Page 809 - The Toxicology of Fishes

Mining Impacts on Fish in the Clark Fork River, Montana: A Field Ecotoxicology Case Study 789

10000

1000
Copper (ppm) 100 CF

FC
GC
LB
10 RC
SB
MT
1
500 550 600 650 700 750 800 850
River Kilometers
FIGURE 19.5 Exponential decline in copper concentrations (µg/g dry wt) in the <63-µm fraction of sediments, as a
function of river mile, from Silverbow Creek (SB), tributaries of the Clark Fork (RC, FC, GC, LB), the Clark Fork (CF),
and Milltown Reservoir (MT) (Moore et al., unpublished data). Note the persistent high concentrations in Silverbow Creek,
the low copper concentrations in tributaries, and the exponential decline of contamination along the Clark Fork River.
Despite dilution from major tributaries, copper concentrations in the Clark Fork sediments are ≥10× higher than concen-
trations in tributary sediments over the entire 250 km of river.

concentration of metals in the stream sediment at every tributary juncture. A simplified estimate of the
amount of sediment supplied from a particular drainage can be obtained from the cumulative area of
the basin (Helgen and Moore, 1996), as long as the climate and geology are similar.
On a scale of 600 km, bed sediment metal concentrations in Clark Fork River decline downstream
from below the pond anomaly, following a single logarithmic function (Figure 19.5) (Axtmann and
Luoma, 1991; Helgen and Moore, 1996). Copper concentrations in the uppermost Clark Fork River are
>100 times higher than in tributaries without a major history of mining. Concentrations of cadmium are
67 times higher. At a site 368 km from the mine, copper concentrations were found to be 10 to 20 times
higher than concentrations in tributaries. A strong fit to the cumulative basin area model (Helgen and
Moore, 1996) indicates that tributary sediments progressively dilute the upstream source. The banks in
the first 50 km of the river appear to be the likely source (Hornberger et al., 1995). The function describing
the decline predicts that the distance necessary to dilute metal concentrations by half was 100 to 180
km; contamination from the mine could extend 475 to 750 km in an unimpeded river system.
Although downstream trends were quite distinct in the 600-km spatial scale, distribution of metal
contamination was more complex on smaller scales. Axtmann et al. (1997) noted that metal concen-
trations in the middle reaches of the Upper Clark Fork River slightly exceeded those predicted by
watershed dilution of metal contamination. They attributed this to inputs by contaminated local banks.
Hydrologic and geomorphologic processes, such as floods, new bank cuts, and variable mobilization
of fine-grained deposits within the river bed, also will change local contributions to the gradient, in
unpredictable ways. Ice jams in some years and not others can redistribute contamination spatially and
affect year-to-year comparability (Moore and Landrigan, 1999). Incomplete mixing of sediments near
tributary confluences can reduce metal concentrations in the immediate vicinity, only to have them
return to higher values further downstream (this is a common feature around tributaries in the upper
Clark Fork River) (Axtmann et al., 1997). Monthly metal (copper) concentrations in fine-grained
sediments show that as much as twofold variability in sediment-metal concentration is common within
a site, following seasonal patterns in some cases (Figure 19.6). With less frequent data, it is difficult
to differentiate metal concentrations between two stations as far apart as 50 km in the upper Clark
Fork (Axtmann and Luoma, 1991).