Page 808 - The Toxicology of Fishes

788 The Toxicology of Fishes

characterize the types of variability typical of mine-impacted streams; for example, the hyporheic zones
under the bed of a contaminated stream or in the subsurface of a slicken deposit are microbially and
geochemically heterogeneous environments (Wielinga et al., 1999) that contain acidic, metal-rich pore
water (Benner et al., 1995). Snowmelt, early spring flushing events, or post-rain surges can transport
these metals to the stream (Nagorski et al., 2003). Such variability in metal concentration occurs on
weekly, monthly, or seasonal scales. Concentrations in a contaminated stream also fluctuate regularly
during the day (Nagorski et al., 2003) and can be two- to threefold higher at night than during the day
in the Clark Fork (Brick and Moore, 1996). The short-term variations may be as great as the longer term
variability, and both may have important implication for toxicity.
Dissolved copper concentrations in the quarterly data for the Clark Fork River ranged from 2 to 20,
with only 2 of 232 observations exceeding the acutely lethal concentration. Chronic dissolved exposures
to metals also did not suggest effects at the average concentrations (Woodward et al., 1994, 1995);
however, survival of trout fry and fingerlings placed in cages in the Clark Fork River for several months
was lower than for reference streams in the late 1980s (Phillips and Lipton, 1995). The USEPA concluded
that acute toxicity rarely if ever occurred in the river, but the Agency could not explain either the caged
fish mortality or the reduced fish populations of the river. Although large overland inputs of contaminated
water were not observed during this period, less visible acute pulses of metal input were suspected, from
groundwaters or the hyporheic zone (USEPA, 1999). The overall conclusion was that survival was
affected by “exposures to pulses, or other high concentration events,” although none was documented.
They also suggested, but could not quantify, a role for nonmetal stressors in affecting fish. Both
conclusions were debatable, given their weak direct documentation, but the conjecture about pulse inputs
was consistent with the complex behavior and challenges of sampling contamination in the water column
of a mine-impacted river.

Sediment Contamination
When contaminated particulate or sedimentary material is dispersed through an ecosystem, it equilibrates
with water, detritus, and living food materials, resulting in ongoing contamination of all environmental
compartments. Organisms (including fish) are exposed to this milieu throughout their lifetime. This type
of chronic sediment contamination may reduce or eliminate populations of fish without killing adults
(e.g., by inhibiting reproduction) or without leaving visually clear evidence of effects. Chronic sediment
contamination can also be widespread. The occurrence, distributions, and geochemistry of the dispersed
material, as well as ecological characteristics, determine biological exposures to contamination.
The heterogeneous mix of mine-derived materials in a river bed usually includes natural silt-clays,
sand, gravel, or cobble. When metal concentrations are compared among samples with such different
particle-size distributions, the results can be very difficult to interpret. The highest concentrations of
metals typically associate with fine-grained sediments. Sediments dominated by sand typically have
lower metal concentrations because of a smaller ratio of surface area to mass. Fine-grained sediments
are also remobilized easily and are most relevant to biological exposures. Separation of the fine fraction
of sediments for analysis reduces the physical variability among sediment samples, and that reduces
the impact of grain size variability on concentrations (Salomons and Forstner, 1984). Interpretation of
much of the early metal data from sediments of the Clark Fork River was highly confounded by grain-
size variability. Characterizations of metal contamination from concentrations in fine-grained, sieved
sediments (<64 µm) resolved that variability and diffused some of the contentious discussions over
contamination trends in the river (Axtmann and Luoma, 1991; Brook and Moore, 1988). In a mine-
impacted river, local precipitation events and subsurface inputs can confound the typical particle size
relationship (Moore et al., 1989), especially in the most contaminated areas. Sieved sediments are still
effective measures of concentration in such areas but should not be used to evaluate metal loads in the
system.
When particle size biases are eliminated, the primary determinants of downstream trends in sediment
contamination away from a mine site include the size of tributaries, their sediment load, and their
buffering. As the sediment moves downstream away from the mineralized zone, it mixes with sediment
from tributaries draining uncontaminated surrounding areas. The unenriched sediments dilute the