 |
||
|
|
||
|
Page Title:
Correlative/Observational Studies (cont.) |
||
| |||||||||||||||
|
|
 ERDC TN-DOER-C23
September 2001
sediments from 25 locations in Baltimore Harbor. Sediments had effects range median quotients
Survival in acute toxicity tests with these sediments ranged from 0 to >129 percent with seven of
the sites showing statistically reduced survival relative to control. For six of the test samples where
survival was in excess of 100 percent the authors indicated that the excess was likely due to the
presence of indigenous amphipods in the samples. It is interesting to note that individual replicates
for the toxicity tests were actually derived from separate independent grabs at a given location. This
is a departure from the standard approach used in sediment toxicity tests where a single homogenized
sample is divided among the laboratory replicates for the test. Only one of these independent grab
samples was evaluated for chemistry.
A plot of the 10-day toxicity test results (including the retest data) against field densities of
amphipods at the test sites suggests a positive correlation (R2 = 0.41, p = 0.001) between amphipod
density and amphipod survival (Figure 5a). The correlation improved slightly (R2 = 0.52, p =
0.0002) with exclusion of a potential outlier value (Figure 5b). In addition, the authors showed a
ΣERM-Q)2 and amphipod survival (Figure 6a). Again the correlation improved slightly (R2 = 0.29,
p = 0.006) with exclusion of a potential outlier (Figure 6b). It is important to note that correlations
might have been stronger if spatial variability in the field had not been incorporated into the sampling
design via the use of independent grab samples for each laboratory toxicity replicate and had
chemical analysis been performed on a composite or all of the sediments rather than a single
replicate.
In a separate study, McGee and Fisher (1999) summarized results of chronic sublethal tests with
L. plumulosus that were conducted in parallel with acute tests on selected samples from many of
the same sites evaluated in their earlier study. Unfortunately, nearly all of the 11 sites included in
this study were depauperate of L. plumulosus. Consequently, a comparison between laboratory
response and field densities of L. plumulosus was not meaningful. A comparison of survival in
these laboratory tests to broader indices of benthic community health (e.g., number of taxa, total
number of organisms, Shannon-Weaver diversity index, and the index of biological integrity (IBI))
also showed little to no relationship (Figure 7). Similarly, a comparison of sublethal end points
(e.g., growth and reproduction) to these same indices also showed no relationship. It is not clear
why this was the case. Perhaps the contamination gradient was not broad enough, or factors other
than contaminants were affecting differences in the benthic communities. Alternatively, it could
be that the community of organisms, showing no gross changes in total abundance, species richness,
or other indices, was populated by organisms that were less sensitive or less exposed to the sediment
contaminants.
1
The effects range median quotient (ERM-Q) is derived by dividing the actual measured concentration in the
sediment by the corresponding effects range median (ERM) value for that particular contaminant. ERM values
were developed by Long et al. (1995) using the distribution of observed mortality in laboratory bioassays for a
large number of sediments collected from a variety of areas. The ERM value represents the concentration of a
particular contaminant that causes significant mortality in the test species 50 percent of the time for the range
of sediments used in deriving the value.
2
ΣERM-Q represents the sum of all ERM-Q values for a particular sediment.
9
|
|
Privacy Statement - Press Release - Copyright Information. - Contact Us - Support Integrated Publishing |
