itat variable at each site.

Water and sediment chemistry. Water-qualityand streambed sediment samples were collectedat each site during low-flow conditions in Au-gust and September 2000 using methods de-scribed by Shelton and Capel (1994). Water-quality samples were collected for analysis ofselected nutrients and trace metals. A portion ofeach sample was filtered through a 0.45-mm fil-ter. Analytes in the filtered aliquot are hereafterreferred to as dissolved. Samples were shippedon ice to the USGS National Water Quality Lab-oratory (NWQL) in Arvada, Colorado, for anal-ysis (Fishman 1993). Water temperature, dis-solved oxygen, pH, and specific conductancealso were measured onsite (Wilde and Radtke1998).

Streambed sediment was sampled at 5 to 10locations within a reach. The upper 2 cm of bedsediment was collected from undisturbed, con-tinuously wetted, depositional zones using aplastic scoop. The subsamples were compositedand wet-sieved through a 63 mm nylon mesh.Sediment ,63 mm was retained to normalizethe size fraction among sites. This sieved sam-ple was placed on ice and submitted to theNWQL for analysis. Metal analysis of the sedi-ment focused on the partially extractable frac-tion (5% hydrochloric acid extraction) (Hornber-ger et al. 1999). This method extracts metalsfrom the sediment surface and represents a met-al fraction that has high potential for exposure


TA

BL

E2.

Env

iron

men

talv

aria

bles

from

the

Coe

ur

d’A

lene

and

St.R

egis

Riv

erba

sins

.Var

iabl

esin

bold

are

sign

ifica

ntly

dif

fere

nt(p

,0.

05)

betw

een

site

typ

es,

base

don

aM

ann–

Whi

tney

t-te

st.A

llm

easu

rem

ents

ofw

ater

and

stre

ambe

dse

dim

ent

phys

icoc

hem

istr

yw

ere

mad

edu

ring

low

flow

.SD

5st

anda

rdde

viat

ion.

Vari

able

Site

typ

e

Ref

eren

ce(n

59)

Mea

n(S

D)

Med

ian

Ran

ge

Test

(n5

9)

Mea

n(S

D)

Med

ian

Ran

ge

Env

iron

men

tal

Bas

inar

ea(k

m2 )

Sit

eel

evat

ion

(m)

Gra

die

nt(%

)D

isch

arge

(m3 /

s)W

idth

(m)

517

(756

)95

4(1

87)

0.89

(0.7

5)2.

05(2

.54)

18.6

(16.

7)

233

1039 0.

521.

3010

.3

17–2

325

664–

1155

0.06

–2.3

00.

13–8

.02

4.8–

53.6

194

(220

)78

3(7

7)1.

00(0

.49)

0.82

(1.0

6)10

.3(4

.4)

100

798 1.

200.

458.

1

52–7

3866

8–90

81.

08–1

.34

0.04

–3.3

46.

1–19

.2D

epth

(m)

Op

enca

nopy

(%)

Velo

city

(m/

s)E

mbe

dde

dne

ss(%

)D

omin

ant

subs

trat

e(m

m)

0.21

(0.0

6)51

(20)

0.63

(0.1

7)7

(7)

78(3

6)

0.20

43 0.63

8 96

0.14

–0.3

224

–80

0.27

–0.8

30–

208–

96

0.22

(0.0

9)57

(5)

0.60

(0.1

5)13

(14)

96(0

)

0.19

57 0.59

5 96

0.12

–0.3

951

–68

0.36

–0.8

00–

4296

Spec

ific

cond

ucta

nce

(mS/

cm)

Dis

solv

edox

ygen

(mg/

L)

Wat

erte

mp

erat

ure

(8C

)p

H

51(2

2)9.

7(0

.5)

11.5

(2.6

)7.

4(0

.3)

52 9.8

11.3 7.4

15–9

38.

9–10

.38.

0–15

.06.

8–7.

7

89(8

0)9.

2(0

.5)

15.1

(2.2

)7.

1(0

.4)

50 9.0

15.0 7.0

30–2

728.

5–10

.012

.6–1

8.9

6.6–

7.8

Har

dne

ss(m

g/L

)To

tal

NO

3(m

g/L

)To

tal

P(m

g/L

)N

H3,

dis

solv

ed(m

g/L

)

21(9

)0.

015

(0.0

07)

,0.

008

0.00

3(0

.001

)

22 0.00

9,

0.00

80.

002

6–33

0–0.

021

,0.

008

0.00

2–0.

004

38(3

7)0.

043

(0.0

67)

,0.

008

(0.0

24)

0.03

0(0

.084

)

19 0.01

3,

0.00

80.

002

10–1

200–

0.20

8,

0.00

8–0.

071

0.00

2–0.

254

Cd D

isso

lved

,wat

er(m

g/L

)S

trea

mb

edse

dim

ent

(mg/

g)a

,1

1(0

.5)

,1 0.

9,

10.

3–1.

93

(5)

24(4

1),

1 3,

1–13

2–12

4

Pb D

isso

lved

,wat

er(m

g/L

)S

trea

mb

edse

dim

ent

(mg/

g)a

,1

52(4

9),

1 41,

13–

125

5(1

0)18

51(2

394)

,1

358

0.5–

3013

2–62

52

Zn D

isso

lved

,wat

er(m

g/L

)S

trea

mb

edse

dim

ent

(mg/

g)a

4(4

)69

(71)

3 660–

122–

199

503

(630

)20

22(2

724)

110

322

16–1

510

112–

7346

aPa

rtia

llyre

cove

rabl

ean

alys

is(w

eak

acid

extr

acti

on)

of,

63-m

msi

zefr

acti

on


to and ingestion by resident biota (Luoma 1989,Cain et al. 1992, Luoma et al. 1995). Metals weredetermined by inductively coupled plasma op-tical emission spectrophotometry (Hornbergeret al. 1999).

Quality assurance. Field quality assurance(QA) consisted of blank water samples and du-plicate sediment samples. Laboratory QA con-sisted of routine blank and replicate samples forwater analyses (Fishman 1993) and analyses ofreference material for sediment samples (Arbo-gast 1990). Concentrations of metals analyzed inall blank water samples were below the mini-mum reporting limit. Concentrations of Cd, Pb,and Zn in duplicate sediment samples had rel-ative differences of 0 to 15%, and concentrationsin reference material were within QA guidelinesof 63 SD. All other analytical results for QAlaboratory samples were within acceptableUSGS method standards.

Benthic invertebrate collection and analysis

Invertebrate sampling. Benthic invertebrateswere collected at all sites during low streamflowconditions in August and September 2000. Sam-ples from 5 separate riffles were collected andcomposited throughout each reach using a Slackrectangular kicknet sampler (total area of 1.25m2) equipped with a 425-mm-mesh net (Cuffneyet al. 1993). Large gravel and cobbles within thesample collection area were brushed to dislodgeorganisms, and the entire area was disturbed bykicking for 30 s. Onsite processing consisted ofbucket elutriation of each sample by repeatedwashing through a 425-mm-mesh sieve. Sampleswere placed in 1-L plastic jars and fixed with10% buffered formalin. Invertebrate taxonomicand abundance data were provided by the Bio-logical Group at the NWQL (Moulton et al.2000). Organisms were identified to the lowestpractical taxonomic level (genus or species formost taxa). Voucher collections of invertebratesare deposited at the NWQL.

Metrics. Benthic invertebrate assemblageswere analyzed using relative abundance dataand metrics useful in evaluating structuralchanges resulting from elevated concentrationsof metals in Rocky Mountain streams: 1) num-ber of total taxa; 2) number and % Ephemer-optera, Plecoptera, and Trichoptera (EPT) taxa;3) number and % of Ephemeroptera taxa; 4)number of metal-sensitive Ephemeroptera taxa;

5) density (number/m2) of all taxa; 6) density ofEPT taxa; 7) density of metal-sensitive Ephem-eroptera (Hoiland et al. 1994, Kiffney and Cle-ments 1994, Carlisle and Clements 1999, Cle-ments et al. 2000). We used metal-tolerancerankings from 0 (sensitive) to 10 (tolerant) de-veloped by McGuire (2001) for Ephemeropterain the Clark Fork basin, Montana. In our study,metal-sensitive Ephemeroptera (tolerance values#2) included Ameletus spp., Caudatella spp., Dru-nella spp., Serratella spp., Timpanoga hecuba, Cin-ygmula spp., Epeorus spp., Rhithrogena spp., andLeptophlebiidae.

Taxa were assigned to functional-feedingguilds (i.e., filterers, gatherers, predators, scrap-ers, and shredders) following guidelines of Mer-ritt and Cummins (1996). Relative abundancesof each feeding guild were used in the analyses.

Tissue analysis. The net-spinning caddisfliesCeratopsyche spp., Hydropsyche spp., and Arcto-psyche grandis were collected with kicknets andby hand from riffle habitats throughout eachreach for tissue analysis. Caddisflies initiallywere placed in plastic containers (previouslyacid rinsed), cleaned with stream water, andsorted onsite. Sorted samples were transferredto sealable plastic bags with a small volume ofriver water, placed on dry ice, and stored at2708C in the laboratory.

Caddisfly larvae were partially thawed inbatches, rinsed with cold, deionized water to re-move adhering particles, and then transferred toa chilled sorting dish for identification (Merrittand Cummins 1996). Larvae were then imme-diately transferred to an ice cooler, and com-bined randomly into subsamples, each weighingbetween 0.4 and 1.6 g wet mass. One to 6 sub-samples from each site were processed for metalanalysis.

Larvae were homogenized with a stainless-steel, high-speed, tissue homogenizer in cold,0.05-M Tris-hydrochloric buffer (pH 7.4, previ-ously degassed and bubbled with N2) under anN2 atmosphere for 1 min. An aliquot of the ho-mogenate was collected to obtain a whole-bodyconcentration, weighed, and then frozen at2408C. The cytosol then was isolated from a2nd aliquot of the homogenate by differentialcentrifugation following procedures modifiedfrom Wallace et al. (1997) and Cain and Luoma(1998).

All tissue samples were prepared for elemen-tal analysis using procedures described by Cain


and Luoma (1998). Samples were analyzed byinductively coupled plasma optical emissionspectrophotometry. The median CV for replicateanalysis of whole-body and cytosol samples forthe reported metals ranged from 5 to 15% and6 to 19%, respectively. QA consisted of analysesof procedural blanks, National Reference Coun-cil of Canada Tort-2 reference material, and met-al-spiked (predigestion) blanks and samples.Sample concentrations were not corrected forblanks because procedural blanks for Cd and Pbwere below the method detection limits (0.3 and2.0 mg/L, respectively), and measurable Znblanks (.0.3 mg/L) were ,5% of the sampleconcentrations. Measured concentrations inspiked blanks and samples were within 7% ofnominal concentrations. Relative SD for mea-sured concentrations of Cd and Zn in Tort-2 ref-erence material were 2 and 4%, respectively, andthe mean concentrations 63 SD were within thecertified concentrations. Lead concentrations inprepared Tort-2 reference materials were at ornear the method detection limits and, thus, werenot quantifiable.

Data analysis

Site categorization. Sampling sites were cate-gorized as either reference or test sites on thebasis of production mine density upstream fromeach site and concentrations of metals in waterand sediment. The median mine density of 0.2mines/km2 in the 500-m buffer was initiallyused to divide sites into 2 equal groups of ref-erence (low mining density) and test (high minedensity) sites. Test sites were finally selected bysatisfying 2 criteria: 1) production mine densityof $0.2 mines/km2 in the stream buffer (Table1) and, 2) elevated metal concentrations in waterand (or) sediment that exceeded one or more ofthe guidelines for the protection of aquatic life.Although the North Fork CDA River near Ena-ville (site 10) is located downstream of 4 testsites, it was categorized as a reference site be-cause metal concentrations did not exceed Ida-ho’s Water-Quality Standards (Woods 2000).Only Cd, Pb, and Zn concentrations are report-ed in our study because they have been identi-fied as the metals most elevated in CDA Riverbasin streams and are of greatest ecological con-cern (Brennan et al. 1999).

Water and sediment toxicity. Metals in waterwere compared with USEPA (2000) acute

AWQC for the protection of aquatic life. Dis-solved, rather than total recoverable, concentra-tion was evaluated because the dissolved frac-tion more accurately reflects site-specific metalbioavailability and toxicity (Prothro 1993). Be-cause water hardness affects bioavailability andtoxicity of most metals, criterion values weremodified to account for variation in water hard-ness among streams (USEPA 2000). Chronic cri-teria were not evaluated because they are basedon 4-d average metal concentrations, whichwere not collected during our study. Our studymay not accurately describe Cd criteria exceed-ances because of the low water hardness andhigh laboratory detection limit (1.0 mg/L). Onoccasions where hardness is ,30 mg/L, acutecriteria of ,1.0 mg/L for Cd are possible.

National Oceanic and Atmospheric Adminis-tration (NOAA) screening values were used toevaluate the severity of the metal contaminationin sediment (Buchman 1999). The Probable Ef-fect Level (PEL) for each metal, defined as theconcentration above which adverse effects toaquatic life are predicted to occur frequently,was used. Because the ,63-mm size fraction,rather than bulk sediment (used to developthese guidelines), was analyzed, the likelihoodof exceeding a PEL may be increased.

Cumulative toxic units (CTU) for Cd, Pb, andZn in water and sediment were calculated forour study similarly to Clements et al. (2000).Acute AWQC and the NOAA PEL values wereused as toxicity reference values (TRV), similarto the approach used by Dyer et al. (2000). EachCTU value represents the sum of Cd, Pb, andZn concentrations in water and sediment at eachsite divided by the appropriate TRV. A CTU val-ue .1.0 represents potentially toxic levels of cu-mulative Cd, Pb, and Zn concentrations in wateror sediment at a site, and provides resourcemanagers with a measure of the severity of met-al contamination in the streams sampled.

Bioaccumulation data. Not all genera of cad-disflies were collected at every site. Previousstudies (Cain et al. 1992, Cain and Luoma 1998),and analysis of samples from our study (AN-OVA) indicated that interspecific differences inbioaccumulation would not confound among-site comparisons of metal exposure. Therefore,Ceratopsyche spp. and Hydropsyche spp. weretreated as a single taxon, Hydropsyche spp., andmetal data for Hydropsyche and A. grandis were


combined at sites where both taxa were collect-ed.

Statistics. All statistical analyses were per-formed using SYSTAT (L. Wilkinson. 1999. SYS-TAT for Windows statistics, version 9.0, SYSTAT,Inc., Evanston, Illinois.). All possible combina-tions of responses and explanatory variableswere examined using Spearman’s rank correla-tion matrices. Regression analyses betweenmine densities and metal concentrations wereused to predict metal concentrations and poten-tial biological impacts. Correlations betweenmine density and the sum of Cd, Pb, and Znconcentrations in water and streambed sedi-ment were described using regression analyses.The LOWESS (LOcally WEighted ScatterplotSmoothing) technique, a robust, nonparametricdescription of data patterns (Helsel and Hirsch1992), was used to evaluate trends between ben-thic invertebrate metrics and CTUs for metals inwater and sediment. Significant differences inthe environmental variables and benthic inver-tebrate metric data between reference and testsites were determined with the nonparametricMann–Whitney test. Nondetections of chemicalconstituents were set to ½ the minimum report-ing limit prior to statistical analysis. Results ofstatistical tests were considered significant if p, 0.05.

Results

Environmental variables

Site elevation and water temperature were theonly environmental variables (excluding Cd, Pb,and Zn concentrations) significantly differentbetween reference and test sites (Table 2). Me-dian elevations were significantly higher at ref-erence sites (1039 m) than at test sites (798 m)because some reference sites necessarily werelocated upstream from mining areas. Medianwater temperatures were significantly lower inreference sites (11.38C) than test sites (15.08C).Moreover, unlike many Rocky Mountainstreams affected by acid-mine waste, pH of thestream water at the test sites was near neutraland ranged from 6.6 to 7.8.

Cd and Zn in water and streambed sedimentand Pb in streambed sediment were signifi-cantly higher at test sites than at reference sites(Table 2). More than 75% of the concentrationsof dissolved Cd and Pb in water were less than

the reporting limit; only at test sites 13 (CanyonCreek at Woodland Park), 15 (East Fork PineCreek), and 14 and 17 (South Fork CDA at Sil-verton and near Pinehurst) were concentrationsof one or both of these metals quantifiable. Met-al concentrations in water and sediment oftenexceeded the acute AWQC and PEL, and werehighest at test sites 13, 15, 14, and 17. However,metal concentrations in sediment at referencesites 1, 10, 11, and 12 also exceeded the PEL of91.3 and 315 mg/g for Pb and Zn, respectively.Concentrations of dissolved metals in water atreference sites did not exceed acute AWQC.

Concentrations of Zn in water and Pb in sed-iment were significantly correlated (r2 5 0.69and 0.65, p , 0.01) with the number of minesin close proximity to streams upstream fromsampling sites (Fig. 2A, B). Metal concentrationsin both water and sediment at test sites gener-ally exceeded concentrations at reference sites,some by as much as 1 to 2 orders of magnitude.Mean Zn concentrations in water were 503 and4 mg/L at test and reference sites, respectively(Table 2). Similarly, mean Pb concentrations insediment were 1851 and 52 mg/g at test andreference sites, respectively.

The relative severity of Cd, Pb, and Zn con-tamination in water and sediment is shown byranking all sites using the CTU values (Fig. 3A,B). CTUs for water and sediment at all siteswere significantly correlated (r 5 0.93). Zn con-tributed the most CTUs for water and Pb con-tributed the most CTUs for sediment. CTUs var-ied greatly among sampling sites, ranging from1 to 35 for water and 1 to 111 for sediment.CTUs for most reference sites did not exceed 1,indicating that cumulative effects from Cd, Pb,and Zn contamination are unlikely. CTUs forwater and sediment were highest at sites 13, 14,and 17.

Metal concentrations in caddisflies

Cd, Pb, and Zn concentrations in whole-bodycaddisfly tissue were significantly higher at testsites than at reference sites (Table 3). Pb in thecytosolic fraction was also significantly higherat test sites than at reference sites. Caddisflieswere absent at 5 of the 9 test sites. Metal con-centrations in the whole body and cytosol werehighest at test sites 14 and 17 (Fig. 4A–C). Con-centrations in the cytosol of caddisflies at thosesites were generally 1 to 2 orders of magnitude


FIG. 2. Relationship between production mine density (in 500-m stream buffer) and Zn concentrations in water(A) and Pb in streambed sediment (B), Coeur d’Alene and St. Regis River basins. For details on sampling sitessee Table 1 and Fig. 1. Data were fitted using a linear regression. Dashed lines are the 95% confidence intervals.


FIG. 3. Ranking of sampling sites on the basis of cumulative toxic unit (CTU) values for Cd, Pb, and Znconcentrations in water (A) and sediment (B), Coeur d’Alene and St. Regis River basins. A CTU value .1.0represents potential cumulative toxicity of metals at a site. * 5 reference site.


TABLE 3. Metal concentrations in caddisfly tissue, Coeur d’Alene and St. Regis River Basins. Variables inbold are significantly different (p , 0.05) between site types, based on a Mann–Whitney t-test. SD 5 standarddeviation.

Caddisfly tissuea,b

Site type

Reference(n 5 9)

Mean (SD) Median Range

Test(n 5 9)


Cd

Whole body (mg/g)Cytosol fraction (mg/g)

0.58 (0.70)0.28 (0.36)

0.300.15

0.08–2.320.05–1.21

6.68 (6.61)1.10 (1.20)

6.800.74

0.42–12.70.13–2.78

Pb


2.94 (2.83)0.61 (0.80)

1.370.24

0.20–7.090.04–2.56

142 (157)7.20 (10.90)

1142.48

11–3310.42–23.4

Zn


181 (69)68 (35)

17258

82–32627–148

707 (560)108 (82)

64284

180–136541–221

a Includes test sites 8(AH), 9(A), 14(AH), 17(AH); A 5 Arctopsyche grandis, H 5 Hydropsyche spp.b Includes references sites 1(A), 2(A), 3(A), 4(H), 5(A), 10(H), 11(A), 12(A), 18(A)

higher than at other sites. Cytosolic Pb concen-trations at site 14 were particularly high, ex-ceeding 20 mg/g. Metal concentrations in thewhole body and cytosol were significantly cor-related (r $ 0.85). However, the proportionalcontribution of the cytosol tended to decline aswhole body concentrations increased, reflectingthe importance of other metal accumulationsites (Wallace et al. 1997, Cain and Luoma 1998).Cd and Zn concentrations in caddisflies at ref-erence sites 1 (St. Regis River above RainyCreek), 4 (North Fork CDA River near Prichard),and 12 (Canyon Creek near Burke) were higherthan in samples at other reference sites. As men-tioned above, metal concentrations in sedimentat sites 1 and 12 also exceeded the PEL.

Metal concentrations in caddisfly tissue weresignificantly correlated with mine density andmetal concentrations in water and sediment (Ta-ble 4). Metal concentrations in the whole bodiesof caddisflies were most strongly correlatedwith sediment concentrations (r 5 0.88, 0.90,and 0.67, for Cd, Pb, and Zn, respectively). Cor-relations with water were significant for all met-als (r 5 0.62, 0.63, and 0.47, for Cd, Pb, and Zn,respectively), while correlations with produc-tion mine density were significant for Cd andPb (r 5 0.56, and 0.64, respectively). Similarly,metal concentrations in the cytosol increased inrelation to increasing concentrations of metals insediment and water; however, corresponding

changes in metal concentrations were not highlycorrelated (r , 0.40). These correlations proba-bly were affected by site-specific differences inthe partitioning of metals between the cytosoland whole-body fractions.

Benthic invertebrates and metrics

Ninety-one benthic invertebrate taxa wereidentified. The number of taxa collected rangedfrom 19 at site 13, the most contaminated site,to 38 at site 8 (Prichard Creek at Prichard). Baetistricaudatus was the most common taxon (collect-ed at all sites) and one of the most abundantorganisms collected (mean density of 1173/m2).Only Simulium spp. were more abundant (meandensity 1254/m2) at the 17 sites where theywere collected.

Total taxa richness, EPT taxa richness, anddensities of total individuals, EPT individuals,and metal-sensitive Ephemeroptera individualswere significantly higher at reference sites thanat test sites (Table 5). In addition, abundances ofinvertebrates at reference sites were approxi-mately twice those at test sites, for all densitymetrics. No significant differences in the per-centages of functional feeding groups were ev-ident between site types (data not shown).

Comparisons of paired basins revealed that,for reference sites 1, 2, and 3 in the St. RegisRiver basin and site 11 upstream from major


TABLE 4. Spearman correlation coefficients among production mine densities; metal concentrations in water,sediment, and caddisfly tissue; and selected benthic invertebrate metrics, Coeur d’Alene and St. Regis Riverbasins. * 5 p , 0.05.

Productionmines/km2 in500-m stream

buffer

Water (mg/L)

Cd Pb Zn

Concentration (mg/g) in caddisfly tissuea

CdPbZn

0.56*0.64*0.35

0.62*0.63*

0.47*

Metric

Total taxaEPT taxab

Total individuals (no./m2)EPT individualsb (no./m2)Metal-sensitive Ephemeroptera individuals (no./m2)

20.66*20.66*20.56*20.47*20.56*

20.63*20.64*20.3820.2820.66*

20.60*20.61*20.3420.2920.65*

20.75*20.77*20.4520.3820.62*

a Whole-body concentrations at 13 sitesb Sum of Ephemeroptera, Plecoptera, and Trichoptera taxa

mining impacts in the South Fork CDA Riverbasin, EPT taxa richness ranged from 10 to 19,and metal-sensitive Ephemeroptera taxa rich-ness ranged from 3 to 7 (Fig. 5). Test sites 14and 17 in the CDA River basin are affected bymining in Canyon and Nine Mile Creeks thatflow into the South Fork CDA River upstream.These tributaries carry high concentrations ofCd, Pb, and Zn, which increase metal loads intothe South Fork CDA River by an order of mag-nitude (Woods 2000). Fewer EPT (7 to 9) andmetal-sensitive Ephemeroptera taxa (0 to 1)were collected at test sites than at referencesites; however, the % composition of EPT washigher at test sites (86–93%) than at referencesites (41–76%). Baetis tricaudatus, Hydropsychespp., and A. grandis, which are relatively tolerantof metals (Clements 1994), were the predomi-nant EPT taxa at test sites 14 and 17. Metal-sensitive Ephemeroptera taxa composed 16 to22% of the assemblage at the reference sites butonly 0 to 3% at the test sites. Drunella doddsi wasthe only metal-sensitive Ephemeroptera taxoncollected at either of these test sites. This speciesalso was collected at site 13, the most metals-contaminated site (Fig. 3A).

Benthic invertebrate measures and contaminantrelations

The 5 benthic invertebrate metrics shown inTable 5 that were significantly different between

reference and test sites were sometimes signifi-cantly correlated with mine density and Cd, Pb,and Zn concentrations in water, sediment, andcaddisfly tissue (Table 4). Each metric was neg-atively correlated with $1 of the variables as-sociated with mining. Some of the strongest cor-relations were between the total number of taxa(r 5 20.75), EPT taxa (r 5 20.77), and Zn con-centrations in water. In addition, density of met-al-sensitive Ephemeroptera was significantly,negatively correlated with production minedensity and metal concentrations in caddisflies(whole-body), water, and streambed sediment(r 5 20.54 to 20.70).

Scatterplots (including LOWESS smoothing)summarizing relations among benthic inverte-brate metrics and CTUs for Cd, Pb, and Zn inwater clearly showed a declining trend in bioticcondition along a gradient of increasing metaltoxicity (Fig. 6). Because water and sedimentCTUs were strongly correlated (r 5 0.93), therewere also similar declining trends in biotic con-dition with increasing sediment CTUs. Corre-lation coefficients among CTUs for metals inwater and selected benthic invertebrate metricswere all significant. Total taxa, EPT taxa, anddensity of metal-sensitive Ephemeroptera taxawere the most responsive metrics to site-specificchanges in water and sediment toxicity. Thosemetrics were less variable than the abundancemetrics, total number of individuals and EPTindividuals.


TABLE 4. Extended.

Streambed sediment(mg/g)

Cd Pb Zn

Caddisfly tissuea

(mg/g)

Cd Pb Zn

0.88*0.90*

0.67*

20.62*20.56*20.56*20.4120.63*

20.62*20.58*20.64*20.50*20.62*

20.69*20.68*20.48*20.3820.60*

20.1220.2220.4320.47*20.70*

20.2120.2820.420.4220.70*

20.3220.50*20.0820.1520.54*

The number of invertebrate taxa generally de-clined with increasing water and sedimentCTUs, but measures of invertebrate densitywere highly variable among reference and testsites. For example, total and EPT densities wererelatively high at site 17 in spite of the fact thatCTUs for water and sediment were among thehighest of any test site. Unlike habitat at othersites, the predominant habitat at site 17 consist-ed of dense mats of filamentous algae. This sitewas dominated by hydropsychid caddisflies,composing ;67% of all organisms in the sam-ple.

Discussion

Our study showed broad agreement amongthe multiple lines of evidence (mine density,metal concentrations in water and sediment,bioaccumulation in caddisfly tissue, and benthicinvertebrate assemblage structure) used to eval-uate hard-rock mining impacts. The most pro-nounced differences were among reference sitesand some of the most metal-contaminated testsites, including sites 13, 14 and 17. Our studydemonstrated that streams downstream fromareas of intensive hard-rock mining containedfewer benthic invertebrate taxa and lower den-sities, and that these changes in assemblagestructure were associated with elevated metalcontaminants in both water and sediment.There were few significant differences betweenreference and test sites for the instream physical

habitat variables measured. Although watertemperatures differed significantly between sitetypes, temperatures at all sites were below Ida-ho’s coldwater temperature criterion of 198C(Grafe et al. 2002). Moreover, based on occur-rence patterns of Idaho’s coldwater obligate in-vertebrates, defined as those taxa that have,10% probability of occurring in streams wherewater temperatures exceeded 198C (Grafe et al.2002), thermal conditions at the test sites did notappear to be biologically limiting. For example,reference and test sites had at least 15 and 13coldwater obligate invertebrate taxa, respective-ly. In addition, coldwater fish species are pres-ent at all sites, except in lower Canyon Creek(Maret and MacCoy 2002). Furthermore, Reiser(1999) measured continuous summer water tem-peratures at numerous sites in the CDA Riverbasin and also concluded that water tempera-tures are likely not limiting to aquatic life.

Metal exposure

Production mine density in the upstream ba-sin and within a 500-m buffer was a useful in-dicator of metal exposure. Test sites, defined ashaving $0.2 mines/km2 in the 500-m buffer,had elevated concentrations of Cd, Pb, and Znin $1 exposure indicators (water, sediment, cad-disflies). Although production data for individ-ual mines might have enhanced our analysis,this information was not available.

Concentrations of Cd, Pb, and Zn at reference


TABLE 5. Benthic invertebrate metrics for all sampling sites, Coeur d’Alene and St. Regis River basins. Var-iables in bold are significantly different (p , 0.05) between sites, based on a Mann–Whitney t-test. SD 5standard deviation.

Metric

Site type

Reference(n 5 9)


Richness

Total taxaEPT taxaa

% EPTEphemeroptera taxa% Ephemeroptera

34 (5.2)17 (3.1)67 (14)7 (1.7)

38 (12)

351870

637

21–3710–2141–825–921–58

Metal-sensitiveEphemeroptera taxa 5 (2.1) 5 3–8

Density (no./m2)

Total individualsEPT individuals

9356 (4294)5919 (2204)

83045790

3564–18,1332734–10,258

Metal-sensitiveEphemeroptera individuals 2045 (553) 2005 1135–2823

a Sum of Ephemeroptera, Plecoptera, and Trichoptera taxa

sites in the study area were typically belowAWQC. Elevated concentrations of Pb and Zn insediment at reference sites 1, 10, 11, and 12 maybe the result of limited mining upstream, nat-ural mineral deposits, and (or) prevailing windsthat historically transported airborne metal par-ticulates from the nearby smelter operations atthe Bunker Hill Superfund site near Kellogg,Idaho.

At those sites where benthic invertebrateswere used for tissue analysis, metal concentra-tions in whole-body and cytosol samples gen-erally corresponded with contamination levelsmeasured in sediment and water, confirmingthat metals originally associated with minewaste were biologically available. The correla-tion between metal concentrations in whole-body and cytosol samples suggests that, in gen-eral, site-specific exposures to bioavailable met-als in the CDA River basin can be inferred fromthe concentrations in whole-body samples.

Previous studies showed that metal contami-nation of streambed sediment and biota extends.70 km from major mining sources in the SouthFork CDA River basin to the Spokane Riverdownstream from Lake CDA (USGS 1999, Gros-bois et al. 2001). Mean concentrations of Cd, Pb,and Zn in tissue of caddisflies ;40 km down-stream from the lake were 3 mg/g, 52 mg/g, and

480 mg/g, respectively (USGS 1999). These tis-sue concentrations are ;3 to 18 times higherthan mean concentrations in caddisflies at ref-erence sites sampled in our study (Table 3).

Maret and Skinner (2000) noted that Cd, Pb,and Zn concentrations in livers and fillets of fishfrom the lower South Fork CDA River wereamong the highest from any sites sampled inthe Spokane River basin. Farag et al. (1998) alsonoted that concentrations of metals in tissue ofvarious fish and macroinvertebrates from theSouth Fork CDA River were higher than in mac-roinvertebrates at reference sites, and that metalconcentrations measured in Arctopsyche spp.were comparable to those at sites 14 and 17 onthe lower South Fork CDA River.

Effect of metals on invertebrate assemblages anddietary exposure to fish

Elevated metal exposure associated with min-ing density appeared to be directly related toeffects on benthic invertebrate assemblages.Comparisons of reference and test sites, usingboth upstream-versus-downstream and paired-basin approaches, provided evidence that met-als associated with mining have adversely af-fected benthic invertebrate assemblages. Num-bers of total taxa and EPT taxa at sites down-


TABLE 5. Extended.

Site type

Test(n 5 9)


Mann–Whitney

t-test p-value

25 (6.3)13 (3.3)73 (24)

5 (2.8)48 (26)

221484

648

19–388–18

28–931–9

15–80

0.0190.0140.1450.2610.566

4 (2.6) 4 0–8 0.326

5323 (3160)3746 (2609)

43303224

1333–10,521711–9097

0.0480.047

1054 (814) 968 0–2199 0.015

FIG. 4. Metal concentrations in whole-body caddisf-ly tissue and cytosol (Arctopsyche grandis and Hydrop-syche spp.), Coeur d’Alene and St. Regis River basins.A.—Cd. B.—Pb. C.—Zn. Means were used to representsites where both taxa were collected. * 5 reference site.

stream from areas of mining activity weresignificantly lower than at reference sites. Vari-ous measures of abundance, including densityof all individuals, density of EPT individuals,and density of metal-sensitive mayflies werealso reduced. Other studies have documentedreductions in total taxa with increasing metalcontamination (Beltman et al. 1999, Carlisle andClements 1999). Contrary to our results, Belt-man et al. (1999) did not identify a decrease intotal density of individuals, which Beltman etal. (1999) attributed to the highly variable num-bers of chironomids in their samples.

Metal-sensitive Ephemeroptera metrics wereless responsive to metal exposure. Reductions inthe number and % of metal-sensitive Ephemer-optera were evident only at the most contami-nated sites (13, 14, and 17) downstream fromareas of intensive mining activity. Kiffney andClements (1993) suggested that the loss of met-al-sensitive Ephemeroptera taxa from metal-contaminated systems is possibly a direct resultof high body burdens of metals. Our results areconsistent with this notion. The exceptionallyhigh metal exposures at sites 14 and 17, indi-cated by elevated concentrations in tissue of hy-dropsychid caddisflies, could prohibit coloniza-tion of these sites by metal-sensitive taxa. How-ever, contaminated sites may be occupied tem-porarily by insects drifting downstream from


FIG. 5. Comparisons of taxa number (A) and % composition (B) of EPT and metal-sensitive Ephemeropteraamong 4 reference and 2 test sites along the South Fork Coeur d’Alene River, Idaho, and St. Regis River, Montana.See Fig. 1B for site locations. * 5 reference site.

uncontaminated areas (Clements 1994, Beltmanet al. 1999). Even though these organisms mayoccupy contaminated sites for only a short time,the immigration of individuals from upstreammay affect sampling results. Our measures ofmetal-sensitive Ephemeroptera could have beenaffected by the presence or absence of a few raretaxa. Thus, measures of invertebrate drift maybe useful where drift may affect assessment re-sults. Also, samples collected in different sea-sons may help determine local survival of rarespecies (i.e., metal-sensitive Ephemeropterataxa) over the year.

Percent Ephemeroptera was not an effectivemetric, primarily because B. tricaudatus wasfound at all sites and was often the most abun-

dant taxon at test sites. This finding is consistentwith Clements (1994), who noted the occurrenceof large numbers of Baetis spp. during the sum-mer downstream from areas of intensive miningactivity along the Arkansas River in Colorado.Clements and Kiffney (1995) noted that the use-fulness of Baetis spp. as an indicator of metalcontamination in Rocky Mountain streams isquestionable because of its potential for rapidrecolonization. However, Deacon et al. (2001)noted that the abundances of Baetis spp. werereduced in streams in the Colorado River basinthat contained elevated trace metals.

Functional-feeding groups did not appear todiscriminate between reference and test sites.Carlisle and Clements (1999) also determined


FIG. 6. Correlations between benthic invertebrate metrics and cumulative toxic units (CTU) for Cd, Pb, andZn in water for all sampling sites, Coeur d’Alene and St. Regis River basins. Trend lines are based on the LOWESS(Locally Weighted Scatterplot Smoothing) technique (Helsel and Hirsch 1992).


that spatial assemblage patterns of functional-feeding groups were highly variable and not in-dicative of metal contamination.

Total and EPT taxa metrics at reference site 2were much lower than at other reference sites;however, density metrics had some of the high-est values (Fig. 6). This difference could be part-ly caused by the high density of Simulium spp.(9500 individuals/m2) collected at this site. Rea-sons for this unusually high number of blackflies at this location are unclear. This taxon oc-casionally has been found in high densities(.3000 individuals /m2) in riffle habitats of Ida-ho reference streams (Maret et al. 2001).

Species groups and community metrics ex-hibit a variety of response patterns to increasingmetal toxicity (Hickey and Golding 2002). LOW-ESS smoothing of the responses of taxa richnessand density to increasing CTUs for water andstreambed sediment suggested a threshold re-sponse for most metrics. We did not calculateeffects threshold levels, but CTUs for water andsediment at 7 of 9 test sites exceeded 1.0, indi-cating that cumulative metal concentrations maybe toxic to aquatic life, especially benthic inver-tebrates living on or in sediment. Invertebrateassemblages at moderately contaminated testsites 8 and 9 were not significantly differentfrom those at reference sites. Even though sed-iment at these 2 test sites was enriched in Pband Zn, concentrations in water did not exceedAWQC, and bioaccumulation of metals in cad-disfly tissue was relatively low. Maret andMacCoy (2002) also reported that fish assem-blages at these sites were similar to those at ref-erence sites. Even if metals are affecting bioticassemblages at these sites, detection of these ef-fects is probably confounded by the inherentvariability in these assemblages. Likewise, dem-onstrating biological recovery associated withremediation efforts at these sites will be diffi-cult.

In spite of elevated concentrations of metalsin water and streambed sediment, some benthicinvertebrate taxa still persisted at the most con-taminated sites. For example, 19 benthic inver-tebrate taxa were collected at site 13, and 29 in-vertebrate taxa were collected at sites 14 and 17.These numbers are higher than reported in ear-lier studies (Savage and Rabe 1973, Hoiland andRabe 1992, Hoiland et al. 1994), indicating thatconditions at these sites have improved overtime. Hoiland et al. (1994) attributed increases

in taxa richness at sites 14 and 17 to mine clo-sures and improved wastewater treatment.However, a recent study showed that fish do notinhabit site 13 (Maret and MacCoy 2002).

Abundances of benthic invertebrates werelower at test sites than at reference sites, butabundances were generally comparable to thosein other Western trout streams. The lowest den-sities occurred at test sites site 13 and 14 (,1600organisms/m2). Binns and Eiserman (1979) de-scribed streams with .5000 organisms/m2 asproductive trout streams. McGuire (2001) con-sidered densities of .5500 organisms/m2 astypical of Montana streams containing high-quality fish habitat. However, densities alonemay not be indicative of high-quality fish food,especially if the assemblage is made up of onlya few abundant taxa such as chironomids. Sur-prisingly, site 17 contained a much higher den-sity than expected (.10,000 organisms/m2).This large increase in density relative to othertest sites may be the result of organic materialfrom a sewage treatment outfall upstream (G.Harvey, Idaho Department of EnvironmentalQuality, personal communication). This samplewas predominantly composed of Hydropsychespp., which are tolerant of metals (Cain et al.1992, McGuire 2001) and organic pollution(Winner et al. 1980). Dense, filamentous algalgrowth was observed at this site in riffle habitatswhere benthic invertebrate samples were col-lected. Moderate nutrient increases from organicinflows increase benthic invertebrate abundanc-es (Quinn and Hickey 1993). In addition, dis-solved organic matter from this source may de-crease metal bioavailability and, subsequently,decrease toxicity to invertebrates (McCarthy1989).

Temporal variability in metal toxicity can beexpected as a result of seasonal influences(Nimmo et al. 1998). Clements (1994) noted thatthe greatest effects on invertebrates in the Ar-kansas River occurred during spring, whenmetal concentrations were highest. In contrast,dissolved Cd and Zn concentrations in streamsof the CDA River basin would be highest duringlow-flow periods (summer, autumn, and winter)(Woods 2000). Additional studies of temporalvariability in benthic invertebrate assemblagesin the CDA River basin are needed to better as-sess mining impacts throughout the year.

Feeding on contaminated invertebrates expo-ses fish to high concentrations of metals that are


potentially harmful. Farag et al. (1999) attribut-ed impaired health of fish in the South ForkCDA River to elevated metal concentrations inthe tissue of macroinvertebrate prey. Woodwardet al. (1994) determined that metals in benthicinvertebrates reduced survival and growth inhatchery rainbow trout at concentrations thatwere lower than Cd, Pb, and Zn in caddsifliesfound in our study at sites 14 and 17. Moreover,metal concentrations of the large-sized filterfeeders in our study may underestimate actualdietary exposure of fish because these concen-trations can be lower than in smaller-sized taxasuch as Baetis (Kiffney and Clements 1993, Faraget al. 1998). Thus, our results support the notionthat dietary exposures to metals pose the threatof reduced growth and survival to resident fish,especially early life stages of sculpin and troutthat feed almost exclusively on small-sized in-vertebrates. This notion is consistent with Maretand MacCoy (2002) who reported that sculpinswere absent at most test sites (except 8 and 9).The most contaminated test sites, 13, 14, and 17,also had lower trout biomass than most refer-ence sites.

Our study used a spatially extensive design,and included paired-basin comparisons andsampling along metal-contamination gradientsto assess the response of benthic invertebratesto mining impacts. We conclude that metal ex-posure affected benthic invertebrate assemblag-es in the CDA River basin. Several measures ofassemblage structure, including total number oftaxa, number of EPT taxa, and the density ofmetal-sensitive mayflies, were inversely relatedto indicators of metal exposure and toxicity.Production mine density within a 500-m streambuffer was a significant explanatory variable ofmining impacts in our study. Maret and Skinner(2000) noted significant correlations betweenproduction mine density and metal concentra-tions in sediment, but mine density previouslyhas not been used to describe basin disturbanceor metal concentrations that may be detrimentalto benthic invertebrates. This basin descriptorcan be generated from GIS, and we suggest thatit may have general utility as an indicator of theseverity of metal contamination and effects onaquatic life in Rocky Mountain streams.

Acknowledgements

We thank the following individuals from theUSGS for assistance in collecting and processing

data during the course of this study: Rick L.Backsen, Mike A. Beckwith, Craig L. Bowers,Lawrence R. Deweese, Angela K. Frandsen,Keith L. Hein, Douglas S. Ott, Kenneth D. Skin-ner, and Zachary L. Taylor. Steve E. Box provid-ed useful information on mine locations and ge-ology. Jack N. Goldstein assisted with fish col-lections. Michelle I. Hornberger did the metalanalyses on streambed sediment. Reviews byGregory M. Clark, Aida M. Farag, and David A.Peterson improved the quality of earlier ver-sions of the manuscript. We also appreciate thecomments and advice of William H. Clementsand 2 anonymous reviewers. Funding was pro-vided by the USGS NAWQA Program.

Literature Cited

ARBOGAST, B. F. 1990. Quality assurance manual forthe Branch of Geochemistry. U. S. Geological Sur-vey Open-File Report 90-668. US Geological Sur-vey, Denver, Colorado.

BECKWITH, M. A. 1998. Concepts for monitoring waterquality in the Spokane River basin, north Idahoand eastern Washington. U. S. Geological SurveyOpen-File Report 98-534. US Geological Survey,Boise, Idaho.

BELTMAN, D. J., W. H. CLEMENTS, J. LIPTON, AND D.CACELA. 1999. Benthic invertebrate metals expo-sure, accumulation, and community-level effectsdownstream from a hard-rock mine site. Environ-mental Toxicology and Chemistry 18:299–307.

BINNS, N. A., AND F. M. EISERMAN. 1979. Quantifica-tion of fluvial trout habitat in Wyoming. Trans-actions of the American Fisheries Society 108:215–228.

BRENNAN, T. S., A. M. CAMPBELL, A. K. LEHMANN,AND I. O’DELL. 1999. Water resources data, Idaho,water year 1998, volume 2, Upper Columbia Riverbasin and Snake River basin below King Hill. U.S. Geological Survey Water-Data Report ID-99-2.US Geological Survey, Boise, Idaho.

BUCHMAN, M. F. 1999. NOAA screening quick refer-ence tables: NOAA HAZMAT Report 99-1. Coast-al Protection and Restoration Division, NationalOceanic and Atmospheric Administration. Seattle,Washington.

CAIN, D. J., AND S. N. LUOMA. 1998. Metal exposureto native populations of the caddisfly Hydropsyche(Trichoptera: Hydropsychidae) determined fromcytosolic and whole body metal concentrations.Hydrobiologia 386:103–117.

CAIN, D. J., S. N. LUOMA, J. L. CARTER, AND S. V. FEND.1992. Aquatic insects as bioindicators of trace el-ement contamination in cobble-bottom rivers and


streams. Canadian Journal of Fisheries andAquatic Sciences 49:2141–2154.

CARLISLE, D. M., AND W. H. CLEMENTS. 1999. Sensitiv-ity and variability of metrics used in biologicalassessments of running waters. EnvironmentalToxicology and Chemistry 18:285–291.

CLEMENTS, W. H. 1994. Benthic invertebrate commu-nity response to heavy metals in the Upper Ar-kansas River basin, Colorado. Journal of the NorthAmerican Benthological Society 13:30–44.

CLEMENTS, W. H., D. M. CARLISLE, J. M. LAZORCHAK,AND P. C. JOHNSON. 2000. Heavy metals structurebenthic communities in Colorado mountainstreams. Ecological Applications 10:626–638.

CLEMENTS, W. H., AND P. M. KIFFNEY. 1995. The influ-ence of elevation on benthic community respons-es to heavy metals in Rocky Mountain streams.Canadian Journal of Fisheries and Aquatic Scienc-es 52:1966–1977.

CUFFNEY, T. F., M. E. GURTZ, AND M. R. MEADOR. 1993.Methods for collecting benthic invertebrate sam-ples as part of the National Water-Quality As-sessment Program. U.S. Geological Survey Open-File Report 93-406. US Geological Survey, Denver,Colorado.

DEACON, J. R., N. E. SPAHR, S. V. MIZE, AND R. W.BOULGER. 2001. Using water, bryophytes, andmacroinvertebrates to assess trace element con-centrations in the Upper Colorado River basin.Hydrobiologia 455:29–39.

DOPPELT, R., M. SCURLOCK, C. A. FRISSELL, AND J. R.KARR. 1993. Entering the watershed: a new ap-proach to save America’s river ecosystem. IslandPress, Washington, DC.

DYER, S. D., C. E. WHITE-HULL, AND B. K. SHEPARD.2000. Assessments of chemical mixtures via tox-icity reference values overpredict hazard to Ohiofish communities. Environmental Science andTechnology 34:2518–2524.

ELLIS, M. M. 1940. Pollution of the Coeur d’Alene Riv-er and adjacent waters by mine wastes. U.S. Bu-reau of Fisheries Special Report Number 1. USBureau of Fisheries, Washington, DC.

FARAG, A. M., D. F. WOODWARD, W. BRUMBAUGH, J. N.GOLDSTEIN, E. MACCONNELL, C. HOGSTRAND,AND F. T. BARROWS. 1999. Dietary effects of met-als-contaminated invertebrates from the Coeurd’Alene River, Idaho, on cutthroat trout. Trans-actions of the American Fisheries Society 128:578–592.

FARAG, A. M., D. F. WOODWARD, J. N. GOLDSTEIN, W.BRUMBAUGH, AND J. S. MEYER. 1998. Concentra-tions of metals associated with mining waste insediments, biofilm, benthic macroinvertebrates,and fish from the Coeur d’Alene River basin, Ida-ho. Archives of Environmental Contaminationand Toxicology 34:119–127.

FISHMAN, M. J. (EDITOR). 1993. Methods of analysis by

the US Geological Survey National Water QualityLaboratory—determination of inorganic and or-ganic constituents in water and fluvial sediments.U. S. Geological Survey Open-File Report 93-125.US Geological Survey, Denver, Colorado.

FITZPATRICK, F. A., I. R. WAITE, P. J. D’ARCONTE, M. R.MEADOR, M. A. MAUPIN, AND M. E. GURTZ. 1998.Revised methods for characterizing stream habi-tat in the National Water-Quality Assessment Pro-gram. U. S. Geological Survey Water-ResourcesInvestigations Report 98–4052. US GeologicalSurvey, Middleton, Wisconsin.

GRAFE, C. S., C. A., MEBANE, M. J. MCINTYRE, D. A.ESSIG, D. H. BRANDT, AND D. T. MOSIER. 2002. TheIdaho Department of Environmental Quality wa-ter body assessment guidance, second edition—final. Idaho Department of Environmental Qual-ity, Boise, Idaho.

GROSBOIS, C. A., A. J. HOROWITZ, J. J. SMITH, AND K.A. ELRICK. 2001. The effects of mining and relatedactivities on the sediment-trace element geochem-istry of Lake Coeur d’Alene, Idaho, USA. PartIII—downstream effects: the Spokane River basin.Hydrological Processes 15:855–875.

HARE, L. 1992. Aquatic insects and trace metals: bio-availability, bioaccumulation, and toxicity. CriticalReviews in Toxicology 22:327–369.

HELSEL, D. R., AND R. M. HIRSCH. 1992. Statisticalmethods in water resources. Elsevier, New York.

HICKEY, C. W., AND L. A. GOLDING. 2002. Response ofmacroinvertebrates to copper and zinc in astream mesocosm. Environmental Toxicology andChemistry 21:1854–1863.

HOILAND, W. K., AND F. W. RABE. 1992. Effects of in-creasing zinc levels and habitat degradation onmacroinvertebrate communities in three northIdaho streams. Journal of Freshwater Ecology 7:373–380.

HOILAND, W. K., F. W. RABE, AND R. C. BIGGAM. 1994.Recovery of macroinvertebrate communities frommetal pollution in the South Fork and mainstemof the Coeur d’Alene River, Idaho. Water Environ-ment Research 66:84–88.

HORNBERGER, M. I., S. N. LUOMA, A. VAN GEEN, C.FULLER, AND R. ANIMA. 1999. Historical trends ofmetals in the sediments of San Francisco Bay, Cal-ifornia. Marine Chemistry 64:39–55.

HOROWITZ, A. J., K. A. ELRICK, J. A. ROBBINS, AND R.B. COOK. 1995. Effect of mining and related activ-ities on the sediment trace element geochemistryof Lake Coeur d’Alene, Idaho, USA. Part II—subsurface sediments. Hydrological Processes9:35–54.

HUGHES, R. M., D. P. LARSEN, AND J. M. OMERNIK.1986. Regional reference sites: a method for as-sessing stream potentials. Environmental Man-agement 10:629–635.

KENDY, E., AND R. E. TRESCH. 1996. Geographic, geo-


logic, and hydrologic summaries of intermontanebasins of the Northern Rocky Mountains, Mon-tana. U. S. Geological Survey Water-Resources In-vestigations Report 96-4025. US Geological Sur-vey, Helena, Montana.

KIFFNEY, P. M., AND W. H. CLEMENTS. 1993. Bioaccu-mulation of heavy metals by benthic invertebratesat the Arkansas River, Colorado. EnvironmentalToxicology and Chemistry 12:1507–1517.

KIFFNEY, P. M., AND W. H. CLEMENTS. 1994. Structuralresponse of benthic macroinvertebrate communi-ties from different stream orders to zinc. Environ-mental Toxicology and Chemistry 13:389–395.

LUOMA, S. N. 1989. Can we determine the biologicalavailability of sediment-bound trace elements?Hydrobiologia 176/177:379–396.

LUOMA, S. N., Y. B. YO, AND G. BRYAN. 1995. Fate,bioavailability and toxicity of silver in estuarineenvironments. Marine Pollution Bulletin 31:44–54.

MARET, T. R., AND D. M. DUTTON. 1999. Summary ofinformation on synthetic organic compounds andtrace elements in tissue of aquatic biota, ClarkFork-Pend Oreille and Spokane River basins,Montana, Idaho, and Washington, 1974–96. U. S.Geological Survey Water-Resources InvestigationsReport 98-4254. US Geological Survey, Boise, Ida-ho.

MARET, T. R., AND D. E. MACCOY. 2002. Fish assem-blages and environmental variables associatedwith hard-rock mining in the Coeur d’Alene Riv-er basin, Idaho. Transactions of the AmericanFisheries Society 131:865–884.

MARET, T. R., D. E. MACCOY, K. D. SKINNER, S. E.MOORE, AND I. O’DELL. 2001. Evaluation of mac-roinvertebrate assemblages in Idaho rivers usingmultimetric and multivariate techniques, 1996–98.U. S. Geological Survey Water-Resources Investi-gations Report 01-4145. US Geological Survey,Boise, Idaho.

MARET, T. R., AND K. D. SKINNER. 2000. Concentrationsof selected trace elements in fish tissue andstreambed sediment in the Clark Fork-PendOreille and Spokane River basins, Washington,Idaho, and Montana, 1998. U. S. Geological Sur-vey Water-Resources Investigations Report 00–4159. US Geological Survey, Boise, Idaho.

MCCARTHY, J. F. 1989. Bioavailability and toxicity ofmetals and hydrophobic organic contaminants.Pages 263–277 in I. H. Suffet and P. McCarthy(editors). Aquatic humic substances: influence onfate and treatment of pollutants. Elsevier andAmerican Chemical Society, Washington, DC.

MCGUIRE, D. L. 2001. Clark Fork River macroinverte-brate community biointegrity: 2000 assessments.Prepared for Montana Department of Environ-mental Quality, Planning, Prevention, and Assis-tance Division, by McGuire Consulting, Espano-la, New Mexico. (Available from: Planning, Pre-

vention, and Assistance Division, Montana De-partment of Environmental Quality, 1520 E 6thAvenue, Helena, Montana 59629 USA.)

MEBANE, C. A. 2001. Testing bioassessment metrics:macroinvertebrate, sculpin, and salmonid re-sponses to stream habitat, sediment, and metals.Environmental Monitoring and Assessment 67:293–322.

MERRITT, R. W., AND K. W. CUMMINS (EDITORS). 1996.An introduction to the aquatic insects of NorthAmerica. 3rd edition. Kendall/Hunt, Dubuque,Iowa.

MOULTON, S. R., J. L. CARTER, S. A. GROTHEER, T. F.CUFFNEY, AND T. M. SHORT. 2000. Methods ofanalysis by the U.S. Geological Survey NationalWater Quality Laboratory—processing, taxono-my, and quality control of benthic macroinverte-brate samples. U. S. Geological Survey Open-FileReport 00-212. US Geological Survey, Denver,Colorado.

NIMMO, D. R., M. J. WILLOX, T. D. LAFRANCOIS, P. L.CHAPMAN, S. F. BRINKMAN, AND J. C. GREENE.1998. Effects of metal mining and milling onboundary waters of Yellowstone National Park,USA Environmental Management 22:913–926.

OMERNIK, J. M., AND A. L. GALLANT. 1986. Ecoregionsof the Pacific Northwest. EPA 600/3–86/033. USEnvironmental Protection Agency, Corvallis,Oregon.

POULTON, B. C., D. P. MONDA, D. F. WOODWARD, M.L. WILDHABER, AND W. G. BRUMBAUGH. 1995. Re-lations between benthic community structure andmetals concentrations in aquatic macroinverte-brates: Clark Fork River, Montana. Journal ofFreshwater Ecology 10:277–293.

PROTHRO, M. G. 1993. Office of water policy and tech-nical guidance on interpretation and implemen-tation of aquatic life metals criteria. Memoran-dum from Acting Assistant Administrator forWater. Office of Water, U.S. Environmental Protec-tion Agency, Washington, DC. (Available from:US Geological Survey Library, 345 MiddlefieldRoad, MS-955, Menlo Park, California 94025 USA,or US Environmental Protection Agency Library,Region 10, 1200 Sixth Avenue, Seattle, Washing-ton 98101 USA.)

QUINN, J. M., AND C. W. HICKEY. 1993. Effects of sew-age waste stabilization lagoon effluent on streaminvertebrates. Journal of Aquatic EcosystemHealth 2:205–219.

RAINBOW, P. S. 1996. Heavy metals in aquatic inver-tebrates. Pages 405–426 in W. N. Beyer, G. H.Heinz, and A. W. Redmon-Norwood (editors).Environmental contaminants in wildlife. CRCPress, Boca Raton, Florida.

REISER, D. W. 1999. United States of America vs.ASARCO Incorporated, et al. Case No. 96-0122-N-EJL in the United States District Court for the


District of Idaho. Expert Report of Dudley W. Rei-ser, Ph.D., Redmond, Washington.

SAVAGE, N. L., AND F. W. RABE. 1973. The effects ofmine and domestic wastes on macroinvertebratecommunity structure in the Coeur d’Alene River.Northwest Science 47:159–168.

SHELTON, L. R., AND P. D. CAPEL. 1994. Guidelines forcollecting and processing samples of streambedsediment for analysis of trace elements and or-ganic contaminants for the National Water-Qual-ity Assessment Program. U. S. Geological SurveyOpen-File Report 94-458. US Geological Survey,Denver, Colorado.

SORENSEN, E. M. 1991. Metal poisoning in fish. CRCPress, Boca Raton, Florida.

STRAHLER, A. N. 1957. Quantitative analysis of water-shed geomorphology. Transactions of the Amer-ican Geophysical Union 38:913–920.

US BUREAU OF MINES. 1995. Minerals availability sys-tem. U. S. Geological Survey Special Publication12-95. US Geological Survey, Denver, Colorado.(Data set on CD-ROM.)

USEPA (US ENVIRONMENTAL PROTECTION AGENCY).2000. Water quality standards, federally promul-gated water quality standards, 40 CFR131.36. USGovernment Printing Office, Washington, DC.(Available from: http://www.access.gpo.gov/nara/)

USGS (US GEOLOGICAL SURVEY). 1975. Hydrologicunit map—1974, State of Idaho, scale 1:500,000.US Geological Survey, Reston, Virginia.

USGS (US GEOLOGICAL SURVEY). 1999. National Wa-ter-Quality Assessment—Northern Rockies StudyUnit (NROK), Spokane River Studies. US Geolog-ical Survey, Boise, Idaho. (Available from: http://idaho.usgs.gov/projects/spokane/.)

WALLACE, W. G., G. R. LOPEZ, AND J. S. LEVINTON. 1997.Cadmium resistance in an oligochaete and its ef-fect on cadmium trophic transfer to an omnivorousshrimp. Marine Ecology Progress Series 172:225–237.

WILDE, F. D., AND D. B. RADTKE (EDITORS). 1998. Na-tional field manual for the collection of water-qual-ity data. U. S. Geological Survey techniques of wa-ter-resources investigations. Book 9. Handbooksfor water-resources investigations. Chapter A6,Field measurements. US Government Printing Of-fice, Washington, DC.

WINNER, R. W., M. W., BOESEL, AND M. P. FARRELL.1980. Insect community structure as an index ofheavy-metal pollution in lotic ecosytems. CanadianJournal of Fisheries and Aquatic Sciences 37:647–655.

WOODS, P. F. 2000. Concentrations and loads of cad-mium, lead, zinc, and nutrients measured duringthe 1999 water year within the Spokane River ba-sin, Idaho and Washington. U. S. Geological SurveyOpen-File Report 00-441. US Geological Survey,Boise, Idaho.

WOODS, P. F., AND M. A. BECKWITH. 1997. Nutrient andtrace-element enrichment of Coeur d’Alene Lake,Idaho. U. S. Geological Survey Water-Supply Paper2485. US Geological Survey, Boise, Idaho.

WOODWARD, D. F., W. G. BRUMHAUGH, A. J. DELONAY,E. E. LITTLE, AND C. E. SMITH. 1994. Effects on rain-bow trout fry of a metals-contaminated diet of ben-thic invertebrates from the Clark Fork River, Mon-tana. Transactions of the American Fisheries Soci-ety 123:51–62.

Received: 6 November 2002Accepted: 7 July 2003

response of benthic invertebrate assemblages to metal ... · 1994). some insect taxa accumulate...

Documents