Environmental Engineering Reference
In-Depth Information
The no-effects data sets and the effects data sets for marine and fresh water
sediments are listed in Table
6
above.
Inspection of the four data sets raises the important question of why there are no
effects at thousands of ppb and effects at a few or even less than 1 ppb. As will
become apparent, these discrepancies are not so much the result of species and
ecosystem differences, but rather, errors and misinterpretations in the data sets.
Using the 50th percentile of the no-effects data set ignores the most important
information. The highest half of the data points that were not associated with a toxic
effect are counted only by number and not by concentration. For example, the 50th
percentile of the no-effects data set for total DDT in marine sediments is 5 ppb (aver-
age of two shaded values in Table
6
). However, in the same data set, levels from 5 to
35,300 ppb are also without effect, but do not weigh by concentration of total DDT
into the TEL value. Based on this data set, the threshold for toxicity is likely to be
three orders of magnitude higher than the 50th percentile value. The threshold for
toxicity is not appropriately weighed by the TEL calculation. This point has also
been made by MacDonald, the author of the TELs. When MacDonald assessed total
DDT toxicity in marine sediments from Southern California (MacDonald
1994
),
he reported a lowest observed effect level for total DDT for the most sensitive life
stage of the most sensitive organism to be 7,120 ppb.
The 15th percentile of the effects data set for marine sediments is 2.92 ppb total
DDT (shaded data point in Table
6
above). This means that only six out of 37 data
points infl uence the 15th percentile by concentration. Twenty two data points matter
only by count. Hence, whatever dose-response data are in the data set, only the 15th
percentile value weighs into the TEL. Review of the underlying studies for each data
point in the effects data set for marine sediments revealed highly signifi cant errors
and inconsistencies. The analysis is summarized below.
The data point of 1.6 ppb (Lyman et al.
1987
), one of the 6 data points that count
in the 15th percentile, is a re-publication of the 1.58 ppb data point (JRB Associates
1984
) shown in Table
6
. Hence, a single data point, based on equilibrium partitioning,
is given the weight of two data points (Table
6
).
The lowest fi ve data points all rely on outdated and inaccurate K
ow
values. K
ow
values were used to estimate K
oc
values that were then employed to determine equi-
librium partitioning between water and organic carbon in sediment. The old meth-
odology of determining K
ow
has been superseded by the slow-stir method (de Bruijn
et al.
1989
). There is general consensus among scientists measuring K
ow
that the
slow stir method gives the most accurate value, and deserves to supersede earlier
methodologies. K
ow
values based on the slow stir method are recommended by
U.S. EPA (
2002
) and are listed for the DDTs in the SARWQCB staff TMDL report
(SARWQB
2006
). The fi ve data points relying on outdated K
ow
values display errors
of more than one order of magnitude.
Dealing with mixtures of chemicals becomes very important when evaluating
many of the TEL data points in the total DDT effects data sets, particularly the fi ve
data points ranging from 2.92 to 3.27 ppb in marine sediments from highly polluted
areas in San Francisco Bay (Table
6
). Based on spiked-sediment bioassays and equi-
librium partition calculations, these levels are three orders of magnitude below
Search WWH ::
Custom Search