3-16-95
ANALYSES OF ACUTE AND CHRONIC DATA
FOR AQUATIC LIFE
by
George E. Host
Natural Resources Research Institute
Duluth, MN 55811
Ronald R. Regal
University of Minnesota
Duluth, MN 55812
and
Charles E. Stephan
Environmental Research Laboratory - Duluth
6201 Congdon Blvd.
Duluth, MN 55804
OFFICE OF ENVIRONMENTAL PROCESSES AND EFFECTS RESEARCH
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
-------
NOTICE
The information in this document has been funded wholly or
in part by the United States Environmental Protection Agency
under Cooperative Agreement No. CR-814147-02-O with the
University of Wisconsin - Superior and through Contract No. 68-
03-3544 with AScI Corporation. The extramural portion of this
work was funded by the Criteria and Standards Division of the
Office of Water and by Region III in connection with a RARE
(Regional Applied Research Effort) project concerning pesticides.
This document has been subjected to the Agency’s peer arid
administrative review, and it has NOT been approved for
publication as an EPA document. Mention of trade names or
commercial products does not constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
The U.S. EPA is, responsible for the protection of the
environment in which we live and work An essential part of
exercising this responsibility is the research that allows the
Agency to formulate and implement actions that lead to a
compatible balance between human activities and preservation of
the environment.
The Environmental Research Laboratory - Duluth is the
Agency’s center of expertise in freshwater aquatic toxicology.
To help the Agency fulfill its responsibilities as set forth in
the Clean Water Act, ERL-Duluth is examining ways that might be
used to draw valid and useful conclusions on the basis of a few
data concerning toxicity to aquatic organisms without requiring
the collection of many additional toxicity data in all
situations. The work reported here deals with analyses of acute
values and final acute-chronic ratios that were derived from
aquatic life criteria documents.
iii
-------
ACKNOWLEDGEMENT
The authors thank R.J. Erickson for his help in designing
these analyses and interpreting the results.
iv
-------
ABSTRACT
Data concerning acute toxicity and acute-chronic ratios were
obtained from draft and final aquatic life criteria documents so
that several kinds of analyses could be performed. An analysis
of the dependence of the Final Acute Value (FAV) on the number of
Genus Mean Acute Values (GMAVs) used in its calculation
demonstrated that, although for many datasets the FAV increased
as the number of GMAVs increased above the minimum of eight, the
FAV decreased for a few datasets. Final Acute Value Factors
(FAVF5), which are intended to relate the results of one or a few
acute toxicity tests to a FAV, were derived by empirical and
theoretical methods. The empirical derivation of FAVFs was
accomplished using specially selected subsets (i.e., samples) of
acute values obtained from datasets contained in the criteria
documents. FAVFs for subsets containing 1 to 8 acute values,
where each acute value satisfied a different minimum data
requirement 1 were determined by two methods. Whether calculated
on a worst-case basis or as percentiles from random
sampling, the FAVFS decreased substantially as subset size
increased from 1 to 8 for both fresh and salt water. Also for
both waters, when the subset was required to contain an acute
value for one or more species that are sensitive to many -
chemicals, the FAVF was much smaller, especially at small subset
sizes. The theoretical derivation of FAVFs was based on the log-
triangular distribution. Final Chronic Value Factors (FCVFS)
were calculated by applying the log-normal distribution to Final
Acute-Chronic Ratios given in criteria documents; the 95 th
percentile FCVF was 43 for fresh water and 39 for, salt water.
V
-------
CONTENTS
Page
References . 33
Notice
Foreword
Acknowledgement
Abstract
Figures
Tables
1. Introduction
2. Database Management
3. Dependence of the FAV on the Number of GMAVs
4. FAVFs Based on Highest Acute Values
5. FAVFs Based on Random Samples of Acute Values
6. FAVFs Based on the Log-triangular Distribution
7. Final Chronic Value Factors
8. Discussion
ii
iii
iv
V
vii
Viii
1
4
6
8
11
16
23
25
vi
-------
F IGURES
Number Paq e
1 Probability Plot for Chiordane in Salt Water 34
2 Log-normal Probability Plot of Freshwater FACRS . . . . 35
3 Log-normal Probability Plot of Saltwater FACRs 36
vii
-------
TABLES
Number Page
1 Sources of Datasets 37
2 Fields Used in the Database COMBO 39
3 Datasets Not Satisfying the Minimum Data Requirements . 41
4 Values of FAV and S for Datasets in Source Documents . . 42
5 Dependence of the FAV and S on the Number of GMAVs in
Fresh Water 44
6 Dependence of the FAV and S on the Number of GMAVs in
Salt Water 50
7 Worst-Case Acute Values 55
8 Worst-Case FAVFs 56
9 Minimum Data Requirements Ranked by Acute Value . . . . 57
10 Worst-Case Acute Values When Daphnids Were Required . . 58
13. Worst-Case FAVFs When Daphnids Were Required 59
12 Worst-Case Acute Values When Salmonids Were Required . . 60
13 Worst-Case FAVFs When Salmonids Were Required 61
14 Worst-Case Acute Values When Both Daphnids and
Salmonids Were Required 62
15 Worst-Case FAVFs When Both Daphnids and Salmonids
Were Required 63
16 Summary of Geometric Mean Worst-Case FAVFs 64
17 Freshwater Summary FAVF5 (Version 1) 65
18 Freshwater Summary FAVFs (Version 2) 66
19 Saltwater Summary FAVFs (Version 2) 67
20 Overall Percentiles for Fresh Water 68
21 Effect of Daphnids on Overall Percentiles 70
22 FAVFs Based on Log-triangular Distribution 71
viii
-------
23 Final Acute-Chronic Ratios from Source Documents . . . . 72
24 FCVFs Corresponding to Various Percentiles 73
ix
-------
SECTION 1
INTRODUCTION
The most recent version of the “Guidelines for Deriving
Numerical National Water Quality Criteria for the Protection of
Aquatic Organisms and Their Uses” (Stephan, et al., 1985),
commonly known as the °National Guidelines”, describes a set of
procedures that can be used to derive water quality criteria for
aquatic life. The National Guidelines specify Minimum Data
Requirements (MDRs) that should be satisfied if aquatic life
criteria are to be derived. Each MDR specifies both a type of
test and a type of species for which an acceptable test result
required. If both a freshwater criterion and a saltwater
criterion are to be derived for a chemical, results of sixteen
acute toxicity tests and three chronic toxicity tests are
required. If a criterion is to be derived for a chemical in just
one of the waters, results of only eight acute tests and three
chronic tests are required.
The data required to derive either a freshwater or a
saltwater criterion are available for only a very small
percentage of the commercially important chemicals and the cost
of satisfying the MDRs for a chemical in either water will
usually exceed $100,000. In many situations, however, although a
criterion might be desirable, it might not be necessary. For
example, in order to determine whether a measured or predicted
concentration of a chemical in a body of water is likely to be
toxic to aquatic organisms, it might be sufficient to know that
there is a high probability that the aquatic life criterion would
be higher than the measured or predicted concentration. If the
measured or predicted concentration of the chemical was high
enough to warrant further consideration, a decision might be made
to generate additional toxicity data, but it still might be
possible to adequately evaluate the concentration of the chemical
without generating all the data necessary to derive an aquatic
life criterion. If, however, the chemical was found in enough
bodies of water at concentrations that are cause for concern, it
might be decided that a criterion should be derived for the
chemical. Thus if acceptable ways can be developed to evaluate
the acceptability of measured and predicted concentrations of
chemicals in ambient water using results of one or a few acute
toxicity tests, it might be possible to avoid the generation of
additional toxicity data, which might help ensure that resources
are applied to the chemicals for which data are most needed.
A.n aquatic life criterion derived according to the National
Guidelines consists of two numbers, the lower of which is the
Criterion Continuous Concentration (CCC). For many chemicals the
CCC is equal to the Final Chronic Value (FCV), which is usually
1
-------
derived by dividing the Final Acute Value (FAV) by the Final
Acute-Chronic Ratio (FACR), i.e., it is often true that:
CCC = FCV = FAV/FACR
The FACR is used to extrapolate from acute toxicity to chronic
toxicity. Although the CCC is often equal to a FCV, it might
also be based on field data or data concerning effects of
residues or effects on plants. The higher of the two numbers in
an aquatic life criterion is the Criterion Maximum Concentration
(CMC), which is derived by dividing the FAV by a factor of 2.
Because the FAV is an estimate of an LC5O or EC5O, this factor of
2 is used to reduce the concentration from one that kills or
affects 50 percent of the individuals of a species to a
concentration that kills or affects a much lower percentage of
the individuals.
If it is desirable to design a procedure that can be used to
draw a conclusion concerning the criterion for a chemical on the
basis of results of one or a few acute toxicity tests on that
chemical, it would seem appropriate to mimic the procedure
described in the National Guidelines for calculating a water
quality criterion for aquatic life. A simple way of imitating
this procedure would be to divide the lowest acceptable Genus
Mean Acute Value (GMAV) available for a chemical by a Final Acute
Value Factor (FAVF) in order to calculate an acute Level of
Concern (aLOC). The FAVF would be derived so that the
probability is high (e.g. , 95%) that the FAV f or the chemical
would be higher than the aLOC. Then the aLOC could be divided by
a Final Chronic Value Factor (FCVF) to calculate a chronic Level
of Concern (cLOC), on the assumption that there is a high
probability that the CCC would be higher than the cLOC. This
procedure can be summarized as:
(Lowest GMAV)/FAVF = aLOC and aLOC/FCVF = cLOC.
A measured or predicted concentration that is below the cLOC
should not cause much concern for possible toxicity to aquatic
life. This approach was used in Michigan’s “Rule 57” (Michigan
DNR, 1987) and in the “Guidelines for Deriving Ambient Aquatic
Life Advisory Concentrations” (U.S. EPA, 1987)
An alternative approach would be to develop factors that
would be applied to the mean of the available GMAVs rather than
to the lowest of the available GMAVs. This approach was not
considered here for two reasons. First, the occurrence of
“greater than” values will be more of a problem if factors are
applied to the mean, rather than the lowest, of the available
GMAVs. Second, adding values to a dataset can raise or lower the
mean, whereas additional values can only lower the lowest value.
Thus it should be easier to understand the dependence of the
FAVFs on the number of GMAVs if the FAVFs are applied to the
2
-------
lowest, rather than the mean, of the available acceptable GMAVs.
When only one GMAV is available, the choice between the mean and
the lowest GMAV is moot because the mean and the lowest value are
the same.
The purpose of the work described here was to derive FAVFS
and FCVFS, specifically taking into account the fact that for
chemicals for which few data are available, most available GMAVs
will be an acute value from one acute toxicity test and will
rarely be the mean of a number of acute values. In addition, it
was desired that the FAVFs be dependent on the size of the
dataset, so that the FAVF decreases as the uncertainty decreases.
This was implemented as a dependence on the number of MDRs that
were satisfied by the dataset. Because this work concerned FAVs
and FCVs, it was natural to base the derivation of FAVFs and
FCVFS on data contained in draft and final aquatic life criteria
documents (Table 1).
Section 2 of this report describes the formation of the
database on which the analyses described in Sections 3, 4, and S
were performed. Section 3 describes the analysis of the
dependence of the FAV on the number of GMAVs in datasets that
satisfy the eight MDRs concerning acute toxicity. Sections 4 and
5 describe two methods that were used for empirical calculation
of FAVFs. The method described in Section 4 calculates worst-
case FAVFs. In Section 5, two versions of a different method are
used to calculate FAVFs on the basis of random samples. Section
6 of this report describes the theoretical calculation of FAVFs.
Calculation of FCVFs is described in Section 7, and the results
are discussed in Section 8.
3
-------
SECTION 2
DATABASE MANAGEMENT
Empirical analyses concerning FAVs and FAVFs were based on
the toxicological database WQtJAL and the taxonomic database
TAXON. All toxicity values in WQUAL concern acute toxicity to
aquatic animals and are derived from the draft and final criteria
documents listed in Table 1. Both databases were originally
compiled at the U.S. EPA’S Environmental Research Laboratory in
Narragansett, RI, by David Hansen and Walter Berry with help from
Charles Stephan at the U.S. EPA’s Environmental Research
Laboratory in Duluth, MN. After the databases were transferred
to Duluth, they were checked against the source criteria
documents and some data were also checked against the original
references; all errors found were corrected. To facilitate
analyses, pertinent fields from WQIJAL and TAXON were used to
create a third database named COMBO. Data fields included in
COMBO are described in Table 2.
A computer program was written to determine whether each
dataset in COMBO satisfied the MDRs concerning acute toxicity
specified in the National Guidelines. Eleven freshwater and two
saltwater data sets were rejected by the program (Table 3). Each
rejected dataset was examined manually to verify that it did not
satisfy all eight MDR5. Most of the rejected datasets were from
criteria documents that were based on the 1980 version of the
National Guidelines, because the MDRs were not as strict in the
1980 version of the National Guidelines as they are in the 1985
version.
After the incomplete datasets were removed from
consideration, Species Mean Acute Values (SMAVs) were calculated
for the remaining 29 freshwater and 28 saltwater datasets
according to the following rules:
1. Acute values with B, E, H, or tv1 in the Remark field (Table 2)
were not used.
2. If, after application of rule 1, a dataset contained one or
more “flow-through measured” acute values (indicated by an “F”
in the Technique field and an “M” in the Measurement field)
for a species, only those values were used to calculate the
SMAV for that species, as specified in the National
Guidelines. If the dataset did not contain a “flow-through
measured” value for the species, then all acute values, except
as per rule 1, in the dataset for that species were used to
calculate the SMAV.
3. The freshwater datasets for ammonia, cadmium, chromium(III),
copper, lead, nickel, pentachiorophenol, and zinc were not
4
-------
subject to rule 2. In the criteria documents for these
chemicals, acute toxicity in fresh water is related to one or
more water quality characteristics such as hardness, pN, or
temperature. For each of these datasets all acute values,
except as per rule 1, were first adjusted to the same value(s)
of the charac’teristic(s) using the relationship given in the
criteria document and then all adjusted acute values were used
to calculate SMAVs.
For each of the 57 datasets, GMAVs and FAVs were then calculated
using the procedures described in the National Guidelines, except
that the FAV was never lowered to equal the SMAV for a sensitive
impbrtant species, as was done, for example, in the criteria
document for ammonia in fresh water. For the source documents
that were based on the 1985 version of the National Guidelines,
the SMAVs, GMAVS, and FAVs calculated by the computer were
compared with those given in the documents, and all discrepancies
were examined. (In the criteria documents prepared according to
the 1980 version of the National Guidelines, the FAV was
calculated from SMAVs, not GMAVs, and so only SMAVs could be
compared.) WQUAL, TA.XON, and COMBO were corrected so that all
remaining discrepancies were due to errors in the source
documents.
The scale factor S, which is the slope of the probability
plot of the four lowest GMAVs, was also calculated for each
dataset. The scale factor is an estimate of a iT , where a is
the standard deviation of the population from which the GMAVs are
assumed to have been drawn, which popu1 tion is assumed to have a
log-triangular distribution. The scale factor was calculated
because it might be useful in interpreting results presented in
Section 3 and so that it could be used in the theoretical
derivation of Final Acute Value Factors (FAVFs) in Section 6.
The scale factor is based on the four lowest GMAVs because these
determine the slope used in the calculation of the FAV.
The values of the FAV and S calculated from each dataset are
presented in Table 4.
5
-------
SECTION 3
DEPENDENCE OF THE FAV ON THE NUMBER OF GMAVs
The procedure described in the National Guidelines for
calculating an FAV requires at least eight GMAVs such that the
eight MDRs concerning acute toxicity are satisfied. This
procedure was designed to be conservative in the sense that, on
the average, greater uncertainty (i.e., fewer data) should result
in more protection (i.e., a lower FAV). The intent was to
establish this relation between uncertainty and conservatism by
(a) specifically requiring data for one or more generally
sensitive species in the MDRs in addition to requiring data for a
variety of species, and (b) calculating the FAV as a
percentile so that the amount of extrapolation below the lowest
GMAV depends on the number of GMAVs that can be calculated from
the dataset. (When there are more than twenty GMAVs, the FAV is
usually calculated by interpolation between the two lowest GMAVs
rather than by extrapolation below the lowest GMAV). Thus the
procedure was intended to produce, on the average, higher FAVs as
the number of GMAVs increases above 8. The intended relation
between uncertainty and conservatism will be accomplished,
however, only if the MDRs result in bias toward an overabundance
of sensitive species for all or nearly all chemicals.
The actual relationship of the FAV to n (= the number of
GMAVs) was determined by calculating FAVs from subsets (i.e.,
samples) of acute values that were randomly selected from the
dataset. For each subset the first eight acute values were
randomly selected, subject to the restriction that each acute
value satisfied a different MDR. Additional acute values, for
n>8, were selected so that each new value was for a genus which
was not already represented in the subset. Because the mean FAV
obtained here forn=8 was to be used in Sections 4 and 5, this
sampling was designed to be compatible with the sampling used in
Sections 4 and 5. Thus acute values, not SMAVs or GMAVs, were
randomly sampled and the presence of a B or M in the Remarks
field and the “flow-through measured” rule were ignored, as
explained in Section 4.
For each dataset the subset sizes used began with n=8 and
progressively increased by 2 (i.e., 8, 10, 12,...) up to, but not
including, the total number of GMAVs available from the dataset.
One hundred subsets were randomly selected for each of the subset
sizes that could be used with any one dataset and then 100 FAVs
were calculated. The geometric mean and the minimum and maximum
values of the FAV are presented in Tables S and 6 for each subset
size used with each dataset. Tables 5 and 6 also contain similar
data concerning the scale factor S. The values of FAV and S
given in Table 4 were calculated using a different set of rules
and so are not comparable to the values given in Tables 5 and 6.
6
-------
No values are in Table 5 for phenanthrene, thallium, and
tributyltin or in Table 6 f or cyanide because only eight GMAVs
could be calculated from these datasets.
Many datasets showed the trend of higher mean FAVs with
higher values of n, although for some datasets the mean FAV
increased very little as n increased. For a few datasets,
however, the geometric mean FAV decreased as n increased.
Especially for pentachiorophenol in fresh water and copper in
salt water, the mean FAV decreased as n increased over a wide
range of values of n. The difference between increasing and
decreasing trends did not seem to be related to the scale factor.
For most datasets, the standard deviations of both the FAV and S
decreased as n increased.
One way to summarize the data presented in Tables 5 and 6 is
to examine the ratio of the mean FAV for n=].8 to the mean FAV for
n=8 for those datasets for which at least 18 GMAVs could be
calculated. (The subset size of 18 is used merely because it is
a reasonably large size for which many ratios can be calculated.)
Twelve freshwater datasets and ten saltwater datasets contained
acute values for more than 18 genera; the 22 ratios are:
RATIO NUMBER
6.8 1
4.4 1
3-4 2
2-3 4
1 - 2 11
0.6 - 1 3
The range of the ratios is about a factor of ten. For most
datasets there was a small to moderate increase in the FAV when n
increased from 8 to 18, but for a few datasets the increase was
more than a factor of 3. These results show that the intended
conservatism was achieved by a factor of two or more for only 8
of the 22 datasets. For 3 of the 22 datasets the relationship
between the FAV and the number of GMAVs was the opposite of that
intended.
7
-------
SECTION 4
FAVFs BASED ON HIGHEST ACUTE VALUES
A Final Acute Value Factor (FAVF) is intended to be used to
calculate an aLOC from the lowest GMAV in a dataset. The FAVF is
intended to be conservative in the sense that, on the average,
greater uncertainly (i.e., fewer data) should result in more
protection, i.e., in a lower aLOC. In other words, a lack of
data should rarely result in underprotection, which can only be
achieved by having overprotection most of the time. Both
empirical and theoretical methods were used to derive FAVFs and
if both kinds of methods are used appropriately, the results
should be comparable. Two kinds of empirical calculations were
performed. The worst-case approach is described here in Section
4, whereas a statistical approach is described in Section 5.
After these two kinds of empirical approaches are discussed in
Sections 4 and 5, a theoretical approach is discussed in Section
6.
The objective of the analyses presented in this section was
to determine the worst-case (i.e., largest) FAVF that could be
obtained from each dataset by using one through eight acute
values that are actually in the dataset, where each selected
acute value had to satisfy a different MDR. To accomplish this,
the eight highest acute values that collectively satisfied the
eight MDRs were selected from each dataset. This selection of
acute values ignored the presence of a B or M in the Remarks
field and also ignored the “flow-through measured” rule. These
portions of the rules discussed in Section 2 were ignored here
because FAVFs are intended to be used with chemicals for which
few toxicity data are available. Some types of decisions that
might be possible with large datasets are unlikely to be possible
with small datasets. For example, when few data are available,
the B and M remarks and the “flow-through measured” rule will
rarely be applicable. (Acute values adjusted for water quality
were used for ammonia, pentachiorophenol, and some metals even
though such adjustments are not likely to be possible when
dealing with small datasets.) FAVFs are intended, however, to be
applied only to acute values that have been subjected to
screening procedures similar to those described in the National
Guidelines. The GMAVs and FAV5 discussed in Section 2 and
presented in Table 4 are intended to be calculated from datasets
that satisfy the MDRs and therefore will generally be based on
more acute values.
The eight highest acute values selected from one dataset,
with the restriction that each acute value had to satisfy a
different MDR, were ranked in descending order. These ranked
acute values were the worst-case subsets of sizes N=l to N=8,
where the highest of the eight acute values was considered the
8
-------
subset of 1, the two highest values were considered the subset of
2, and so on. The acute value that was the subset size of 1 was
the highest acute value in the data set because each acute value
wil,l satisfy a MDR. For each dataset, the lowest value in each
of the eight subsets was then divided by the geometric mean FAV
for the subset size of n=8 given in Table 5 for the same dataset.
(Because Table 5 does not contain values for phenanthrene,
thallium, and tributyltin, FAVs for these chemicals were obtained
from Table 4.) These eight quotients are, therefore, the worst-
case FAVFS that can be generated from the dataset using acute
values that are actually in the dataset. This analysis was
conducted on freshwater datasets only.
For each of the 29 freshwater datasets, the ranked eight
highest acute values that satisfied the eight MDRs are presented
in Table 7. Table 8 presents the worst-case FAVFs, which were
calculated by dividing the values for a dataset in Table 7 by the
mean FAV given in Table 5 (or Table 4 when necessary) for the
same dataset. Table 8 also shows the arithmetic and geometric
means across datasets. The values were not tested for normality
on either an arithmetic or logarithmic scale, but quotients are
usually more like a log-normal than a normal distribution. The
geometric mean worst-case FAVFs decreased from 541 for only one
(i.e., the highest) acute value to 6 when all eight MDRs were
satisfied. Interestingly, the worst-case FAVFs in Table 8 for
antimony(III), chromium(VI), copper, parathion, and thallium for
subset size of 7 are more than 10 times greater than those for
subset size of 8. Also, in Table 8, nine of the 29 FAVFs for
subset size = 8 are greater than 10.
Analyses were also conducted on the effects of requiring
that each subset, regardless of subset size, contain an acute
value for one or more sensitive species. Daphnids and salmonids
were tested as likely sensitive species. In the case of
daphnids, for example, when the highest acute values were
selected from a dataset, the highest acute value for a daphnid
was selected as the subset of 1 and thus was contained in each
successively larger subset in the sequence of eight subsets from
that dataset. Table 9 shows the MDR that was satisfied by each
of the acute values that are in Table 7. For example, the
positions of the daphnid values (code Dl; see Table 2) within
rows shows the ranking of daphnids relative to the other types of
species. In this worst-case analysis, the daphnid was the most
resistant type of species to acenaphthene and phenanthrene, but
the most sensitive to cadmium, chlorine, chrornium(VI), copper,
2,4-dimethyiphenol, mercury, nickel, parathion, selenium(IV), and
zinc. (This indication of relative sensitivity might not agree
with that seen in the source documents because a B and/or M in
the Remark field and the “flow-through measured” rule were
ignored in this worst-case analysis.) Table 10 presents the
highest acute values with the restriction that each subset had to
contain the highest acute value available for a dapimid, and
9
-------
Table 11 presents the corresponding worst-case FAVFs. For most
datasets, requiring that the highest value for a daphnid be in
each subset reduced the FAVFs substantially. The geometric mean
FAVF averaged across datasets for N=1, for example, was reduced
from 541 to 22 (Tables 8 and 11). The impact of requiring the
highest acute value for a daphnid decreased as subset size
increased.
A similar analysis was conducted to examine the effects of
requiring that each subset, regardless of size, contain the
highest acute value for a salmonid. Table 12 presents the
highest acute values with the restriction that each subset,
regardless of subset size, had to contain the highest acute value
for a salmonid in the dataset, and Table 13 presents the
corresponding worst-case FAVFS. As with the daphnid, the mean
FAVF was substantially reduced by the presence of the highest
acute value for a sairnonid, although the reduction was not as
great as that produced by daphnids. For N=1, the geometric mean
FAVF across datasets was reduced from 541 to 5]. by the inclusion
of the value for a salmonid.
The last analysis of worst-case FAVFs was to examine the
effects of requiring both the highest acute value for a salmonid
and the highest acute value for a daphnid in each subset of size
N=2. through N=8. Table 14 presents the highest acute values,
subject to the requirement that each subset in the sequence of
subsets from a dataset must contain both the highest acute value
for a daphnid and the highest acute value for a salmonid; Table
15 presents the corresponding worst-case FAVFs. This dual
requirement resulted in the lowest mean FAVFs, because salmonids
could be sensitive when daphnids were resistant and vice versa.
Requiring both a value for a daphnid and a value for a salmonid
yielded only a slight reduction in the mean FAVF compared to the
mean FAVF derived when a value for only a daphnid was required
(Table 16). The maximum reductior occurred at N=2, where the
FAVF was reduced from 19 (only daphnid required) to 12 (both
daphnid and salmonid required). Even when a value for a salmonid
was required (Tables 12 and 13), the more than a factor of ten
decrease when subset size increased from 7 to 8 occurred for the
same five chemicals noted above for Tables 7 and 8; when a value
for a daphnid was required, the large decrease only occurred for
antimony(III) and thallium (Tables 10, 11, 14, and 15) .
All four mean worst-case FAVFs at N=8 were 6 (Table 19) . At
subset size = 8, the mean worst-case FAVFs are the same for all
four cases because the four separate worst-case subsets obtained
from any one dataset will contain the same eight highest acute
values regardless of what, if any, of the conditions tested here
are imposed concerning sensitive species.
10
-------
SECTION 5
FAVFs BASED ON RANDOM SAMPLES OF ACUTE VALUES
Whereas the previous section concerned empirical FAVFS that
were deliberately calculated to be worst-case FAVFs, the
empirical FAVFs discussed in this section were based on subsets
(i.e., samples) obtained by random sampling. The empirical FAVFs
discussed in this’ section were calculated by dividing the lowest
of N randomly selected acute values by a FAV. Two versions of
this second empirical methodology were used. The major
difference between the two versions concerned how the FAV used as
the divisor was calculated:
1. In the first version the divisor was the geometric mean FAV
presented for n=8 in Tables 5 and 6, except that the FAVs in
Table 4 were used for phenanthrene, thallium, and
tributyltin in fresh water and for cyanide in salt water.
2. In the second version the divisor used for each of the eight
subsets in a sequence was the FAV that was calculated from
the subset of size N=8 in that sequence.
Thus, in the first version the same mean FAV was used with each
subset obtained from a particular dataset, whereas in the second
version a FAV was separately calculated for use with each
sequence of 8 subsets.
The first version of this random selection method was
applied to freshwater data sets only, whereas the second version
was applied to both freshwater and saltwater datasets. In both
versions of this method, subsets of eight acute values were
randomly selected from each dataset, with the restriction that
each selected acute value within the subset of eight had to
satisfy a different MDR. A variety of randomly selected subsets
were manually examined to verify that the MDRs were satisfied.
Each time a subset of eight acute values was created, subsets of
sizes 1 through 7 were also created, with each new acute value
increasing the subset size by 1. Thus when a subset of size
eight was selected, a sequence of eight subsets of sizes 1
through 8 was selected.
To avoid the possibility that the MDRs might confound the
effect of subset size (e.g., by always selecting an acute value
for a salmonid first) , the MDRs themselves were randomized before
each random selection of acute values; one acute value was then
selected at random to satisfy each of the randomized MDRs. The B
and M remarks and the “flow-through measured” rule were ignored
here for the same reason they were ignored in Section 4; adjusted
values were used here just as they were in Section 4.
The lowest acute value present in each subset of size 1’ to 8
was used as the numerator when calculating the FAVF. In the
first version of this method the random sampling procedure was
11
-------
repeated 99 times to generate 99 sequences of eight subsets from
each dataset; then the lowest acute value in each subset was
divided by the geometric mean FAV presented for n=8 in Tables 5
and 6, except that the FAVS given in Table 4 were used for the
four datasets named above. In the second version, 199 sequences
of eight subsets were generated from each dataset and the lowest
acute value in each subset was divided by the FAV calculated from
the subset of size N=8 in that sequence. The sampling was
repeated 199 times in the second version to account for the
probable increase in variability caused by using an individual
PAy, rather than a mean FAV, as the divisor.
The design and results of these analyses can be summarized
as follows:
Version 1 Version 2 Version 2
Freshwater Freshwater Saltwater
B subset sizes 8 subset sizes 8 subset sizes
29 datasets 29 datasets 24 datasets
99 subsets per 199 subsets per 199 subsets per
subset size subset size subset size
For each group of 99 FAVFs calculated in version 1, the median
( 50 th percentile) FAVF and the 95 th percentile FAVF were
calculated, resulting in twenty-nine medians and twenty-nine 95 th
percentiles for each subset size. Then the median and 95 th
percentile of the twenty-nine medians arid the median and 95 th
percentile of the twenty-nine 95 th percentiles were calculated.
Thus these calculations resulted in four summary FAVFs:
a. median of 29 medians.
b. 95 th perce ti1e of 29 medians.
c. median of twenty-nine 95 th percentiles.
95 percentile of twenty-nine g 5 h percentiles.
In addition to these two medians (a and c) and two 95 th
percentiles (b and d) that were calculated for each subset size,
one more median and one more 95 th percentile were calculated
using all 29.99 = 2871 FAVFs that had been calculated for each
subset size. These last two summary FAVFs are:
e. the overall median = the median of the 2871 FAVFs.
f. the overall 95 th percentile = the 95 th percentile of the 2871
FAVFs.
In the second version of the method, these same six summary FAVFS
were calculated from 199 FAVFs, rather than from 99 FAVFs, for
each subset size for each dataset. These six summary FAVFs were
calculated for each of the eight subset sizes for version 1 in
fresh water (Table 17), version 2 in fresh water (Table 18), and
version 2 in salt water (Table 19)
Of the six summary FAVFB, “a t ’ and “e” are expected to be
similar and robust because they are weighted and unweighted
medians of the dataset. Indeed, the medians of the 29 medians
are similar to the overall medians in Table 17; the same is true
12
-------
for Table 18 and for Table 19. In contrast, the 95 percentile
of the 95 th percentiles is an extreme summary FAVF; it is always
the largest of the six summary FAVFs, and often by a large
margin. The interpretation of the 95 th pe...centile of the 29
medians is somewhat difficult and less relevant. Thus, the focus
here will be on the median of the twenty-nine 95 th percentiles
(i.e., “c”) and the overall 95 th percentile (i.e., “f”).
In the course of these analyses, a concern arose as to
whether there was a “weighting” effect of taking 99 random
subsets from datasets for which the total number of possible
combinations was fewer than 99, because some combinations would
necessarily be obtained more than once. To evaluate this,
results obtained using 99 random samples were compared with
results obtained using each possible combination once and only
once. For example, if a dataset contained 16 acute values and
sampling was done for the subset size of N=1, this dataset would
contribute only 16 subsets if each possible combination was used
only once, whereas each combination would be used about 6.5 times
on the average if 99 subsets were taken. Calculations of medians
and 95 th percentiles showed that the two sampling strategies did
not produce substantially different results; for example, summary
FAVFs calculated as overall 95 percentiles using the first
version at a subset size of N=]. were 10,500 and 13,600 for the
two sampling strategies.
The effect of requiring that each subset, regardless of
subset size, contain an acute value for one or more sensitive
species was also examined using this empirical method. For
freshwater datasets daphnids and salmonids were used as sensitive
species, whereas for saltwater datasets crustacean species in the
families Mysidae and Penaeidae and fish species in the genus
Menidia were used as sensitive species. The saltwater MDRs
require an acute value for either a mysid. or a penaeid or both,
partly because these are considered sensitive species. Most of
the saltwater datasets also contain a value for a species in the
genus Menidia . An examination of the results obtained with
salmonids and daphnids in fresh water and a review of the
saltwater datasets indicated that the three kinds of saltwater
species named above probably would have a major impact on the
saltwater FAVF. In order to have the best comparison it was
desirable to use only datasets that contain acute values for all
the kinds of species that were to be tested as sensitive species.
Many saltwater datasets contain a value for a mysid or a penaeid
but not for both, so these two were considered together as one
kind of sensitive species, just as they are considered together
in the saltwater MDRS. All but four of the saltwater datasets
contained a value for a species in the genus Menidia , so these
four data sets were not used in this analysis. Therefore the
FAVFs in salt water were examined using 24 datasets.
13
-------
Similar to the procedure used in Section 4 to assess the
impact of sensitive species, additional 99 or 199 sequences of 8
subsets were randomly selected with the restriction that an acute
value for the sensitive species was chosen first (e.g., at N=1
the acute value chosen was a randomly selected value for a
daphnid). Each subset of 8 was completed with the remaining M]JR5
being satisfied in random order. Then, in order to obtain a
better indication of the effect of the sensitive species with the
second version of the method, in another group of 199 sequences,
each subset of 8 was drawn by excluding the sensitive species
until the eighth (i.e., the last) acute value was drawn (e.g.,
the requirement for a daphnid was satisfied last). This allowed
comparisons of subsets of sizes 1 to 7 that did and did not
contain an acute value for the sensitive species.
Table 17 summarizes the results obtained with the first
version of this method, which was used only with freshwater acute
values. For the four different sampling strategies, which
differed in their consideration of sensitive species, the six
summary FAVF5 are presented for each of eight subset sizes.
Subset size had a strong effect on the FAVF. In several cases,
FAVFS showed an approximately exponential decrease as subset size
increased. Requiring a randomly selected acute value for a
daphnid in every subset greatly reduced the FAVF for small subset
sizes; the reductions that occurred at subset sizes of 2 through
8 were less dramatic. Adding a value for a salmonid to a subset
that already contained a value for a daphnid usually had little
effect on the FAVF.
Results obtained using the second version of this method
with freshwater acute values (Table 18) were quite similar to
those obtained using the first version of the method (Table 17).
The overall 95 percentile FAVFs generated using the two
versions were generally within a factor of two, as were the
medians of the 95 th percentiles. The same exponential
relationship of decreasing FAVF with increasing subset size was
evident with the second version. FAVFS generated when daphnids
were required were consistently much lower than FAVFs generated
without concern for sensitive species, and FAVFs generated when
daphnids were excluded were consistently higher than FAVFs
generated without concern for sensitive species. As with the
first version of the method, adding a value for a salmonid to a
subset that already contained a value for a daphni.d had little
effect on the FAVF. As before, the median of the 95 percentile
was always substantially lower than the overall 9 gth percentile.
Saltwater acute values were studied using only the second
version of the method and generally produced FAVFs (Table 19)
that were somewhat lower than the comparable freshwater FAVFs
(Table 18). It is interest ing to compare Menidia or mysid-
penaeid required to both Menidia and mysid-penaeid excluded. In
this case, requiring Menidia and mysid-penaeid, respectively,
14
-------
even with the other excluded reduced the FAVF. The FAVF was
affected very little by adding a value for a species in the genus
Menidia to subsets that already contained a value for either a
mysid or a penaeid.
With both versions of the method, subsets of size 8 are
complete datasets for calculating FAVs. At N=8 with the second
version, the overall 95 th percentile FAVF is 8 in fresh water
(Table 18) and 4 in salt water (Table 19). For the first version
of the method, the overall 95 th percentile FAVF was 15, probably
because the mean FAV was less appropriate to individual subsets
than was the individually calculated PAy. Also, because the
sample of 8 is a complete dataset, the fact that the range of the
four overall 95 th percentile FAVFs in Table 17 is very small
indicates that the four separate sets of 99 samples of size 8
gave nearly identical results. The comparable overall 95 th
percentile FAVFs in Tables 18 and 19 indicate that the same
conclusions apply to the separate sets of 199 subsets. The
medians of the 95 th percentiles also had small ranges and the
values were substantially lower than the overall 95th percentiles
in Tables 17, 18, and 19. This low degree of variation at N=8
indicates that the 99 and 199 random samples used with the first
and second versions of this method, respectively, were
sufficiently large to produce robust results.
Version 2 of the method is probably the better version
because the FAV used with each sequence of subsets is derived
from the subset of size N=8 that is in the sequence itself.
Within sampling error, the two versions should give the same mean
FAV for each chemical and therefore the same median FAVF for a
chemical. Thus the median of the medians and the overall median
should be the same, within sampling error, for the two versions.
Indeed, the medians of the medians are in close agreement between
Tables 17 and 18, and the overall medians are also similar in the
two tables.
The results of the analyses have been presented mostly in
terms of the overall 95 th percentiles and medians of 95
percentiles. The data for overall percentiles from 50 to 95 (in
increments of 5) are presented in Table 20. The data for the
overall 50 th and 95 th percentiles are the same as those given in
Table 18. The effect of daphnids on the overall percentiles is
presented in Table 21.
15
-------
SECTION 6
FAVFs BASED ON THE LOG-TRIANGULAR DISTRIBUTION
The empirical simulations discussed in the previous section
gave FAVF5 by which the lowest GMAV in a subset should be divided
in order to have a known high probability that the resulting aLOC
is below the FAV. The FAV itself is an estimate of the 5 th
percentile assuming a log-triangular distribution in the lower
tail (Erickson and. Seephan, 1988). This section describes a
method for determining exact theoretical 95 percent confidence
limits on the 5 th percentile assuming a log-triangular
distribution. This exact theoretical method gives FAVFS that can
be compared with the empirical FAVFs.
The Triangular Distribution
The procedure used to calculate FAVs applies a log-
triangular distribution to the values in the lower tail (Erickson
and Stephan, 1988). The probability density function of the
standard triangular distribution at any value z is given by
i . +!z - J zO
f(z) =
__!Z Oz /
Integrating this density function and simplifying gives the
cumulative distribution function
+ — f z 0
12
F(z) =
The inverse of the cumulative distribution function for any
probability p is then
].2p -v’ 0 p 1/2
F’(p) =
- p) + ../ 1/2 p 1
Parameter and percentile Estimation
If a random variable X has a triangular distribution with
mean a and variance a 2 , then X can be viewed as a linear
16
-------
rescaling of z, such that X = + oz . Given data X( 1 )
a probability plot can be formed by plotting the ordered values
X(r) versus standardized triangular quantiles Qr = F 1(Nrl)
(Rice, 1988, p. 292). For points below the median, this
corresponds to plotting ordered values of X versus ‘i2p -
Thus by plotting X(r) versus Q(r) = 4 12(Nrl) the intercept is
1, which ±s an estimate of , and the slope is 0, which is an
estimat,e of a.
When an FAV is calculated using the procedure described by
Erickson and Stephan (1988), the FAV is usually calculated from
the four lowest ordered GMAVs X( 1 )x( 2 )x( 3 )x( 4 ) and their
standardized quantiles, i.e., their cumulative probabilities.
The FAV is then , which is the estimate of the pth percentile;
thus the FAV is the estimate of the value ‘that would be greater
than p percent of the values of X. If X is the average of X( 1 ),
X( 2 ), X( 3 ), and X( 4 ) and Q is the average of the corresponding
standardized quantiles, a line fitted to a plot of X(r) versus QCr)
must pass through (Q,X) . Therefore
and
= X -
The lower limit L of the resulting triangular distribution is
estimated as
f . = -
= X - (Q +
To estimate 2 ,, the pth percentile of a triangularly distributed
X, for p 0.5,
= $1 + O( 12p -
= + Oy’12p
If, instead of plotting X r) versus \ 12(N) X is plotted
versus N+l 1 the resulting slope is S = 8 /i . Hence
= + svI
The relationship between percentiles of a triangular
distribution can be used to derive the confidence bound on the
17
-------
5 percentile. Suppose 2 is the p percentile estimate from a
triangular distribution with scale factor S = â iT . For p 1
0.5 ____
= + -
If is the corresponding estimator of another percentile with
P2 < 0.5, ____
= + 8( /12p 2 -
Then ____ ____
= + 0 f 2 /I p 1 )
= + S( / -
For example, j estimates the 40 th percentile, then the 5
percentile is estimated by
= x 040 + S( / 05 — I0. 40)
= R 040 - 0.41S
If, however, p 1 > 0.5, then:
= I L + 8(- f12(l—p 1 ) + f )
Therefore, for p 1 > 0.5 and p 2 < 0.5:
X , = + 8 (,/I 2p 2 + /I2 (l—p 1 ) - 2J )
= + S( / + f1—p 1 —
If, for example, X 060 estimates the 60 th percentile, then:
x 005 = R 060 + S({0.05 + 0.40 —
= 0 60 - 0.56S
Calculating an LBFAV from the Lowest Acute Value When N<8
Let X( 1 ) ,X(N, be a random sample from a triangular
distribution, and let X be the lowest X. In the notation used
above, X( 1 ) was the lowest value, but X , will be used here for
clarity. An upper 95% confidence limit for the percentile
corresponding to X is 1 - 0.05’ . If X , corresponds at most
to percentile U (for upper limit), then X provides a lower
bound for the Uth percentile. A lower bound for the gth
percentile, is then:
18
-------
+ S(fO.05 - /t7) U 0.5
LBX 0 05 =
+ s( Io.o5 + — / ) u 0.5
where U = 1 - 005 1/N
The acute values in the lower tails of the aatasets are
assumed to be log-triangular. Because these values are
concentrations C 1 ,C 2 ,C , then X = ln(C ) , where “in”
signifies the natural logarithm, i.e., to base e. In particular,
Xmin = lfl(Cmin). Therefore:
LB Cmin/(e ) 5 U 0 . 5
LBFAV=e X 005
Cm 1 n/(e / )S U 0. 5
For example, assume a dataset that consists of the eight
GMAVs available for chiordarie in salt water: 120, 17.5, 16.9,
11.8, 6.4, 6.2, 4.8, 0.4 (Erickson and Stephan, 1988) . Then
imagine, for illustration, that the only data available happened
to be the highest four values: 120, 17.5, 16.9, and 11.8. To
calculate a IJBFAV, it would be necessary to select a value of S
large enough to’ give a conservative estimate of the 5 th
percentile. A large S means that the 5 th percentile potentially
could be far below the 1owe t observed value. Based on the
values of S in Table 4, a conservative value of S = 10 (see
below) would achieve 95% confidence in bounding the 5 th
percentile. Using this value of = 10 for illustration:
= ln(ll.8)= 2.47
U = 1 - 0.05’ = 0.527
ln(LBFAV) = 2.47 + ( /0.05 + f1 - 0.527 —
= 2.47 + 10 (0.223 + 0.688 — v’ )
= —2.56
LBFAV = e 2 56 = 0.077
Equivalently:
LBFAV =
= 0.077.
19
-------
When all 8 values are used, the FAV is 0.20. Hence even using
the highest four values with this conservative method gives a
LBFAV below the FAV. In this example this occurs partly because
X( 5 ) = 11.8 is less than one would predict on the basis of the
probability plot (Figure 1) of the lowest 4 values, and partly
because >
As another example, suppose that the only value available
was 120, which is the highest of the saltwater chiordane values.
Then U = 0.95, ln(120) = 4.79, and
ln(LBFAV) = 4.79 + ( fO.05 + /1 — 0.95 -
= 4.79 + 10 (0.223 + 0.223 —
= -4.88.
LBFAV = e 4 88 = 0.008
Equivalently:
LBFAV = 120/(e0.967)lO
= 0.008.
Again the LBFAV is below the FAV. This is also partly because
X( 8 ) = 120 is less than would be predicted from the probability
plot of the lowest four values and partly because > . In this
case, however, the LBFAV is less than the FAV also partly because
one of eight values is at most the 89 th percentile of the
dataset, whereas the upper limit for the percentile of a random
sample of size N=]. is U = 0.95.
Table 22 shows FAVFs by which the lowest value in the
dataset must be divided in order to have the specified percent
confidence that the LBFAV is less than the true 5 percentile.
The FAVFs for a given value of S>]. are the FAVFs for S=1 raised
to the S power. For example, the FAVF for 95% confidence with
S=2.5 and N=1 is (2.63)2 5 = 11.2. The FAVFS for 50% confidence
correspond to estimating the 5 th percentile, rather than
estirt’ating an upper limit on the 5 th percentile. Thus if S=lO
and there is one value (i.e., N=1), then that one value would
have to be divided by 126 in order to have a 50-50 chance of the
result being below the true 5 th percentile. If S=10 and N=8, an
estimate for the 5 percentile could be found by dividing the
lowest acute value by 1.9. In order to be 95% confident that the
resulting value is below the 5 th percentile, the lowest acute
value at N=8 would have to divided by 28.6.
Using a FAVF based on 95% confidence of being below the 5 th
percentile is conservative compared to calculating an FAV. When
20
-------
eight MDRs are satisfied, the FAV is an estimate of the 5 th
percentile. When fewer than eight requirements are satisfied,
the LBFAV will usually be below the 5 th percentile if S is chosen
to be sufficiently large.
Choosing a Value of S from a Range Of Values
If a number of different chemicals have different values of
a, it is necessary to determine what value of to use with a
randomly chosen chethical. The probability that a LBFAV computed
using a fixed is below the true 5 percentile is a mixture of
the probabilities for the chemicals. When a is used to compute
the LBFAV and a is the true standard deviation and UO.5 iS:
p { ; + a /r (yo.o5 + V1-u - v’ ) < + a( fi /0.O5 -
which can be rearranged to:
{ < -R l-U - + + (1-R) .05 }
The probabilities p { < z } are computed by the standard
log-triangular cumulative distribution function F(z) above.
For example, if R = a/a = 1.5, then the assumed scale
factor is 50% too large, and the computed LBFAV will usually
be even lower than desired, so the probability that the LBFAV is
21
-------
less than the true 5 th percentile will be above the targeted 95%.
If there are N=4 acute values, then tJ=O.527. Using the formula
given above for U0.5 and R=1.5, the actual probability that the
computed LBFAV is less than the true 5 th percentile is 0.999. If
there are N=7 values, then U=0.35. Using the formula given above
for U
-------
SECTION 7
FINAL CHRONIC VALUE FACTORS
Because a Final Chronic Value Factor (FCVF) is intended to
be used when a Final Acute-Chronic Ratio (FACR) cannot be
derived, the FACRS that are in the draft and final aquatic life
water quality criteria documents (Table 1) are the moat
appropriate data for use in determining a FCVF. The available
FACRS are listed in Table 23. This list contains only the FACRS,
not all the acute-chronic ratios (ACRs) given in the documents.
Only the FACRs are listed because ACRs are sometimes higher or
lower for acutely resistant species than for acutely sensitive
species and the FCVF is intended to be appropriate for use with
sensitive species. In the footnoted cases, however, the value
given in Table 23 is not the same as the FACR published in the
criteria document. For example, the value given in Table 23 for
aluminum is larger than the published FACR because the FCV in the
criteria document was lowered to protect an important sensitive
species. For aluminum, therefore, the value given in Table 23
was obtained by dividing the published FAV by the published
lowered FCV. No such changes were made because a Final Residue
Value was lower than a FCV. All values in Table 23 are rounded
to two significant figures.
Table 23 gives a “greater than” value for the FACR for
toxaphene in fresh water. This “greater than” value is a “right-
censored” value because the only information available is that
the FACR is somewhere to the right of 38 on the number line.
Because this right-censored value is below the empirical 95th
percentile, a nonparametric method coul’d not be used to estimate
high percentiles by counting up to a certain ordered value. To
estimate high percentiles in fresh water, therefore, the FACRs
were modeled with a parametric distribution and the parameters of
the distribution were estimated. A parametric method also
smoothes the data in the tails and allows a point estimate with
small sample sizes; sample sizes of 24 and 19 are considered
small for estimating 95th percentiles. The distributions of both
the freshwater and the saltwater FACRS are modeled well by a log-
normal distribution, particularly in the upper end of the
distribution, which is the most relevant portion here. Figures 2
and 3 show log-normal probability plots of the FACRs for fresh
and salt water, respectively. To estimate percentiles, the FACRs
are modeled as a log-normal distribution for values above 10.
Values below 10 are treated as left-censored so that they
contribute to the analysis but do not affect the shape of the
curve for the high percentiles that are most relevant. In fresh
water 62 percent of the FACRs are below 10, whereas 74 percent of
the saltwater FACRs are below 10.
23
-------
Due to the right-censored toxaphene value and the left-
censored values below 10, a computer program that could handle
both types of censoring was needed. In order to perform the
necessary analyses, maximum likelihood routines given in chapter
3 of Kalbfleisch and Prentice (1980) were . .evised to handle both
types of censoring simultaneously. Originally these routines
were for right censoring but a FORTRAN program was written to
handle right and left censored data for both normal and log-
normal distributions. For fresh water this program estimated the
mean and standard deviation to be 1.96 and 1.10, respectively, on
the natural log scale. Using a normal probability function and
transforming back to the original scale, the 95 th percentile was
calculated to be 43. For salt water the mean and standard
deviation were estimated to be 1.48 and 1.33, respectively,
resulting in an estimated 95 th percentile of 39. The 95 th
percentiles are quite similar, partly because of fortuitous
compensating effects. For most chemicals the freshwater and
saltwater FACRs are similar. The major exceptions are toxaphene,
which has a much larger FACR in fresh water, and chromium(VI),
which has a much larger FACR in salt water. Thus in both waters
there is one large value that is not in the other water, and the
resulting 95 th percentiles are quite similar. FCVFB
corresponding to percentiles from the 50 th to the g 8 th are given
for both fresh and salt waters in Table 24. The value of 25,
which was recommended by Kenaga (1982) and Call, et al. (1985),
was the 87 th percentile for fresh water and the 9O percentile
for salt water.
Although the median is the same as the 50 th percentile, the
median (i.e., the 50 th percentile) can be calculated in different
ways. The parametric method described above produced medians of
7.1 and 4.4 in fresh water and salt water, respectively (Table
24). A nonparametric median is often defined as the middle value
if the number of values is odd and as the average of the two
middle values if the number of values is even. This
noriparametric median is 7.8 in fresh water and 4.0 in salt water,
both of which are below 10. The freshwater nonparametric median
of 7.8 is between the 50 th and 55 th percentiles for fresh water,
whereas the saltwater nonparametric median of 4.0 is below the
50 th percentile for salt water.
24
-------
SECTION 8
DISCUSSION
The acute values in some of the datasets used in these
analyses could not be reviewed as much as desired. Additional
review would have been desirable because the screening procedures
used in the preparation of the 1980 criteria documents were not
as stringent as those used for the later criteria documents. It
is hoped that additional review would not ‘have substantially
impacted the results of these analyses.
The results presented in this report are based on all of the
datasets that were available in draft and final aquatic life
criteria documents at the time this work was begun. Many of the
these pollutants are metals and pesticides, but some are
industrial organic chemicals and others are such pollutants as
ammonia, chloride, chlorine, and cyanide. Thus these results are
based on datasets for a wide variety of pollutants in both fresh
and salt water and are likely to be’ applicable to a wide variety
of pollutants.
The results concerning the relationship of the FAV to the
number of GMAV5 from which it is calculated indicate that the
MDRs do not provide the intended conservatism for all datasets.
In Tables 5 and 6, the mean FAV decreased (rather than increased)
as “ r i ’ 1 increased for some chemicals, especially pentachiorophenol
in fresh water and copper in salt water. Another problem is that
nine of 29 worst-case FAVFs for n=8 in Table 8 were greater than
10; one was 835. Additional analyses might be able to indicate
desirable changes in the MDRs.
Another major result of these analyses is that the range of
acute values for any one chemical is likely to be much larger
than the range of ACRs among chemicals. The forty-three FACRS,
24 of which were in fresh water and 19 of which were in salt
water, ranged from 2 to 51, although one value is only known to
be greater than 38. In contrast, the overall 95 percentile
FAVF5 ranged from 4 to 20,000. Thus there is much more
uncertainty in the FAVFs than in the FCVFs. This can be
considered fortunate because acute values are certainly cheaper
to determine than ACRs. The large range in FAVFs is not the
result of assumptions about distributiois, etc.; each FAVF used
in the calculations for Tables 8, 11, 13, and 15 through 21
corresponds to real data.
The range in the FAVF ‘is due in part to variation among
replicate tests on a chemical using the same species. In a few
cases the agreement among replicate results is very poor. When
many data are available for a chemical, and especially for a
species, the chances of detecting “outliers” is much greater.
25
-------
For chemicals for which few data are available, SMAVs and GMAVs
will usually be the result of one test and will have greater
variability and uncertainty than SMAVs and GMAVs based on results
of many tests. This could be taken into account in the
determination of FAVFS.
The studies of sensitive species produced some unexpected
results, especially for salmonids. The summary FAVFs were so
much lower when daphnids were required than when salmonids were
required that it was deemed necessary to understand why these
results did not agree with the conventional wisdom that both
daphnids and sairnonids are sensitive freshwater species.
Examination of the ranked GMAVs for the 29 freshwater datasets
revealed that both daphnids and salmonids were amOng the most
sensitive species in about the same percentage of the datasets.
It was discovered, however, that even when a daphnid species
appeared to be resistant, the species was rarely more than 200
times more resistant than the most sensitive species in the
dataset. In contrast, salmonids were sometimes up to 44,000
times more resistant than the most sensitive species in the
dataset. Thus whether a species (or group of species) will have
a major impact on the FAVF depends mostly on the percent of
datasets in which the species is very resistant.
Judgments concerning the relative sensitivities of species
will depend on the criteria used to define “sensitive”. For
example, salmonids would probably be considered sensitive
freshwater species based on the worst-case FAVFs given in Table
16, but they would not be considered sensitive species based on
the FAVFs given in Tables 17 and 18. Another consideration is
the comparison that is made. Probably the best way to judge the
impact of a sensitive species is to compare results obtained when
the species is required to be present and when the species is
required to be absent. Such comparisons were made for daphnids
in Table 18 and for sensitive saltwater species in Table 19.
Studies of sensitive saltwater species indicated that two
kinds of species can cause a large reduction in the FAVF.
Although salmonids did not reduce the freshwater FAVFs, fish
species in the genus Menidia did reduce saltwater FAVFs. In
addition, crustacean species in the families Mysidae and
penaeidae also caused large reductions. When two kinds of
species deserve special consideration, there are more possible
combinations of inclusion and exclusion that can be tested (Table
19)
Results of computerized studies concerning the kinds of
species that can cause large reductions in FAVFs need to be
examined closely to determine exactly which species are being
studied. For example, in the freshwater family Daphnidae the
only genera for which many acute values are available are
Ceriodaphnia, Daphnia , and Simocephalus . Similarly for saltwater
26
-------
crustaceans, Mysidopsig and Penaeus are the only genera in the
families Mysidae and Penaeidae, respectively, for which many data
are available. Species within a genus probably have sufficiently
similar toxicological characteristics that generalization within
a genus is possible, but generalizations within a family might be
more risky.
The theoretical bounds on the 5 th percentile in Table 22
correspond most directly with the cases in Tables 18 and 19
labeled “any family” where no consideration was given to
sensitive species. Application of the iterative numerical
calculation procedure discussed at the end of Section 6 to the
values of S given in Table 4 indicated that value of S=8.2 would
give a 95 percent confidence level. The FAVFS in Table 22 for
S=7.5 and 95 percent are between the medians of the 95 th
percentile and the overall 95 th percentiles in Tables 18 and 19
at small sample sizes, but are higher at large sample sizes. The
FAVFs are not expected to be exactly the same for three reasons.
First, data for freshwater and saltwater species are considered
separately in Tables 18 and 19, but are considered togetijer in
Table 22. Second, the FAVFs in Tables 18 and 19 answer slightly
different questions than the FAVFs in Table 22. The overall g 5 th
percentile FAVFs in Tables 18 and 19 give divisors such that one
is 95 percent confident that the resulting L2FAV is less than the
FAV that would be obtained from the sample when the eight MDRS
are satisfied. The g 5 th percentile FAVFs in Table 22 give bounds
on the true population 5 th percentile. Third, the acute values
chosen to satisfy different of the eight MDRs are not truly
independent as assumed by the derivation using the log-triangular
distribution. Thus, the simulations will not match any
theoretical results exactly. Despite these three reasons why the
values in Table 22 will not agree with those in Tables 18 and 19,
the overall 95 th percentile FAVFs and the medians of the 95 th
percentiles for “any family” are similar in magnitude to the
FAVFs in Table 22 for S=5.O and S=7.5 at 95 percent. The overall
medians and the medians of medians for “any family” in Tables 18
and 19 are between the FAVFs in Table 22 for S=5 and S=7.5 at 50
percent.
Similarly, the geometric mean worst-case FAVFs given in
Table 16 are quite close to the medians of the 95 th percentiles
given in Tables 17 and 18. This is in spite of the fact that
this is a comparison of geometric means in Table 16 with medians
in Tables 17 and 18; in addition, this is a comparison of “worst-
case” value in Table 16 with 95tl percentiles in Tables 17 and
18. The general agreement between the values in Tables 16, 17,
18, 19, and 22 confirms that all of these different approaches
are reasonably appropriate and that none of them give results
that are “outliers”. For example, the very large FAVFs given in
Tables 18, 19, and 22 for small subset sizes are not an artifact
of the analyses. These very large FAVFS are due to a very large
range of acute values in more than 5 percent of the datasets.
27
-------
When eight MDRS are satisfied, FAVF5 would not be used
because the FAV calculation procedure would be used. It is
interesting, however, that when all eight MDRs are satisfied, the
medians of medians and the overall medians in Tables 17, 18, and
19 range from 1.7 to 2.1. Also, the FAVFs for fifty percent in
Table 22 range from 1.0 to 1.9. on the average, the FAV is
within a factor of 2 of the lowest GMAV even when there are only
eight GMAVs. (It is comforting that the results obtained using
empirical and theoretical approaches are similar.)
Even at the 95 th percentile, most of the FAVFs are not large
when all eight MDRs are satisfied. The overall percentile
FAVFs obtained at N=8 range from 4.4 to 8.5 and the medians of
the g 5 th percentiles range from 2.8 to 3.4 (Tables 17, 18, and
19). Similarly,, the mean worst-case FAVFs at N=8 are 6 (Table
16) and the FAVFs for 95 percent in Table 22 are above 8 only for
the larger values of S.
These FAVFs for eight GMAVs are small considering that the
extrapolations are from n=8 to the 95 th percentile. The lowest
GMAV is an estimate of the p = 100/(n+1) = 100/9 = l1
percentile when n=8, whereas 19 values would be needed in order
for the lowest value to be an estimate of the p = 100/20 =
percentile. These small factors’ might be due in part to the use
of the log-triangular distribution because other distributions
such as the log-normal and log-logistic would have longer tails
and would give lower values when extrapolating from n=8 to the
95 th percentile. The ratios of the FAVs for n=18 and n=8 given
on page 7, however, indicate that large factors are not necessary
when n=8. This is probably due in part to the conservatism built
into the MDRs, but it does indicate that the small factors and
the log-triangular distribution are appropriate.
Data in Table 18 can be used to illustrate a relation
between FAVFs based on medians and FAVFs based on high
percentiles. Not only will each additional GMA.V lower the FAVF
at a high percentile, it will also lower the FAVF for the 50 th
percentile because a new GMAV can lower, but not raise, the
lowest GMAV. Each additional GMAV will make the lowest GMAV
closer to the FAV on the average. The overall medians with
daphnid required give the values that, on the average, should be
divided into the lowest acceptable GMAV in order to get the best
prediction of the FAV. Thus if the FAV of a chemical is 100
ug/L, the average acute value obtained with a daphnid would be
489 ug/L, because this would give 100 ug/L when divided by 4.89.
Similarly, the average lowest GMAV obtained when succeeding tests
are conducted to satisfy MDRs would be 323, 260, 236, 218, 202,
and 192 ug/L. Each of these, when divided by the FAVFs of 3.23,
2.60,2.36, 2.18, 2.02, and 1.92, respectively, would give a FAV
of 100 ug/L. If this series of lowest GMAVS was not used to
estimate the FAVs, but was used to derive aLOCs using the overall
95 th percentile with daphriid required, the results would be:
28
-------
These aLOCs consistently increase, but there is little difference
between one and two GMAVs. As expected, these aLOCs are larger
than those calculated using the overall 95 percentile and are
smaller than the FAV of 100 ug/L.
The aLOC increases somewhat as each additional MDR is satisfied.
of
Number
GMAVs
1
Average Lowest
GMAV (u /L)
FAVF
94.0
aLOC
(ug/L)
5.2
489
2
323
58.0
5.6
3
260
50.5
5.1
4
236
41.8
5.6
5
218
31.0
7.0
6
202
22.0
9.2
7
192
13.1
14.7
It can be seen that the aLOC does not smoothly increase as the
number of GMAVs increases.
Similarly if the data for the overall 80 th percentile with
daphnid required in Table 20 are used, the results are:
of
Number
GMAVs
1
Average Lowest
GMAV (u /L)
FAVF
20.50
aLOC
(u IL)
24
489
2
323
13.12
25
3
260
8.59
30
4
236
6.53
36
5
218
5.04
43
6
202
4.00
50
7
192
3.61
53
If the data for the median
calculate the lowest Gr4AVs,
percentiles is used as the
results are:
of medians in Table 18 are used to
and then the median of the 95th
FAVF to calculate the aLOCs, the
Number
Average
Lowest FAVF aLOC
of
GMAVS
GMAV
(u /L) (ug/L)
1
396
21.89
18.1
2
319
12.95
24.6
3
240
7.98
30.].
4
226
6.99
32.3
5
213
6.09
35.0
6
207
5.20
39.8
7
196
4.30
45.6
29
-------
Rather than using all of the ACRs in each source document,
only the FACR given in a source document was used. The choice
between using the FACR and using all of the ACRs in a document
cart be considered from the standpoint that using all of the ACRs
would increase “n and make the derivation more robust. A much
more important consideration, however, is that some ACR5 in some
documents are not relevant to sensitive species. Therefore, the
FACR needs to be derived from the ACRS that are available on a
chemical-by-chemical basis so that the FACR is as appropriate as
possible for use with the species that are sensitive to that
chemical. Thus it is more appropriate to use the FACR, rather
than the individual ACRs, in the derivation of the FCVFs. In the
context of this project, the toxicological consideration is much
more important than the statistical consideration.
The results presented in this report should be useful for
interpreting data concerning the toxicity of chemicals to aquatic
life when insufficient data are available to allow derivation of
water quality criteria for aquatic life using the procedures
described in the National Guidelines. Use of the results
presented herein will usually be in a risk management context,
and so choosing the FAVFs and FCVFS to use will be a risk
management decision. Some relevant considerations are:
a. More people are likely to be able to understand the empirical
results than the theoretical results.
b. Even though they are empirical results, the worst-case results
will, by definition, occur rarely.
c. To be toxicologically appropriate, the FCVFs must be derived
on a by-chemical basis. The FAVFs will usually be used with
FCVFs and so, to be consistent, the FAVF9 should also be on a
by-chemical basis. The summary FAVFs that are on a by-
chemical basis (i.e., based on a percentage of chemicals and
therefore weighted by chemical) are the ones that are medians
of percentiles, such as the median of medians and the median
of 95th percentiles. Aithough these are the only two
percentiles for which such summary FAVFs are given in this
report, those corresponding to other percentiles can be
derived.
d. The percentile on which the FAVF is based does not have to be
the same as the percentile on which the FCVF is based, but the
selection of one percentile should probably take into account
the other. If, for example, both the FAVF and the FCVF are
calculated as 95 th percentiles, the probability of the cLOC
being lower than the CCC will be greater than 95 percent.
Because the relative range of the FAVFS is much greater than
that of the FCVFs, it is likely that the probability of the
cLOC being below the CCC will depend almost entirely on the
larger range and so the probability might not be too much
greater than 95 percent. Additional analyses that combine the
FAVF and FCVF into one factor might be of interest. Having
two separate factors is desirable, however, because there will
30
-------
probably be chemicals for which a FAV can be calculated but a
FACR cannot. When a FACR cannot be calculated, the number of
available ACRs might be 0, 1, or 2. The possibility that a
FCVF might be used with a FAV might be taken into account in
the selection of the percentile that is to be the basis of the
FCVF. An alternative might be to derive two FCVFs: one to be
used when a FAV is available and one to be used when a FAV is
not available.
e. An important consideration in the selection of the percentiles
should be the percent of times that the acute and/or chronic
Level of Concern (aLOC and/or cLOC) is allowed to provide less
protection than would be provided by the FAV and/or CCC if
they could be derived.
f. A LOC is not a very good basis for evaluating the
concentration of a chemical in a body of water unless there is
a high probability of the evaluation being correct; a
statement concerning a high probability can be made only if
the LOC is based on a high or low percentile.
g. The fiftieth percentile is the most robust.
h. The fiftieth percentile FAVF provides the best prediction of
the FAV.
i. The magnitude of an FAVF or FCVF is directly related to the
uncertainty and to the level of protection. There are two
ways to reduce the uncertainty and the magnitude of an FAVF
without changing the level of protection (i.e., without
changing the percentile):
1. Conduct pertinent acute toxicity tests to satisfy
additional MDRs.
2. Use FAVFs that require the presence of one or more species
whose inclusion has been shown to result in smaller FAVFs.
For example, it has been shown that lower FAVFs apply to
datasets that contain certain species of daphnids.
j. The closer the percentile used is to the 50 th, the more cases
there will be in which the aLOC will provide a lower level of
protection than would be provided by the FAV. The closer the
percentile used is to the 50 th percentile, the less useful the
cLOC will be for identifying situations in which more data
should be generated, even thought the cLOC might not be a very
good prediction of the CCC.
k. The closer the percentile used is to the 50 th the less
definitive are the statements that can be made concerning the
effect of satisfying more MDRs. If the aLOC is a predicted
FAV, each additional MDR that is satisfied has an equal chance
of raising or lowering the predicted FAV.
1. To be consistent with the concept of establishment of MDRs for
deriving aquatic life criteria, the interpretation of fewer
data should take into account the greater uncertainty.
m. An aLOC or cLOC that is substantially below the FAV or CCC
will not cause additional data generation if the measured or
predicted concentration being evaluated is above the aLOC or
cLOC.
31
-------
n. The higher the percentile, the more cases in which additional
data will be generated unnecessarily.
Additional work could be done to provide more definitive
information concerning the items examined in this report. Such
additional work could involve the following:
1. Additional datasets could probably be added to COMBO, and
literature searches could be performed for data that could be
added to the datasets that are in COMBO. All data in COMBO
could be screened using updated procedures for reviewing the
acceptability of results of acute toxicity tests. Toxicity
‘tests could be conducted to fill gaps in datasets that satisfy
only six or seven MDRs.
2. An alternative approach could be used to study the dependence
of the FAV on the number of GMAVs by randomly selecting
subsets (n8) of GMAVs, with the restriction that the GMAVs in
each subset must satisfy the MDRs. The geometric mean FAVs
for the various subset sizes from n=8 to the maximum number of
GMAVs in the dataset could then be compared. This approach
would result in an FAV for the maximum number of GMA Vs that
would be the same as the FAV given in Table 4. Preliminary
results obtained using this approach with the freshwater
datasets were similar to the results reported in Section 3.
The relationship between the FAV and the number of GMAVs could
be used to study the impact of specific changes in the MDRs.
3. Analyses could be performed on an updated version of COMBO
using version 2 of the empirical method to calculate summary
FAVFs.
4. Variance could be compared within species, genera, and
families, and among families.
5. Analyses could be performed for specific kinds of chemicals,
such as organic chemicals whose mode of action is narcosis, if
sufficient data are available. Use of the results, however,
would require knowing that the mode of action of a particular
chemical is narcosis.
32
-------
REFERENCE S
Call, D.J., L.T. Brooke, M.L. Knuth, S.H. Poirier, and M.D.
Hogland. 1985. Fish Subchronic Toxicity Prediction Model for
Industrial Organic Chemicals that Produce Narcosis. Environ.
Toxicol. Chem. 4:335-341.
Erickson, R.J., and C.E. Stephan. 1988. Calculation of the
Final Acute Value for Water Quality Criteria for Aquatic
Organisms. PB88-214994. National Technical Information Service,
Springfield, VA.
Kalbfleisch, J.D., and R.L. Prentice. 1980. The Statistical
Analysis of Failure Time Data. John Wiley, New York, NY. 321
pp.
Kenaga, E.E. 1982. Predictability of Chronic Toxicity from
Acute Toxicity of Chemicals in Fish and Invertebrates. Environ.
Toxicol. Chem. 1:347-358.
Michigan Department of Natural Resources. 1987. Updated Support
Document for the Aquatic Chronic Value of the Rule 57(2)
Guidelines. Surface Water Quality Division, Lansing, MI.
Rice, J.A. 1988. Mathematical Statistics and Data Analysis.
Wadsworth, Pacific Grove, CA.
Stephan, C.E., D.I. Mount, D.J. Hansen, J.H. Gentile, G.A.
Chapman, and W.A. Brungs. 1985. Guidelines for Deriving
Numerical National Water Quality Criteria for the Protection of
Aquatic Organisms and Their Uses. PBB5-227049. National
Technical Information Service, Springfield, VA.
U.S. EPA. 1987. Guidelines for Deriving Ambient Aquatic Life
Advisory Concentrations. Office of Jater Regulations and
Standards, Washington, DC, and Office of Research and
Development, Duluth, MN.
33
-------
Figure 1. Probability Plot for Chiordane in Salt Water
>
C /)
0
0
L i,
C
C )
1
0
U,
0
L()
-1.5
-1.0 -0.5 0.0 0.5 1.0
Li)
d
d
1.5
34
-------
Figure 2. Log-normal Probability Plot of Freshwater FACRs
0
0
0 *
*
*
4
*
*
0 *
II **
*
*
*
L I,
*
**
*
**
* *
*
I I I
-3 -2 -1 0 1 2 3
35
-------
Figure 3. Log-normal Probability Plot of Saltwater FACRS
0
0
0 *
U-)
*
*
*
*
cc
00
* *
LL *
U , -
* *
* * *
* *
*
* * *
I I
-2 -1 0 1 2
36
-------
Table 1. Sources of Datasets
Chemical Watera Source ,Documentb
Acenaphthene
Acrolein
Aidrin
Aluminum
Ammonia
Ammonia
Antimony C III)
Arsenic ( III)
Cadmium
Chiordane
Chloride
Chlorine
Chiorpyri fos
Chromium ( III)
Chromium (VI)
Copper
Cyanide
DDT
Dieldrin
2,4 -Dimethyiphenol
Endosul fan
Endriri
Heptachior
Lead
Lindane
Mercury
Methyl parathion
Nickel
Parathion
Pentachiorophenol
Phenant hrene
Phenol
Selenium (IV)
Selenium (VI)
Silver
F
F
F
F
F
S
F S
F S
F S
F S
F
F
F
F
F
F S
F S
F S
F S
F S
F S
F S
F S
F S
F S
F
F
F
F
F
F
F
F
F
F
Draft: 10-11-88
Draft: 9-21-87
EPA 440/5-80-019
EPA 440/5/86/008
EPA 440/5-85-001
EPA 440/5-88-004
Draft: 8-30-88
EPA 440/5-84-033
EPA 440/5-84-032
EPA 440/5-80-027
EPA 440/5-88-001
EPA 440/5-84-030
EPA 440/5-8,6-005
EPA 440/5-84-029
EPA 440/5-84-029
EPA 440/5-84-031
EPA 440/5-84-028
EPA 440/5-80-038
EPA 440/5-80-019
Draft: 9-1-88
EPA 440/5-80-046
EPA 440/5-80-047
EPA 440/5-80-052
EPA 440/5-84-027
EPA 440/5-80-054
EPA 440/5-84-026
Draft: 6-7-88
EPA 440/5-86-004
EPA 440/5-86-007
EPA 440/5-86-009
Draft: 8-16-88
Draft: 9-13-88
EPA 440/5-87-006
EPA 440/5-87-006
Draft: 9-24-87
S
S
S
S
S
S
S
S
S
S
37
-------
Draft: 3-4-88
EPA 440/5-86-006
Draft: 6-16-88
Draft: undated
Draft: 9-25-87
EPA 440/5-87-003
a F = freshwater; S = saltwater.
b Published documents are titled “Anthient Water Quality Criteria
for “ and are available from the National Technical
Information Service, Springfield, VA. For unpublished
documents, the drafts used were the newest ones available as of
10-12-88.
Thallium
Toxaphene
Tributyltin
1, 2,4 -Trichlorobenzene
2,4, 5-Trichlorophenol
Zinc
F
F S
F S
F S
F S
F S
38
-------
Table 2. Fields Used in the Database COMBO
Record Number
Chemical
Water
F = Freshwater
S = Saltwater
MDR Codesa for freshwater species:
A = A salmonid, i.e., a species in the family
Salmonidae.
B]. - B13 = Families other than Salmonidae in the class
Osteichthyes.
C = A species that is in the phylum Chordata but is
not in the class Osteichthyes.
Dl = A daphnid, i.e., a species in the family
Daphnidae.
D2 = Other planktonic crustaceans.
E = A species that is a benthic crustacean.
F]. - F7 = Orders of insects.
Gi - G6 = Families in phyla other than Arthropoda or
Chordata, e.g., hydra, planaria, rotifers, worms,
snails, clams, mussels, bryozoans, etc.
X = A species that is nonresident and therefore should
- not have been used in the source document.
MDR Codesa for saltwater species:
Ml - M22 = Families in the phylum Chordata.
Ni = A mysid, i.e., a species in the family Mysidae.
N2 = A penaeid, i.e., a species in the family
Penaeidae.
N3 - N26 = Other families in the phylum Arthropoda.
P1 - P8 = Families in the phylum Annelida.
Si - S13 = Families in the phylum Mollusca.
Ri - R4 = Families in the phylum Echinodermata.
X = A species that is nonresident and therefore should
not have been used in the source document.
Genus
Species
Life Stage
39
-------
Remark
A = No remark (default).
B = This acute value is in the source document but was not
used in the calculation of the FAV iii the source document
or here for the reason given in the source document.
E = This acute value should not be used because it should not
have been included in the source document. (This remark
was usually because the species is not a resident
species.)
G = This acute value is a “greater than” value.
H = The hardness, or another pertinent water quality
characteristic, of the dilution water is unknown, so this
acute value cannot be adjusted and therefore was not used
in calculations in the source document or here. (This
remark is used only with those chemicals for which the FAV
is calculated from adjusted acute values.)
L = This acute value is a “less than” value.
M = This acute value was not used in the calculations in the
source document or here because an acceptable acute value
for this chemical is also available for a more sensitive
life stage of this species.
R = This acute value has been changed from that in the source
document, based on either (a) the reference used in the
source document or (b) a newer reference for the same
value.
Technique
F = Flow-through
S = Static
Measurement
M = Measured
U = Unmeasured
Acute Value
Adjusted Acute Value
a MDR Codes were designed so that an algorithm could be written
to determine which of the minimum data requirements concerning
acute toxicity were satisfied by any group of species.
40
-------
Table 3. Datasets Not Satisfying the Minimum Data Requirements
Chemical
Water
Requirement(s) not Satisfied’
Aidrin
Fresh
a
Chiordane
Fresh
a,b
Chiorpyrifos
Fresh
c
DDT
Fresh
a
Dieldrin
Fresh
a
Endosulfan
Fresh
a
Endrin
Fresh
a
Heptachior
Fresh
a,b
Lead
Fresh
b
Lindane
Fresh
ab
Toxaphene
Fresh
a
Aidrin
Salt
d
Chiordane
Salt
e
The minimum data requirements not satisfied are:
a. A species in a phylum other than Arthropoda or Chordata.
b. A species in any order of insect or any phylum not already
represented.
c. A planktonic crustacean.
d. A species in the family Mysidae or the family Penaeidae.
e. Several requirements were not satisfied.
41
-------
Table 4. Values of FAV and S for Datasets in Source Documents
Chemical Water Number S FAV
________________ ____ of GMAVs ____ ____
Acenaphtherie F 9 4 82 80 02
Acrolein F 12 3.83 6.15
Aluminum F 14 8 77 1496
Ammonia F 34 3 89 0 70
Ammonia S 18 2.41 0 46
Ant imony(III) F 9 12.80 176
Ant3.mony(III) S 11 2 28 2959
Arsenic(III) F 14 4 09 718
Ai-senic(III) 5 11 7 74 137
Cadmium F 44 19 38 8 93
Cadmium S 33 7.14 85 09
Chloride F 12 2 29 1720
Chlorine P 28 5.71 38 32
Chlorine S 21 3 29 25 24
Ch lorpyrifos S 12 5 71 0 02
Chromium(III) F 18 6.53 1968
Chromium(VI) F 27 4 40 31 49
Chromiuin(VI) S 21 2.59 2158
Copper F 41 1 90 18 09
Copper S 20 8 29 5.83
Cyanide F 15 1 83 62 67
Cyanide S 8 954 203
DOT S 16 S 58 0 14
Dieldrin S 19 5.62 0.62
2 ,4-Dimethyiphenol F 12 3.59 2669
2 ,4-Dimethyiphenol S 9 6 53 549
Endosulfan S 12 4 50 0.03
Endrin S 19 4.11 0.03
Heptachior S 18 13 98 0 07
Lead S 11. 2.20 287
42
-------
TABLE 4. (CONT INUED)
L.indane S 17 17 07 0.21
Mercury F 28 11 93 4 86
Mercury S 29 3.57 4.12
Methyl Parathion F 22 15 09 0.63
Nickel F 18 7.11 1577
Nickel S 21 9 51 151
Parathion F 30 17 88 0.12
Pentachiorophenol F 33 13 06 10.86
Pentachiorophenol S 17 3.91 25.06
Phenanthrene F 8 2.76 59.62
Phertanthrene S 10 7.19 15.46
Phenol F 28 5 19 7445
Selenium (IV) F 22 9 05 370
Selenium (IV) S 15 3 30 588
Seleniun (VI) F 1]. 14 59 25 65
Silver F 18 3 89 1.83
Silver S 19 2 02 14 60
Thallium F 8 9 82 46.85
Toxaphene S 15 3 47 0 42
Tributylt in F 8 6.14 0.30
Tributyltin S 18 3 37 0 53
1,2.4-Tr ich lorobeflzefle F 14 2 51 889
1,2 ,4-Trich lorobeflZene S 15 3 23 280
2.4.S .Trich lOrOpheflO l F 10 2.84 199
2,45-TrichlorophenOl S 10 0 67 473
Zinc F 35 6 32 127
Zinc S 28 3.57 177
FAV Final Acute Value.
S Slope factor.
GMAVs Genus Mean Acute Valuea.
43
-------
Table 5. Dependence of FAV and S on Number of Freshwater GMAVs
Chemical Number Final Acute Value S
of Mm. Mean Max Std Mm Mean Max. Std
_______________ GMAVs Dev Dev
Acenaphthene 8 189 189 189 0 00 2.84 2.84 2.84 0.00
Acrolein 8 2 57 6 64 23 67 6.84 2.74 5.37 8.02 1.84
Acro lejn 10 3.28 5 23 21.53 3.37 2.99 6 21 8 40 1.29
Aluminum 8 819 2644 14132 4678 1.30 4 73 8 04 2 53
Aluminum io 990 1887 14571 2694 1 43 6 65 8.89 1.75
Aluminum 12 1182 1861 10709 1847 2.44 7.34 9.62 1.25
Ammonia 8 0 31 0 67 1 46 0 29 0.33 1 97 4 81 1 20
Aznmona.a 10 0.36 0,71 1 57 0 27 0 22 1 91 4.34 1.04
Ammonia 12 0 40 0.74 1.20 0 24 0.43 1.73 5.02 1 14
Ammonia 14 0 48 0 70 1 30 0.22 0 47 2 09 4.35 1.02
Ammonia 16 0 49 0 68 1.18 0 19 0 44 2.35 4.46 0.96
Ammonia 19 0.53 0 70 1 17 0 18 0.48 2.30 4.66 1 12
Ammonia 20 0.56 0 67 1 21 0.15 0 34 2.75 4.38 0.96
Ammonia 22 0 58 0 67 1.21 0.12 0.33 2 98 4.54 0 82
Ammonia 24 0 60 0 66 0.94 0.09 1 40 3.29 4.63 0.76
Ammonia 26 0 62 0 68 1.14 0 1]. 0.58 3 23 4.45 0.80
Ammonia 28 0 64 0 69 0 96 0.08 1 51 3 40 4.57 0.65
Ammonia 30 0 66 0.69 0 84 0 06 2.36 3.55 4.31 0 55
Ammonia 32 0.68 0 70 0.85 0 05 2.44 3 72 4.65 0.47
Antimony(III) 8 142 169 192 25 12.53 12 73 12.88 0.17
Arsenic(III) 8 152 212 296 45 9 58 10 46 11.49 0.54
Arsenjc(III) 10 213 281 630 97 3.50 9.44 12.23 1.86
Arsenic(III) 12 260 441 674 178 3 81 6 73 12.77 3 32
Cadmium 8 0 52 3 62 38.11 8 22 3 52 9 27 17 13 3.42
Cadmium 10 0.90 5 30 42.96 10 43 2.96 8.44 18.69 3.43
Cadmium 12 1.49 6.57 70.61 11.98 1.90 8.23 15.29 3.23
Cadmium 14 1.86 6.80 44.69 9.45 1.79 8.91 18.53 3.11
Cadmium i 3 36 8.41 62.96 11.95 2 15 8.29 15.42 2.98
Cadmium 18 3.51 9.68 57.78 11.47 2.33 8.55 15.40 3.07
Cadmium 20 4 10 9 32 46.33 9.37 1.57 9.31 17 02 2.93
44
-------
TABLE 5. (CONTINUED)
Cadmium
22
4 78
9 11
38.14
9.11
0 84
9 39
16.93
3 18
Cadmium
24
5 58
10 62
87.71
12 05
0 87
9 24
19.75
3 59
Cadmium
26
6 42
10 45
41 67
6.98
0.91
10 26
15.69
2.79
Cadmium
28
7 21
11 01
45.03
8.77
0 94
10 24
15.82
3.20
Cadmium
30
7 89
11 56
40.02
7.23
1.91
1.0.63
16.42
2.99
Cadmium
32
8.62
11.33
39 23
S 10
1 97
12.09
14.77
2.06
Cadmium
34
9.39
11.06
19 10
3 26
9.61
13 02
16.71
1.43
Cadmium
36
10 20
12.72
35 33
4 82
2.09
12 64
15.52
2 24
Cadmium
38
11 06
12.70
22 46
3 43
10 15
13 63
15 94
1 46
Cadmium
40
11 97
13 10
23 07
2 99
10 40
14.15
16.34
1 15
Cadmium
42
12.93
13 57
24 SO
2.47
10 65
14.51
15 80
0 90
Cadmium
44
13 94
14.12
23.99
1.19
10.90
14.98
16.16
0 49
Chloride
8
231
346
583
140
3 66
5 29
6.68
1.07
Chloride
10
270
369
635
128
4 05
5 41
7 39
0 89
Chloride
12
330
385
687
104
4 40
5 46
6 77
0 64
Chlorine
8
17 62
40 19
69 14
17 27
0.31
1.76
5 00
1.41
Chlorine
10
21 89
40 66
71 44
19 01
0.28
1.73
4.88
1.55
Chlorine
12
23 43
42 38
71 11
3.8 98
0.31
1 63
5.30
1 68
Chlorine
14
25.73
39 28
71 49
19 10
0 33
1.99
S 22
1 83
Chlorine
16
28.08
37 26
73.00
17.12
0.29
2.47
4.91
1.80
Chlorine
18
29.87
38.56
72.07
16.44
0.22
2.60
5.22
1 90
Chlorine
20
31 66
39 31
72.50
14 69
0.32
3.04
5.49
1 86
Chlorine
22
33 32
36 73
70 17
11 16
0.34
3.91
5 46
1 SS
Chlorine
24
34 98
37 06
65 55
8 62
0 79
4 51
5.69
1 30
Chlorine
26
36 64
37.55
65 46
5.66
0.83
5.13
5 90
0 92
Chromium(III)
8
1084
1889
2438
583
3.42
4 10
S.40
0.61
Chromium(III)
10
1232
2013
2655
647
3 78
4.56
5.97
0 73
Chromium(III)
12
1383
1951
2875
683
4 11
5 12
6.49
0 79
Chromium(1II)
14
1568
2020
3095
671
4 41
5.50
6 73
0.83
Chromiuin(III)
16
1744
1934
3316
473
4.70
6.12
6.99
0 58
Chromium(VI)
8
0.79
4,19
21.35
4.00
15.92
20.15
26.12
2.37
Chroiniuin(VI)
10
0.89
7.64
56.69
0 99
3.44
18.17
27.43
5 38
Chromium(VI)
12
1 40
11.56
70 88
11 25
3.74
15.84
32.17
6.29
Chrornium(VI)
14
3.91
17.6-1
83.04
13.59
4.02
15.02
29.45
6.69
Chromium(VI)
16
6.39
23.80
97.19
17.94
4.28
13.07
29.78
S.57
45
-------
TABLE 5. (CONTINUED)
2,4 -Dimethylphenol
2 • 4 -Din ethy1pheno1
9 1594 1844 2088 175 4.54 5 17 6.09 0.45
10 1345 1924 2366 295 4.34 5.06 6 38 0.70
Chromiwn(VI)
18
8 26
28 37
114.62
15.45
4 52
11 36
32.01
5 39
C1 n iuui(VI)
20
16 74
31.58
126.40
12.60
4 75
10.00
29 59
4.93
Chromium(VI)
22
25 08
34 35
72.27
7 25
4 97
9.45
20.32
4.58
Chromium(VI)
24
28 75
37 50
52.67
5.15
5 19
8.20
18.14
4.65
Chroeiium(VI)
26
32 96
41.14
46.39
3.57
5.39
6.87
15.36
4.07
Copper
Copper
8
10
2 79
6 19
9 88
11.47
12.72
15.45
2 31
2.00
3.14
3.39
6.21
6.10
13.50
0.07
1.72
0.96
Copper
12
6.52
11.80
17 47
2.61
2.97
6.28
9.25
1.11
Copper
14
8.70
13.16
17.93
2.57
3 54
6.03
8.62
1 13
Copper
16
8.40
14.15
19.16
2 81
3.77
6.24
8 86
1.17
Copper
18
10 79
14 33
21.64
2.74
3.59
6.03
9.24
1.40
Copper
20
11 31
15 55
23 32
3.05
3.77
5.95
9.80
1.40
Copper
22
12 08
16.01
22 05
2 79
3.94
5.64
9.40
1.26
Copper
24
12 87
16.73
23.49
2.73
4.57
5.87
9.54
1.14
Copper
26
13 15
17 43
29.06
3 17
4.27
5.76
10.81
1.26
Copper
28
14.62
17 34
25.73
2.48
4 92
5.75
10 11
1.04
Copper
30
15 82
17.86
25 85
2.55
4 58
5 62
7 62
0.70
copper
32
16.26
18.22
26 55
2.01
5.25
5 81
8 91
0 81
Copper
34
17.08
18.72
30.08
2 87
4.87
5.78
8.49
0.64
Copper
36
17 68
18.91
29.08
2.31
5.00
5.85
8.61
0.57
Copper
38
18 27
18 83
26.54
1.49
5.71
5.84
6.88
0.32
Copper
40
18.87
19.38
23 40
1.38
5.85
5.95
7.06
0.28
Cyanide
Cyanide
8
10
37 84
55 81
57 36
70.21
90 59
97 21
13 03
13.79
1.44
1 09
2 91
1 91
4 81
2.97
0 80
0.55
Cyanide
12
62 42
69 85
99 51
9 61
1.19
1.87
2.74
0.49
Cyanide
14
66.74
70 94
88 13
5.02
1.28
1.75
2.36
0.36
Mercury
8
0 41
0.77
1. 50
0 25
7.83
12 32
15 30
1.68
Mercury
10
0 67
1.15
1 91
0.30
8 65
12.33
15.86
2.01
Mercury
12
0.97
1.61
2.20
0.34
7.66
12.22
16.64
2.32
Mercury
14
1 28
2.08
2.53
0.35
8 22
11.84
17.87
2 35
Mercury
16
1.66
2.53
3.09
0.32
8.75
11.58
17.83
2.29
Mercury
18
2 02
3.03
3.69
0.29
9.26
11.67
18.85
2.12
Mercury
20
2.82
3.44
4 33
0.27
9.73
11.62
18.92
2.02
46
-------
TABLE 5. (CONTINUED)
Mercury
22
3 43
3 89
4 55
0.26
10.18
11.74
17.36
1.72
1.66
Mercury
24
4 04
4.34
5 22
0 28
10.62
11.62
17.18
1.11
Mercury
26
4 64
4 72
5.47
0.20
11.03
11.52
16.04
Methyl parathion
8
0 10
0 34
1.18
0 49
7.13
9 89
13.67
2 98
Methyl parathion
10
0 14
0 43
1 95
0.64
4.51
10.05
15.11
3.16
Methyl parathion
12
0.18
0 49
2.18
0.73
4.90
10.13
16 42
3 31
Methyl parathion
14
0 23
0 54
2.41
0 84
5.26
10 51
17.64
3.58
Methyl parathion
16
0.30
0 60
2.64
0.85
5.60
11 23
18 29
3.55
Methyl parathion
18
0 40
0 55
2.89
0.69
5.93
12 43
17.11
Methyl parathion
20
0.50
0 57
2.62
0 53
6 23
13.29
17.99
2.23
Nickel
8
707
832
960
57
4 89
6 83
7.67
0 47
Nickel
10
865
1002
1100
60
5 41
6 90
7 90
0.70
N Ickel
12
1023
1161
1256
51
5 88
6 89
8.27
0.60
Nickel
14
1250
1315
1421
38
6.32
6 97
8 21
0 63
Nickel
16
1386
1451
1593
32
6.73
7.05
8 47
0.54
Parathion
8
0 00
0.11
0.60
0.17
3.61
10.22
25.23
5.99
Parathion
10
0 02
0.26
0 68
0 16
2.00
7.31
21.71
4.01
Parathion
12
0 16
0.36
0 75
0.14
2.18
6 44
13.53
2.13
Parathion
14
0 21
0 40
0 81
0 13
2.34
5 72
13.03
1.90
Parathion
16
0 21
0 43
0 98
0.12
2.49
5 55
14 72
2 17
Parathion
18
0.32
0 48
0.87
0 13
2.63
5.70
9 95
1 75
Parathion
20
0 34
0 51
0 93
0 11
2 77
4 97
9 39
1.77
Parathion
22
0 40
0 51
0 95
0 07
2 89
4 68
9 23
1 84
Parathion
24
0 47
0 53
0 96
0 07
3.02
4 22
9 54
1 59
Parathion
26
0 49
0.54
0.65
0 03
3.14
4 13
8.S9
1.51.
Parathion
28
0 53
0.55
0 67
0.03
3.25
3.74
6.17
1.14
Pentachiorophenol
8
1 72
17 59
56 39
11 07
0 25
3 04
10 46
2 20
Pentachloropheno].
10
2 38
13 54
55 64
11 91
0 21
3 61
10 40
3 03
Pentachioropheflol
12
2 96
12 75
34 65
10 26
1.14
4 12
10 94
Pentachlorophenol
14
3.62
14 93
56 45
10.82
0 23.
3 96
11.80
3.34
3.88
Pe ntachloropheflol
16
4 25
12.85
57.47
11.69
0 48
4 69
12.45
Pentachiorophenol
18
5 00
11.99
43,06
11.43
1 38
5.37
12.44
Pentachlorophenol
20
5.70
11 29
37.25
10 59
1.50
6.10
13.83
4.20
PentachioropheflOl
22
6.45
11.48
35.15
10 00
1.52
6 29
13.69
4.29
Penta hloropheno1
24
7 14
12.07
32.08
9.93
1.58
6.48
13.58
4.53
47
-------
TABLE 5. (CONTINuED)
Pentachiorophenol 26 7 90 10,79 30.81 8 50 1 64 7.98 14.83 4.14
Peritach lorophenol 28 8 69 10.94 31.23 7.52 1 70 8.60 13.34 4.08
Pentach loropheno l 30 9 52 10 32 31.63 4.84 1 76 10.95 13.57 2.76
Pentachiorophenol 32 10 40 10.75 28.19 3.05 1.98 12.19 13 46 1.90
Phenol 8 2889 5193. 6500 1101 2.78 4.33 5.67 0.59
Phenol 10 3294 5752 7424 1307 3 07 4.42 6.15 0.73
Phenol 12 3834 5978 7901 1450 3.34 4.49 6.58 0.78
Phenol 14 4322 6170 8472 1510 3.59 4.67 6 95 0.83
Phenol 16 4788 6388 9067 1443 3.82 4.71 6.90 0.77
Phenol ie 5281 ‘ 6455 9548 1301 4.04 4.86 7 44 0.79
Phenol 20 5505 6722 10070 1226 4.24 5.02 7.11 0.80
Phenol 22 6107 7056 10632 1222 4 44 4 97 6.86 0.75
Phenol 24 6447 7233 10612 1101 4 63 4 87 7 15 0.59
Phenol 26 6809 7314 11133 891 4.81 5.03 6 66 0 56
Se lenjum(IV) 8 50 272 1276 357 1 50 6 46 13.18 2 90
Selenium(IV) 10 89 308 1322 356 1 65 6.50 12 95 2.71
Selenium(IV) 12 3.36 316 1366 329 1.80 6.66 11.64 2.09
Se lenjum(IV) 14 164 335 1407 232 1.93 7 13 11 70 1.58
Se lenium(IV) 16 195 338 627 135 5.15 7.57 12.47 1.21
Selenium(IV) 18 260 376 1485 226 2 17 7.69 11 10 1.39
Selen ium(Iv) 20 320 345 713 43 5.73 8.65 10.18 0.35
Seleniuni(vI) 8 5 83 37 03 138.52 53.80 9.61 13.36 20 66 3 29
Se lenjum(VI) 10 18 12 31 19 173 78 64 35 10.62 14.07 17.49 2.11
Silver 8 0 75 1 40 10 92 1 32 1.50 4 99 7 87 1 25
Silver to 1 04 1 70 6.66 0.76 1 30 4.19 7.18 1 31
Silver 12, 1 46 1 87 3.52 0 37 1.42 3 71 7.03 1 13
Silver 14 1 65 2 05 3.20 0 26 1 52 3 47 6 01 0 97
Silver 16 1 80 2.18 3 27 0.23 1 62 3 25 5.27 0 70
1.2.4-Trichlorobenzene 8 668 1156 2014 424 0 49 1.70 325 0.89
1,2,4-Trjchlorobenzene 10 721 1017 2037 364 0 53 2.20 3.59 0 89
1 ,2 ,4-Trichloroberizene 12 791 930 1724 262 1.06 2.67 3.64 0.70
2 ,4.5-Trichlorophenol 8 172.4 181.3 187.5 ‘7.4 2 57 2.81 3.21 0.31
48
-------
TABLE 5. (CONTINUED)
ZinC 8 16.74 52.70 146 85 40.05 5 99 9.67 15.96 2 39
Zinc 10 31 39 80 25 169.93 52.42 6 31 9 01 16.97 2 10
Zinc 12 40 19 88 17 192.89 58.74 6.71 9.57 15.06 1.91
Zinc 14 49 OS 103 57 217.22 67.31 7.17 9.89 15.54 2.04
Zinc 16 58 93 113.94 243.11 67.57 7.63 10.41 13.65 1 8S
Zinc 18 67 39 113.53 270.49 70.10 8.07 11.10 14 13 1.86
Zinc 20 76 92 123.83 300.43 68.75 8.48 11.42 14.05 2.00
Zinc 22 96 77 126.44 328.69 70.47 8 88 11.98 14.70 1 85
Zinc 24 97 05 133 39 357.43 68.11 9 26 11.77 15.17 1 90
ZinC 26 108 79 143 64 391.05 68.66 9 62 12 21 15 70 1.97
Zinc 28 121 71 147.35 390.93 52.28 9.97 12 42 16 19 1 82
Zinc 30 134 20 152 30 422.28 44.72 10 36 12.41 16.64 1 69
Zinc 32 148 45 161 66 285 62 36 35 11 42 12 28 15 77 1 44
Zinc 34 161 42 163 85 308 24 16.89 11.76 12.20 16.02 0.80
FAV = Final Acute Value
$ Slope factor.
GMAVS = Genus Mean Acute Values
49
-------
Table 6. Dependence of FAV and S on Number of Saltwater GMAVS
Chemical Number
of
____________________ GMAVa
Final Acute Value
Mm. Mean Max. Std
Dev
S
Miri. Mean Max. Std
13ev.
Antimony(ItI)
Antimony(III)
Ch lorpyrieos
Chiorpyri fos
8 1980 2719 3407 430 1.59 2.22 4 17 0.96
10 2945 3133 3547 12]. 1.75 1.83 2.42 0 20
213 2.76 6 82 10 85 1 51
88 5 74 7.27 8.95 0.85
8 0.02 0 02 0.02 0.00 4 13 4.13 4.13 0.00
10 0.03 0 03 0.03 0 00 4.57 4.57 4.57 0.00
Ammonia
8
0 26
0 50
0 91
0.15
0.25
1.80
4 90
0 84
Ammonia
10
0 33
0 51
0 75
0.13
0.28
1.73
4 04
0.76
Ammonia
12
0.36
0.48
0.75
0.10
0.31
1.87
3.40
0 63
Ammonia
14
0.40
0.48
0.76
0.08
0.33
1.99
3.18
0.47
Ammonia
16
0.44
0.47
0.76
0.06
0.35
2.16
2.62
0.33
Arsenic(III)
8
50
155
1041
Arsenic(III)
10
104
153
360
Cadmium
8
8
63
2507
390
0 90
1.96
14 63
3.42
Cadmium
10
12
55
1246
180
2 47
8 87
16.18
3.16
Cadmium
12
14
55
824
123
2 41
9.47
15 84
2.95
Cadmium
14
20
62
1174
142
3.17
9 04
16 13
2.91
Cadmium
Cadmium
16
18
- 24
34
56
66
233
314
34
35
3.87
4.10
9 22
9.34
15 46
17 50
2 37
2 33
Cadmium
20
34
76
275
43
3 71
8.71
16 38
2.47
Cadmium
22
50
75
329
40
3 57
8 63
14 65
2.15
Cadmium
24
56
77
179
22
4.69
8.51
14.24
1 69
Cadmium
26
63
80
203
29
4.87
8.35
13 79
1 48
Cadmium
28
12
91
140
14
7 63
8 35
11.60
1 08
Cadmium
30
78
82
116
10
7 89
8.24
12.00
0.78
Cadmium
32
84
85
123
8
8.14
8.19
9 55
0 22
Ch lori.ne
8
9 69
15 30
20 23
2 30
2 11
4.37
6 06
0 89
Chlorine
10
12 41
18 28
21 54
2 31
2 33
3 96
6.86
1.01
Chlorine
12
16 05
19 84
23.50
2 35
2 53
3 91
5.89
0 88
Ch1orine
14
18 06
21 39
25.17
189
2 72
3 64
5.41
0 72
Chlorine
16
20 03
22 62
25.51
1 61
2 89
3 56
5.35
0.66
Chlorine
18
21.63
23.75
26.97
1.36
3.06
3 55
5.42
0.69
Chlorine
20
23 33
24.89
27 66
1.10
3.22
3.47
4.56
0.50
50
-------
TABLE 6. (CONTINUED)
2, 4 -Dimethyiphenol
8 416 532 940 161 5 87 6.39 7.58 0 51
8 0 01 0.02 0 08 0 02 3.74 6.31 9.28 1.12
10 0.02 0 03 0 06 0 01 4.14 5 21 8.13 1 32
Chrornium(V!)
8
515
1598
2695
597
1.05
3.03
8.54
3. 92
Chromium(VI)
10
1139
2011
2770
443
0.71
2.36
6.98
1 14
Chromium(VI)
12
1460
2250
2869
433
0.77
1 92
4.01
0 76
Chroinium(VI)
14
1660
2189
2949
423
0.83
2.00
4.50
0.72
Chromium(VI)
16
1904
2231
2879
405
0.88
2.01
3.73
0.72
Chromium(VI)
18
1984
2226
2995
352
0.93
2.06
3.34
0.66
ChromiumiVi)
20
2116
2163
2943
179
0.98
2.45
2.88
0 36
Copper
8
2 09
10.17
56.88
10.96
1.68
5.10
11.15
2.26
Copper
10
2 67
7 88
29 11
8 81
1.38
5.63
11.51
2.62
Copper
12
3 60
9 01
34 46
9 90
1 50
5.35
11.31
2.90
Copper
14
4 67
8.15
36 55
8 93
1.61
5 92
11.46
2.77
Copper
Copper
16
18
5 73
6 60
8 33
‘7 76
2’? 97
30 25
8.29
6 14
1.72
1.82
6.31
7.70
10.52
9.86
2 74
2.10
DOT
8
006
009
0.10
002
4.06
5.74
8.78
154
DOT
10
0 07
0 10
0 11
0.01
4.49
5 64
9 55
1.25
DOT
12
0 09
0 12
0 12
0.01
4.88
5 45
9.03
0.83
DOT
14
013
013
0.13
000
5.24
5.50
6.61
0.37
Die ldrin
8
0 3.7
0 42
3 39
0 69
1 62
6.26
9.48
1.63
Die ldrin
10
0 23
0 43
2.78
0.39
2.16
6.59
9.62
1.39
D leldr ln
12
0 32
0,47
2 07
0.22
3 39
6.53
9 19
1.21
Dieldrin
14
0 39
0 51
1.33
0.11
4.02
6 28
9.11
1. 15
Die ldriri
16
0 46
0 58
1 41
0.12
4.28
6.35
8.76
1.15
D le ldrln
18
0 53
0 60
0 85
0 05
5 47
5.89
‘7.76
0.81
Endosul fan
Endosul fan
Endr].n
Endrin
Endrin
Endrin
Endrin
Endrin
8 0.01
10 001
12 0 01
3,4 0.02
16 0 02
18 0.03
0.01
0 02
0 02
0.03
0.03
0 03
0.02
0.00
6 29
9 39
12.24
0 03
0.01
3.05
7.20
11.59
1,97
0 03
0.01
3.32
6.63
11.37
2.33
0 03
0 01
3 56
5.42
10.23
2.39
0 03
0.00
3 79
4.59
10.71
1.85
0.03 0.00 4.01 4.11 5.00 0.29
5].
-------
461 67 0 97 2.16 3.33 0.74
434 56 1.08 2.01 2.53 0 46
2.30
0 52
3 08
12.75
18.91
3 77
2.09
0 42
2.68
13.28
19 01
3.47
2.29
0.25
2.67
14.91
19.27
2 50
0.49
0.08
6.70
15.64
18 34
1.90
0.84
0.10
6.55
16.06
18.15
1.84
4 .52
4 .64
4 75
4.86
4.59
1.49
1 52
1 48
0 97
0 75
TABLE 6. (CONTINUED)
Heptachior
8
0 01
0 03
Heptachlor
10
0 02
0.04
Heptachior
12
0.03
0.04
Heptachior
14
0 04
0 05
Heptachior
16
0 Os
0 06
Lead
8
235
305
Lead
10
269
297
L.indane
8
0 04
0 16
1.05
8 31
16.24
4.56
Lindane
10
0 06
0.16
1 16
9.64
16 83
4.51
Lindane
12
0 11
0 19
1.26
10 76
16 60
4 53
Lindane
14
0 15
0 18
1.36
13.91
16.93
3.15
Lindane
16
0 19
0.21
1.44
15.44
17.06
2 60
Mercury
8
0 50
2 05
2.50
6.85
12 87
2.35
Mercury
io
0 89
2 54
2 16
6.17
12.33
2.48
Mercury
12
0 96
2 69
1 56
5.23
11.27
2.19
Mercury
14
1 63
3 01
1.67
5.44
13.52
2.13
Mercury
16
1 76
3 41
1.18
4.79
12.62
1.95
Mercury
18
2 61
3 69
1.88
4 39
8.17
1.46
Mercury
20
3 00
3 75
1.98
4.25
8.20
1.49
Mercury
22
3 20
3 84
2.07
4.00
7 92
1.20
Mercury
24
3 41
4.01
2.16
3.76
7.95
1 21
Nickel
8
7 55
57 61
240 05
72.12
9 08
14.01
20 26
2 69
Nickel
io
12 42
66.53
311 44
85.69
6.73
13.74
22.21
2.93
Nickel
12
23 62
76.22
404.05
76.50
7.31
13.42
21 29
3 58
Nickel
14
38.82
101.67
410 61.
89 56
7.86
12.33
21 80
3 73
Nickel
16
59 52
120 53
425 38
71.67
8.36
11.38
20 64
3 14
Nickel
18
78 99
136 31
412 85
52 10
8 84
10.98
19.48
3 10
Nickel
20
112.23
150 18
267 15
22.24
9 30
9.79
14 83
1 57
Pentach] .orophenol
8
8 23
25.27
78.09
12.34
1.79
5.06
10.47
1.67
Pentachiorophenol
10
15 37
28.18
55.17
7.48
1.98
4.40
7.49
1.64
Pentachlorophenol
12
17 59
33.96
59.51
7.16
2.15
3.83
7 86
1 48
Pentach loropheno l
14
2S 36
37.60
63.84
4.25
2.31
3.12
7.87
1.11
Pentach loropheno l
16
34.98
39.37
44.14
1.99
2.46
2.79
4.09
0 64
13 17
8.68’
8.37
6.80
9 47
5.72
5 45
6 71
2.60
2 54
1.18
1 04
0 90
1.26
0.52
0.43
0 50
52
-------
TABLE 6. (CONTINUED)
Ph ’ nanthrene
8 11 62 12.72 13.26 0 57 6 50 6 80 7.52 0.34
1, 2 • 4 -Trichlorobenzene
1,2, 4-Tni.chlorobenzene
1, 2, 4 -Trichlorobenzene
1, 2, 4 -Trlchlorobenzene
8 149 239
10 183 276
12 239 292
14 289 309
355 57 2.98
380 55 3 29
406 47 3.58
431 38 3.85
2,4, 5-Trichiorophenol
8 376 409
466
36 0.61 1 34 2 03 0.45
523 903
602 928
635 952
592 887
187 1.05
167 1.16
152 1 26
80 1.35
2.61
2.21
2.13
2.60
7.52
6.50
3.53
3 20
Selenium (IV)
Selenium (IV)
Selenium ( IV)
Selenium(IV)
Silver
Silver
Si. lye r
Silver
Silver
Silver
Toxaphene
Toxaphene
Toxaphene
Toxaphene
Tributyltin
Tributyltin
Tributyltin
Tributyltin
Tributyltin
8 224
10 283
12 504
14 560
8 3 39
10 4 07
12 5 62
14 5.65
16 7 18
18 13 64
8 0 12
10 0 23
12 0 35
14 0.46
8 0 05
10 0.12
12 0.19
14 0.23
16 0 26
507 965
4 06 9.69
3 27 10.73
2 77 10.33
2 44 9 83
2 22 3.46
21 01 4.38
1684 356
17 48 2.91
1809 228
1802 139
17 80 0.94
1 70 0 26
1 39 0.17
1 02 0.15
0 81. 0.12
7 20
9 .17
11 53
13.45
14 01
14 88
0 44
0 .50
0 .50
0.54
0.24
0.27
0.31
0.28
o 28
1 39
1 54
1.67
1.80
1 91
2.02
1.83
2 02
2.20
2 36
3. 20
3 23
1.44
4 4S
4.73
1.74
1 24
0.65
0.37
2.43
2.83
2 53
1.97
1.19
0.S3
1.59
1 20
1.02
0.63
1 95
1.44
1.42
0.97
0.72
0.83
0.63
0.36
0.22
1.29
0 64
1.35
0 59
0.52
0 27
0.15
0.21
0.09
0 06
4 47
4.16
3.76
3.42
5.30
5.38
5 61
6.07
6 31
4.24
4 22
4 20
4.41
11.27
7.72
6.28
4.53
11.44
9.30
7.84
7.84
8.34
6 26
6.56
5 66
5 00
Zinc
8
18.03
128 50
437 94
88.32
1 28
5.02
12.12
2.06
Zinc
10
38 30
148 35
466.06
95 92
1. 83.
4.69
11.07
1.93
Zinc
Zinc
12
14
76.05
99 68
1S7 07
160 69
445.69
515.41
73.31
78.26
1.57
1.65
4.55
4 56
8.37
7.65
1.33
1.15
Zinc
16
116.67
161 11
425.12
57.76
1.76
4.76
8.04
1.12
Zinc
18
124.68
166 90
408.75
54.07
1 90
4.85
7.24
1.04
53
-------
TABLE 6. (CONTINUED)
Zinc 20 132 79 170 55 389 08 48 84 2 00 4 90 8 16 1 02
Zinc 22 141.01 169 81 397 26 34.18 2.09 4 82 7.96 0 85
Zinc 24 149 35 171 89 270 07 24 18 4 2( 4 92 7.84 0.75
Zinc 26 164 72 171.66 208.78 12 48 4 65 4.84 5.94 0.37
FAV Final Acute Value.
S - Slope factor.
GMAV S Genus Mean Acute Values.
54
-------
Table 7. Worst-Case Acute Values
Chemical Mean Sample Size
__________ FAV 1 2 3 4 S 6 7 8
Acenaphthene 189 41000 2090 2040 1720 1700 670 460 240
Acrolein 6.64 5920 510 368 180 93 90 74 61
Aluminum 2644 79900 55500 50000 49800 40000 38200 23000 22000
Ammonia 0.67 22.80 11.41 8.00 7.99 4.94 4 13 2.76 2.62
Antimony(III) 169 25800 25700 25700 25700 25700 21830 18460 500
Arsenic(III) 212 97000 41760 28130 24500 22040 14960 5278 874
Cadmium 3.62 14650 12220 8400 8100 7921 5395 400 133
Chloride 346 13085 9455 6743 6222 4039 3583 2950 2540
Chlorine 40 19 960 710 400 390 258 179 102 84
Chromium(III) 1899 71064 43100 17776 16371 15800 12436 12400 3200
Chromium(VI) 4 18 1870000 176000 170000 140000 135000 69000 40600 3490
Copper 8 89 10242 9300 6200 6058 5869 1990 636 60
Cyanide 57 36 2490 2326 1800 507 432 426 364 350
2,4-Dimethyiphenol 1844 67600 62500 36300 33400 19300 18100 9200 4800
Mercury 0 77 2100 2000 2000 1000 420 180 50 9
Methyl Parathion 0.34 15200 14800 5300 37 33 15 9 7
Nickel 832 43978 40460 30200 21565 15734 14300 13000 2259
Parathion 0 11 10000 5230 3700 2650 2130 48 32 2
Pentachiorophenol 17 59 43921 18150 1485 1141 663 SOS 308 307
Phenanthrene 59.62 1150 1150 490 419 375 234 126 96
Phenol 5191 2250000 122000 100000 100000 67500 51100 25000 11600
Se lenium(IV) 272 203000 53000 42500 35000 31400 12500 4300 3870
Selenium(Vt) 37 03 442000 193000 66000 63000 47000 20000 5300 760
Silver 1 40 3160 560 280 270 241 60 55 53
Thallium 46.85 1460000 229000 120000 2700 2300 2200 1800 100
Tributyltin 0 30 227.40 66.30 10 20 5.50 5.40 3.90 3.70 0.50
1,2,4-Trichioroben 1156 50000 23800 21200 4333 3900 3400 3010 2140
2 ,4,S-Trichlorophe 181 3 3060 2660 1340 1268 672 611 336 260
Zinc 52 70 144928 88964 27678 20200 19037 18400 2694 547
Values for phenanthrene. thallium, and tributyltin are from Table 4; all other values are from
Table 5 for N=8
FAV Final Acute Value.
55
-------
Table 8. Worst-Case FAVFs
Chemical Mean Sample Size
__________ FAV 1 2 3 4 5 6 7 8
Acenaphthene 189 217 11 11 9 9 4 2 1
Acrolein 6.64 891 77 55 27 14 14 11 9
Aluminum 2644 30 21 19 19 15 14 9 8
Anm on ia 0.67 34 17 12 12 7 6 4 4
Antimony(III) 168 153 153 153 153 153 130 110 3
A.rseriic(III) 212 458 197 133 116 104 71 25 4
Cadmium 3,62 4044 3373 2319 2236 2187 1489 111 37
Chloride 346 38 27 19 18 12 10 9 7
Chlorine 40 19 24 18 10 10 6 4 3 2
Chrom ium(ItI) 1889 38 23 9 9 8 7 7 2
Chrom ium(Vt) 4 18 447159 42086 40651 33477 32282 16499 9708 835
Copper 8.88 1154 1048 699 683 661 224 72 7
Cyanide 57.36 43 41 31 9 8 7 6 6
2,4-Dimethyiphenol 1844 37 34 20 18 10 10 5 3
Mercury 0,77 2732 2602 2602 1301 546 234 65 11
Methyl parathion 0 34 44791 43613 15618 109 97 44 27 20
Nickel 832 53 49 36 26 19 17 16 3
Parathion 0 11 88806 46445 32858 23533 18916 426 284 16
Pentachiorophenol 17 59 2497 1032 84 65 38 29 18 17
Phenanthrerie 59 62 19 19 8 7 6 4 2 2
Phenol 5191 433 24 19 19 13 10 5 2
Seleniuin(IV) 272 746 195 156 129 115 46 16 14
Se len ium(VI) 37 03 11936 5212 1782 1701 1269 540 143 21
Silver 1 40 2252 399 200 192 172 43 39 38
Thallium 46 85 31162 4888 2561 58 49 47 38 2
Tributyltin 0 30 778 227 35 19 18 13 13 2
l,2,4-Trichlorobenzene 1156 43 21 18 4 3 3 3 2
2,4,5-Tr lchlorophenol 181.3 17 15 7 7 4 3 2 1
Zinc 52.70 2750 1688 525 383 361 349 51 10
Arith. mean 22184 5295 3471 2219 1969 700 372 38
Geo. mean 541 235 134 77 60 36 19 6
Values for phenanthrene, thallium, and tributyltin are from Table 4, all other values are from
Table 5 for N 8
FAVPs Final Acute Value Factors.
PAV Final Acute Value
56
-------
Table 9. Mthimum Data Requirements Ranked by Acute Value
Chemical Mean
_________________ FAV
189
6.64
2644
0.67
168
212
3 62
346
40.19
1889
4 18
8 88
57 36
1844
o 77
0.34
Nickel 832
Parathion 0 11
Pentachiorophenol 17 59
Phenanthrene 59 62
Phenol 5191
Se lenium(IV) 272
Selenium(VI) 37 03
Silver 1.40
Thallium 46 85
Tributyltin 0.30
1,2. 4-Trichlorobenzefle 1156
2,4,5-TrichiorophertOl 181 3
Zinc 52 70
Dl P6 G5 34
P3 P6 G5 B
P6 G5 B9 310
B F5 F4 B2
B9 04 B F5
P6 B9 85 G5
B2 B9 G5 F2
C 32 A F6
8 37 P3 310
P5 P2 Dl 88
F3 B B9 P2
P3 05 PS C
P6 8 Dl A
G4 Gi E P6
G5 Fl P3 B12
B5 B2 A G5
85 F3 F5 B9
A G4 32 B4
E P6 A G4
Dl B2 F6 G4
F6 Dl G5 G4
04 G5 F6 32
G4 G5 B4 39
F6 B A B2
F3 P6 B9 G5
39 Dl P6 B4
Dl G5 F6 A
B6 Dl Gi B2
36 P2 E G5
B9 A 8 P3
Dl B9 A 32
A Dl 02 8
Dl 88 G5 A
A 82 Dl 01
F3 A Dl B
G4 A B Dl
PS Dl 3 05
G5 A P1 Dl
89 A G5 3
32 A 05 D l.
88 8 A Dl
G5 P3 89 82
B4 B2 A Dl
A 86 3 Dl
F2 B Dl P6
A 05 3 Dl
B F6 F3 Dl
GS 86 Dl B5
A 39 3 01
02 C S A
B3 A B Dl
A F6 Dl B
G5 B9 Dl 04
A Dl B2 B
G4 A 5 01
B B9 82 04
Fl 04 3 A
89 G4 A Dl
Values for phenanthrene. thallium, and tributyltin are from Table 4; all other values are from
Table 5 for N.8
Ranked from 1 highest to 8 lowest See Table 2 for an explanation of the !WR codes.
FAV Final Acute Value.
Rankb
Acenaphthene
Acrolein
Aluminum
Ammonia
Antimony (III)
Arsenic (111)
Cadmium
Chloride
Chlorine
Chroniunt(tXI)
Chromium (VI)
Copper
Cyanide
2, 4-DirnethyipheflOl
Mercury
Methyl parathion
57
-------
Table 10. Worst-Case Acute Values When Daphnids Were Required
Chemical Mean Sample Size
__________ FA’1 1 2 3 4 5 6 7 8
Acenaphthene 189 41000 2090 2040 1720 1700 670 460 240
Acrolein 6.64 93 93 93 93 93 90 74 62.
Aluminum 2644 38200 38200 38200 38200 38200 38200 23000 22000
Ammonia o 67 4 94 4.94 4.94 4.94 4.94 4.13 2 76 2.62
Antimony(III) 168 18460 18460 18460 18460 18460 18460 18460 ‘500
Arsen ic(III) 212 5278 5278 5270 5278 5278 5278 5278 874
Cadmium 3 62 133 133 133 133 133 133 133 133
Chloride 346 3583 3583 3583 3583 3583 3583 2950 2540
Chlorine 40.19 45 45 45 45 45 45 45 45
Chromium(III) 1889 17776 17776 17776 16371 15800 12436 12400 3200
Chromium(VI) 4 18 3490 3490 3490 3490 3490 3490 3490 3490
Copper 8 88 59.64 59 64 59.64 S9 64 59 64 59 64 59.64 59.64
Cyanide 57 36 1800 1800 1800 507 432 426 364 350
2.4-Dimethylpheno 1844 4800 4800 4800 4800 4800 4800 4800 4800
Mercury 0 77 8 77 8.77 8.77 8 77 8.77 8.77 9.77 8.77
Methyl parathion 0 34 9.1 9.1 9 1 9.1 9 1 9.1 9.1 6.9
icke1 832 2259 2259 2259 2259 2259 2259 2259 2259
Parathion 0 11 1.9 1.8 1 8 1.8 1.8 1.8 1.8 1.8
Pentachlorophenol 17.59 308 308 308 308 308 308 308 307
Phenanthrene 59.62 1150 1150 490 419 375 234 126 96
Phenol 5191 109000 109000 100000 100000 67500 51100 25000 11600
Selenium(IV) 272 3870 3870 3870 3870 3870 3870 3870 3870
Selenium(VI) 37 03 5300 5300 5300 5300 5300 5300 5300 760
Silver 1 40 55 55 55 55 55 SS 55 53
Thallium 46 85 2200 2200 2200 2200 2200 2200 1800 100
Tributyltin 0.30 66.3 66.3 10 2 5 5 S 4 3.9 3.7 0.5
1,2.4-Trichlorobe 1156 50000 23800 21200 4333 3900 3400 3010 2140
2,4,5-Trichioroph 181 3 2660 2660 1340 1268 672 611 336 260
Zinc 52 70 547 547 547 547 547 547 547 547
• Values for phenanthrene, thallium, and tributyltin are from Table 4; all other values are from
Table 5 for N=8
FAV Final Acute Value.
58
-------
Table 11. Worst-Case FAVFs When Daphnids Were Required
Chemical Mean Sample Size
_____________________ FAV 1 2 3 ,•j_ 5 6 7 B
Acenaphthene 189 217 11.08 10.82 9.12 9.02 3 55 2 44 1.27
Acrolein 6.64 14.00 14.00 14.00 14 00 14.00 13 55 11.14 9.18
Aluminum 2644 14 4S 14 45 14 45 14.45 14 45 14.45 8.70 8 32
Ammonia 0.67 ‘7 37 7 37 7.37 7 37 7.37 6.16 4.12 3.91
Aiitimony(III) 168 110 110 110 110 110 110 110 2.97
Arsenic(III) 212 24 91 24 91 24.91 24 91 24 91 24 91 24.91 4 12
Cadmium 3.62 36.68 36.68 36 68 36 68 36.68 36 68 36.68 36 68
Chloride 346 10 35 10 35 10 35 10.35 10.35 10.35 8 52 7 33
Chlorine 40 19 1 12 1,12 1 12 1.12 1.12 1.12 1.12 1.12
Chromium(III) 1889 9 41 9 41 9.41 8 67 8 37 6.58 6.57 1.69
Chrom ium(VI) 4.18 835 835 835 835 835 835 835 835
Copper 8 98 6 72 6 72 6 72 6 72 6 72 6 72 6.72 6.72
Cyanide 57 36 31.38 31.38 31 38 8 84 7 53 7.43 6 35 6 10
2,4-Dimethy lpheno l 1844 2 60 2 60 2 60 2 60 2.60 2.60 2.60 2.60
Mercury 0 77 11 41 11 41 11.41 11.41 11.41 11 41 11.41 11.41
Methyl parathion 0 34 26 82 26 82 26.82 26 82 26.82 26 92 26.82 20.33
icke1 832 2 72 2 72 2.72 2.72 2 72 2.72 2 72 2 72
Parathion 0.11 15.98 15.98 15.98 15.98 15.98 15.98 15.99 15.98
Pentachiorophenol 17.59 17 53 17 53 17.53 17.53 17 53 17.53 17.53 3.7.44
Phenanthrene 59 62 19.29 19.29 8 22 7.03 6 29 3 92 2.11 1.61
Phenol 5191 21 00 21 00 19 26 19.26 13 00 9 84 4.82 2.23
Se lenium(IV) 272 14 23 14 23 14.23 14 23 14 23 14 23 14 23 14 23
Selen ium(VI) 37 03 143 143 143 143 143 143 143 20.52
Silver 1.40 39 19 39 19 39.19 39.19 39 19 39 19 39 19 37.76
Thallium 46 85 46 96 46 96 46 96 46 96 46 96 46.96 38.42 2 13
Tributyltin 0.30 227 227 34 92 18.83 18.49 13 35 12 67 1.71
1,2,4-Trichlorobenzen 1156 43 24 20 58 18 33 3 75 3 37 2 94 2.60 1.85
2 ,4 ,5-TrxchlOrophenol 181 3 14 67 14 67 7 39 7.00 3 71 3.37 1 85 1.43
Zinc 52.70 10 37 10 37 10 37 10.37 10 37 10 37 10 37 10.37
Arith mean 68 08 60 18 52 78 50.80 50 37 49 66 48.55 37 53
Geo. mean 22 01 19.36 17 08 14.91 14.17 12.91 11.19 6.30
Values for phenanthrene. thallium, and tributyltin are from Table 4; all other values are from
Table 5 for H=8.
FAVFs Final Acute Value Factors.
FP.V - Final Acute Value.
59
-------
Table 12. Worst-Case Acute Values When Sairnonids Were Required
Chemical Mean Sample Size
__________ FAV 1 2 3 4 5 6 7 8
Acenaphthene 189 670 670 670 670 670 670 460 240
Acrolein 6 64 74 74 74 74 74 74 74 61
Aluminum 2644 40000 40000 40000 40000 40000 38200 23000 22000
Ammonia 0 67 2 62 2 62 2.62 2 62 2.62 2 62 2.62 2.62
Antimony(III) 168 25700 25700 25700 25700 25700 21830 18460 500
Arsenic(III) 212 14960 14960 14960 14960 14960 14960 5278 874
Cadmium 3 62 5395 5395 S395 5395 5395 5395 400 133
Chloride 346 6743 6743 6743 6222 4039 3583 2950 2540
Chlorine 40.19 179 179 179 179 179 179 102 84
Cbromium(III) 1889 12436 12436 12436 12436 12436 12436 12400 3200
Chromium(VI) 4.18 69000 69000 69000 69000 69000 69000 40600 3490
Copper 8 88 636 636 636 636 636 636 636 60
Cyanide 57 36 507 507 507 507 432 426 364 350
2.4-Dimethyiphenol 1844 9200 9200 9200 9200 9200 9200 9200 4800
Mercury 0.77 420 420 420 420 420 180 50 8.77
Methyl parathion 0 34 5300 5300 5300 37 1 33 15.1 9 1 6 9
Nickel 832 15734 15734 15734 15734 15734 14300 13000 2259
Parathion 0.11 10000 5230 3700 2650 2130 48 32 1 8
Pentachlorophenol 17.59 1485 1485 1485 1141 663 505 308 307
Phenanthrene 59 62 375 375 375 375 375 234 126 96
Phenol 5191 11600 11600 11600 11600 11600 11600 11600 11600
Selenium(IV) 272 12500 12500 12500 12500 12500 12500 4300 3870
Selenium(VI) 37.03 47000 47000 41000 47000 47000 20000 5300 760
Silver 1 40 280 280 280 270 241 60 55 53
Thallium 46.85 2300 2300 2300 2300 2300 2200 1800 100
Tributyltin 0 30 3.9 3.9 3.9 3.9 3.9 3 9 3.7 0.5
1.2,4-Trichlorobenzen 1156 4333 4333 4333 4333 3900 3400 3010 2140
2,4 ,5-Trichiorophenol 181. 3 260 260 260 260 260 260 260 260
Zinc 52 70 2694 2694 2694 2694 2694 2694 2694 547
Values for phenanthrene, thallium, and tributyltin are from Table 4; all other values are from
Table S for N 8
FAV — Final Acute Value.
60
-------
Table 13. Worst-Case FAVFs When Sairnonids Were Required
Chemical Mean Sample Size
___________ FAV 1 2 3 4 . S 6 7 8
Acenaphthene 189 3 55 3.55 3.55 3.55 3.55 3.55 2 44 1.27
Acrolein 6 64 11 14 11.14 11 14 11.14 11 14 11. 14 11 14 9.3.8
Aluminum 2644 15 13 15.13 15.13 15.13 15 13 14.45 8.70 8.32
Ammonia 0 67 3 91 3.91 3.91 3.91 3 91 3 91 3.91 3.91
Antimony(III) 168 153 153 153 153 153 130 110 2.97
Arsenic(III) 212 70 59 70.59 70 59 70.59 70 59 70 59 24 91 4.12
Cadmium 3 62 1489 1489 1489 1489 1489 1489 111 36.68
Chloride 346 19 47 19 47 19.47 17.97 11 66 10 35 8.52 7 33
Chlorine 40 19 4 45 4 45 4 45 4 45 4 45 4.45 2 54 2.09
Chromium(III) 1889 6 58 6 58 6 58 6 58 6 58 6.58 6 57 1 69
Chromium(VI) 4.18 16499 16499 16499 16499 16499 16499 9708 835
Copper 8 89 71 61 71.61 71 61 71 61 71 61 71.61 71.61 6 72
Cyanide 57 36 8 84 8 84 8 84 8 84 7 53 7.43 6 35 6 10
2,4-Dimethylphenol 1844 4 99 4 99 4.99 4 99 4.99 4.99 4 99 2.60
Mercury 0 77 546 546 546 546 546 234 65 04 11 41
Methyl parathion 0 34 15618 15618 15618 109.3 97.24 44.50 26.82 20.33
Nickel 832 18.92 18.92 18.92 18.92 18 92 17 19 15 63 2.72
Parathion 0 11 88806 46445 32858 23533 18916 426 284 15.98
Pentachloropheriol 17.59 84.41 84.41 84.41 64.89 37.67 28.70 17.53 17 44
Phenanthrerie 59.62 6.29 6.29 6.29 6 29 6 29 3 92 2.11 1.61
Phenol 5191 2 23 2 23 2.23 2.23 2 23 2.23 2 23 2 23
Selen ium(IV) 272 45 95 45 95 45 95 45 95 45 95 45 95 15 81 14 23
Seleniurn(VI) 37 03 1269 1269 1269 1269 1269 540 143 20.52
Silver 1 40 200 200 200 192 172 42 75 39 19 37.76
Thallium / 46 85 49 09 49 09 49 09 49 09 49 09 46 96 38.42 2.13
Tributyltin 0.30 13.35 13 35 13 3S 13 35 13 35 13 35 12.67 1 71
1,2,4-Trichlorobenzefl 1156 3.75 3 75 3 75 3 75 3 37 2.94 2.60 1 85
2 .4,S-Tri.chloropheflOl 181 3 1 43 1 43 3. 43 1 43 1 43 1 43 1.43 1.43
Zinc 52.70 51 13 51 13 51.13 51 13 51 13 51 13 51 13 10.37
Arith mean 4313 2852 2384 1526 1365 684 372 38
Geo mean 51 33 50 20 49.60 40.79 38.48 28.18 18.38 6 43
values for phenanthrene. thallium, and tributyltin are from Talde 4, all other valuem are from
Table 5 for N 8
FAVFS — Final Acute Value Factors.
FAV • Final Acute Value.
61
-------
Table 14. Worst-Case Acute Values When Both Daphnids and
Salmonids Were Required
Chemical Mean Sample Size
_____________________ FAV .j _ 2 3 4 5 6 7 8
Acenaphthene 189 670 670 670 670 670 460 240
Acrolein 6 64 74 74 74 74 74 74 61
Aluminum 2644 38200 38200 38200 38200 38200 23000 22000
An onia 067 2 62 2 62 2.62 2 62 2.62 2.62 2 62
Antimony(IXI) 168 18460 18460 18460 18460 18460 18460 500
Arsenic(III) 212 5278 5278 5278 5278 5278 5278 874
Cadmium 3 62 133 133 133 133 133 133 133
Chloride 346 3583 3583 3583 3583 3583 2950 2540
Chlorine 40 19 45 45 45 45 45 45 45
Chrornium(ItI) 1889 12436 12436 12436 12436 12436 12400 3200
Chromjuin(VI) 4 18 3490 3490 3490 3490 3490 3490 3490
Copper 8 88 59 64 59.64 59 64 59 64 59.64 59 64 59.64
Cyanide 57 36 507 507 507 432 426 364 350
2 ,4-Dimethy lpheno l 1844 4800 4800 4800 4800 4800 4800 4800
Merci.iry 0 77 8 77 8 77 8.77 8 77 8.77 8.77 8.77
Methyl parathlon 0 34 9.1 9.1 9 1 9 1 9.1 9.1 6.9
Nickel 832 2259 2259 2259 2259 2259 2259 2259
Parathion 0 11 1 9 1.9 1 8 1.8 1.8 1.8 1.8
Pentachloropheno l 17 S9 308 308 308 308 308 308 307
Phenanthrene 59 62 375 375 375 375 234 126 96
Phenol 5191 11600 11600 11600 11600 11600 11600 11600
Selenium(IV) 272 3870 3870 3870 3870 3870 3870 3870
Selenium(vI) 37.03 5300 5300 5300 5300 5300 5300 760
Silver 1 40 55 55 55 5 55 55 53
Thallium 46 85 2200 2200 2200 2200 2200 1800 100
Tributyltin 0 30 3 9 3 9 3.9 3.9 3.9 3.7 0.5
l,2,4-Trichlorobenzene 1156 4333 4333 4333 3900 3400 303.0 2140
2,4.5-Trichlorophenol 3.83. 3 260 260 260 260 260 260 260
Zinc 52 70 547 547 547 547 547 547 547
Values for phenanthrene, thallium, and tributyltin are from Table 4; all other values are from
Table 5 for N-B.
b Not applicable, two values required.
FAV Final Acute Value.
62
-------
Table 15. Worst-Case FAVFs. When Both Daphnids and Salmonids Were
Required
Chemical Mean Sample Size
____________________ FAV 2 3 4 5 6 7 B
Aceriaphthene 189 3 55 3.55 3.55 3 55 3.55 2.44 1 27
Acrolein 6.64 11.14 11.14 1]. 14 11.14 11.14 11.14 9.18
Aluminum 2644 14.45 14.45 14 45 14.45 14 45 8.70 8.32
Ammonia 0.67 3.91 3.91 3.91 3.91 3.9]. 3.91 3.91
Antimony(III) 168 110 110 110 110 110 110 2 97
A.rsenic(III) 212 24.91 24.91 24.91 24.91 24.91 24.91 4.12
Cadmium 3 62 36.68 36 68 36 68 36 68 36.68 36 68 36 68
Chloride 346 10.35 10.35 10 35 10 35 10.35 8 52 7.33
Chlorine 40 19 1 12 1 12 1.12 1.12 1 12 1.12 1 12
Chromium(IIX) 1889 6 58 6 58 6.58 6.58 6 58 6.57 1 69
Chronu .um(VI) 4.18 835 835 835 835 835 835 835
Copper 8.88 .6 72 6.72 6.72 6.72 6.72 6.72 6.72
Cyanide 57.36 8,84 8,84 8 84 7.53 7.43 635 6.10
2,4-Dimethyiphenol 1844 2 60 2.60 2 60 ‘ 2.60 2.60 2.60 2.60
Mercury 0 77 11.41 11 41 11.41 11.41 11.41 11 4 11 41
Methyl parathion 0.34 26 82 26 82 26 82 26.82 26.82 26.82 20.33
Nickel 832 2.72 2 72 2.72 2.72 2 72 2.72 2.72
Parathion 0 11 15.98 15 98 15.98 15.98 15.98 15.98 15 98
Pentachlorophenol 17.59 17.53 17 53 17.53 17.53 17.53 17.53 17.44
Phenanthrene 59 62 6 29 6.29 6 29 6 29 3.92 2.11 1.61
Phenol 5191 2.23 2.23 2.23 2.23 2.23 2.23 2.23
Seleniu m(IV) 272 14 23 14 23 14.23 14.23 14 23 14.23 14.23
Se len ium(VI) 37.03 143 143 143 143 143 143 20 52
Silver 1 40 39 19 39 19 39.19 39 19 39 19 39 19 37 76
Thalilum 46 85 46 96 46.96 46.96 46 96 46 96 38.42 2.13
Tributyltin 0.30 13 35 13 35 13 35 13.35 13 35 12.67 1.7].
1,2 ,4 -Trichlorobenzene 1156 3.75 3 75 3.75 3.37 2.94 2.60 1 85
2,4.S-Trichlorophenol 181 3 1 43 1 43 1 43 1 43 1 43 1 43 1.43
Zinc 52.70 10 37 10.37 10 37 10 37 10.37 10.37 10 37
Arith. mean 49.33 49.33 49 33 49.27 49 17 48 44 37 53
Geo. mean 12 01 12.01 12 01 11.90 11.65 10.78 6 30
Values for phenanthrene, thallium, and tributyltin are from Table 4; all other values are from
Table 5 for N 8.
Not applicable; two value8 required.
FAVFS - Final Acute Value Factors.
FAV Final Acute Value
63
-------
Table 16. Summary of Geometric Mean Worst-Case FAVFS
Sample Size
1 2 3 4 5
Any family 541 235 134 77 60 36 19 6
Daphnid required 22 19 17 15 14 13 11 6
Salmonid required Si 50 50 41 38 28 18 6
Both daphnid and - - - 12 12 12 12 12 11 6
salmonid required
FAVYS Final Acute Value Factors
64
-------
Table 17. Freshwater Summary FAVFS (Version 1)
FAVP8 - Final Acute Value Factors.
6
85
2.2
62,
8.35
4.20
8.35
4.02
5.11
3.07
4.93
3.41
2.91
2.57
3.43
2.60
1
2
3
4
S
6
7
Median of medians
Any family
2.60
2.16
2.16
1 74
Daphnid required
1.87
1 90
1.82
1.82
Salmonid required
2.58
2.20
2.01
1.82
Both daphnid and salmonid
2 28
1 82
1.82
1.77
95th Percentile of medians
Any family 50000
209.4
39 17
26.14
21 70
15.83
15.83
10.56
Daphnid required 23
21.7
21 06
21.00
15 83
13,89
10.56
15 83
Salmonid required 16389
480 0
94.67
21.70
15.83
15.83
10.56
3.0.56
Both daphnid and salinonid
21 7
21.70
21.06
13.89
10.56
10.56
10 56
Median of 95th percentiles
Any family 444 69
37 20
18.62
13.73
12 62
9.05
7 03
4.01
Daphnid required 20 81
17 25
14.61
9.76
9 76
9 22
5 90
4.01
Salmonid required 24 02
19 52
14.39
12.90
11 71
10 58
7 03
3 48
Both daphnid and salmonid
11.10
11.10
9 76
9 14
9.14
5.90
4.01
95th Percentile of 95th percentiles
Any family 100000
19167
10000
209
112
87 16
28.93
Daphnzd required 229
202
202
112
112
94 67
28 93
Salmonid requ .red 19167
19167
15100
278
202
87 16
28.93
Both daphnid and ealmoriid - - - -
202
202
202
112
94 67
28 93
Overall median
Any family
Daphnid required
Salmonid required
Both daphnid and salmonid
Overall 95th percentile
Any family
Daphnid requited
Salmonid required
Both daphnid and sa].monid
40.1
38.4
32 83
23.81
21.70
19 17
13.89
449
112
4000
202
3.33
2 57
3.03
2 57
87.16
26 .14
87 16
16 74
7.98
11 89
10500
75
16389
2 60 2.21
2 28 1 94
257 221
2 16 1.90
6.81
4 .20
6 .22
3 61
475 9
40 1
913 .6
4 20
3 03
4.25
2.88
175 0
36.7
196.3
1 93
1.82
1.87
1.92
19 17
19.17
19.17
1.82
1.77
1.82
1 80
15.00
15 83
15.83
38.45
21 70
38 45
26 14
21.00
28.93
65
-------
FAVFB Final Acute Value Factors.
Table 18. Freshwater Summary FAVFS (Version
2)
Sample
Size
1.
2
3
4
5
6
7
8
Median of medians
Any family i 63
7.03
3 94
2.82
2.08
1.95
1.90
1.89
Daphnid required 3.96
3.19
2 40
2.26
2.13
2.07
1.96
1.93
Daphnid excluded 18.17
11 37
7.06
4 14
3.41
2.65
2.65
1.89
Salmonid required 7 84
5.64
4.00
3.04
2.35
2.03
1.98
1.93
Both daphnid and salmonid
3.19
2.72
2 48
2.23
2.16
2.02
1.91
95th Percentile of medians
Any family 2258
83
58 11
19.09
7 22
6 46
6.07
5 32
Daphnid required 48
33
22 97
22.42
6.46
6.03
5.76
5.05
Daphnid excluded 4101
262
75.26
29.40
20.73
17.63
15 81
5.05
Salmonid required 22451
1149
97.48
59 64
9.75
6.60
5 98
4.99
Both daphnid and salmonid
47
38.25
25 52
8.75
6.38
6 22
5.1.6
Median of 95th percentiles
Any family 426 49
64 49
42 73
27 55
20 07
12.53
7 49
3.04
Daphnid required 21 89
12.95
7.98
6 99
6 09
5 20
4.30
3 00
Daphnid excluded 423.21
84 80
51.14
45.12
38.41
24.18
7.65
3.10
Salmonid required 54 27
38 18
22 30
20.17,
iS 98
13 14
7.82
3.03
Both daphnid and salmonid
7.89
7.14
7.14
6 06
4.99
4.30
3 02
95th Percentile of 95th percentiles
Any family 165272
39614
8665
1287
804
433
88.2
40.71
Daphnid required 336
166
132
107
98
90
68.9
38.97
Daphnid excluded 174114
35860
17630
1601
981
661
193.6
34 02
Salmonid required 133068
100021
25628
2130
877
370
75 3
49.79
Both daphnid and salmonid
101
98
98
98
97
75.3
48.93
Overall median
Any family 14.66
6.20
3.73
2.81
2 40
2.16
1.96
1.86
Daphnid required 4 89
3 23
2.60
2.36
2,18
2.02
1.92
1.86
Oaphnid excluded 17 98
7 66
4.77
3.69
3 05
2.61
2.36
1.85
Salmonid required 10.63
6.00
3.65
2.70
2.37
2.14
1.94
1.86
Both daphnid and salrnonid
2 93
2,47
2.28
2.16
2.02
1.91
1.86
Overall 95th percentile
Any family 8367
1007
190.0
90.1
53 6
33 17
16.94
7.99
Daphnid required 94
58
50.5
41.8
31 0
22.00
13.12
7.26
Daphnid excluded 10024
1680
424.1
175.6
111.2
64.23
39.90
7.51
Salmonid required 20211
1857
216.2
97.9
59.7
33.73
17.18
8.06
Both daphnid and salmonid ----
52
48.2
38.8
29 0
19.90
13.22
8.42
66
-------
Table 19. Saltwater Summary FAVFS (Version 2)
MP includes all species in the family Mysidae
M8 includes all species in the genus Menidia .
FAVFs - Final Acute Value Factors
and all species in the family Penaeidae.
Sample Size
1 2 3 4 S
6
7
8
Median of medians
Any family 20.60
5.94
3.50
2.88
2.41
2.26
2.07
2.06
NP required, MR excluded 3.62
2.88
2.84
2.33
2.11
2.00
1.99
1.97
MR required; NP excluded 20.61
7.47
5 11
2.90
2.53
2.39
2 26
2.01
Both 14? and Ma
3.55
2.68
2.57
2.40
2.10
2.06
1.99
Neither 149 nor MR 30.41
7.16
4.07
3.09
2.93
2.62
95th Percentile of medians
Any family 344.46
76 9
17.99
9.14
4.01
3.66
3.42
3 33.
NP required; MB excluded 31.66
10.9
5 63
3.77
3.51
3.51
3.51
3.51
MB required; NP excluded 190.13
65.6
32.56
30.11
26.55
24.77
22.14
3.80
Both NP and MR ----
20.9
9 90
5.93
4.29
3.31.
3.31
3.31
Neither MP nor MR 425 02
113.7
84.33
68.09
40.53
33.20
Median of 9Sth percentiles
Any family 394 86
120 5
41.43
16 56
11.81
8 19
S.39
3.35
NP required; MB excluded 12 50
11 8
9 52
9 01
6 84
5 45
2.87
2.87
MR required; NP excluded 52 76
43 2
27.17
16.60
11 83
8 56
4.84
2.85
Both NP and MR ----
11.4
9 97
7.76
7 69
7 10
4.75
2.86
Neither MP nor M8 210.84
63.3
34.82
20.81
12 22
6.26
95th Percentile of 95th percentiles
Any family 428705
2557
1425
547.6
198 1
86 3
19.90
8.23
14? required; MR excluded 271
132
95
77.3
71 7
12 5
8.47
8.47
MR required; NP excluded 2515
1735
433
432.2
431 4
385.7
194.01
8.49
Both MP and MB ----
129
109
89.8
73 1
70.4
12.38
8.36
Neither NP nor MB 82424
3153
1337
615.5
385 6
276.7
----
----
Overall median
Any family 16.53
5 10
3.26
2.67
2 44
2 29
2.11
2.04
NP required, 148 excluded 3.85
2 84
2.51
2 32
2.17
2.06
2.04
2.04
MR required, MP excluded 14.81
6.26
4.16
3.17
2 68
2 58
2.43
1.97
Both NP and M8 ----
3.28
2.66
2.52
2 35
2.20
2.09
2.05
Neither NP nor 148 24.06
8.41
4.35
3.28
2.71
2.52
----
“
Overall 95th percentile
Any family 4903
287 7
101.11
63.3
22.61
12.02
6.50
4.42
NP required; MS excluded ‘74
50 7
21.82
13 6
8.82
6.06
4 62
4.57
148 required, NP excluded 314
191.7
138.02
90 6
76 61
62.73
57.20
4.60
Both NP and MR •---
46.8
34.15
22.2
15.01
9.36
6.07
4.47
Neither NP nor MR S419
436.8
174.26
1OS.S
78.23
68.88
67
-------
Table 20. Overall Percentiles for Fresh Water (Version 2)
Sample Size
4 5
Overall
D .,a- ’l1
1
2 3
6 7 8
50th Percentile
Any family
14.66
6.20
3.73
2.81
2.40
2.16
1.96
1.86
Daphnid required
4 89
3.23
2 60
2.36
2 18
2.02
1.92
1.86
Daphnxd excluded
17 98
7.66
4.77
3.69
3.05
2.61.
2.36
1.95
Salmonid required
10 63
6.00
3 65
2.70
2 37
2.14
1.94
1.86
Daphnid and salmonid
2 93
2 47
2 28
2.16
2.02
1.91
1.86
55th Percentile
Any family
19 16
7 87
4 45
3 34
2.65
2.34
2.14
1.97
Daphxu.d required
6 37
3 74
2 91
2 57
2 36
2.20
2.08
1 97
Daphnid excluded
22 60
9 79
6 07
4 45
3 65
3.05
2.63
1.96
Salmonid required
13 12
7.17
4 45
3 20
2.60
2 30
2.13
1.97
Daphnid and salmonid
3 35
2 89
2 53
2.33
2 19
2 09
... 97
60th Percentile
Any family
26 85
10.29
5 67,
3.87
3 06
2.58
2 25
2 15
Daphnid required
7.77
4 51
3 40
2.83
2 56
2.37
2 25
2.15
Daphnid excluded
34 77
12.61
7 78
5 65
4.37
3.65
3.09
2.13
Salmonid required
17 98
9 S7
5 81
3.89
3.01
2.54
2.25
2.14
Daphnid and salmorn.d
4 11
3 34
2 92
2 57
2.36
2 22
2.16
65th Percentile
Any family
40 92
13 04
7 33
4 68
3 58
2.84
2.50
2 25
Daphnid required
10.11
5 74
3 93
3 31
2.91
2.59
2 42
2.26
Daphni.d excluded
52 15
17 48
10 07
7 08
5.54
4.38
3.68
2.25
Salmonid required
29 67
13 04
7 27
4 95
3.59
2.83
2.46
2 25
Daphnld and salmonid
5.43
3 93
3 38
2.92
2 59
2.40
2.25
70th Percentile
Any family
59 72
17 89
9 98
6.18
4 25
3.38
2.68
2.47
Daphnid required
13.75
7.15
4 97
3.87
3.31
2.88
2.65
2.47
Daphnid excluded
79.65
24.75
13.29
9.32
6 82
5.57
4.56
2.47
Salmonid required
50 55
17.34
9.90
6.59
4.38
3.35
2.66
2.44
Daphnid and salmonid
6.90
5 15
3.97
3.38
2.93
2.62
2.47
68
-------
75th Percentile
Any family
107.88
24 69
13 29
8.54
5.59
3.93
3.20
2.65
Daphnid required
15.68
9 03
6.47
4 90
3.87
3.42
3.04
2.65
Daphnid excluded
Salmonid required
134.32
‘ 67.60
41.37
28.23
20 09
13.12
12.77
8.85
8.90
5 98
6,93
3.94
5.80
3.18
2.65
2.65
Daphnid and sairnonid
7.86
6.64
5.09
3.93
3.43
3.05
2.65
80th Percentile
Any family
174 29
44 00
20.29
12.25
7 62
5 21
3.83
3 17
Oaphnid required
20.50
13 12
8 59
6.53
5.04
4.00
3.61
3.17
Daphnid excluded
242 35
64.82
36.19
20 09
12.86
9 16
7.21
3.17
Salmonid required
136 12
S0.39
20 50
12 45
8 12
5.35
3.77
3.13
Daphnid and salmonid
13 12
8.75
7.00
5 24
4 Os
3.63
3 20
85th Percentile
Any family
442 60
75 26
38.97
19 78
11 95
7 15
4.93
3.76
Daphnid required
31 61
18 58
13 12
9.08
7 00
5.64
4.30
3 74
Daphnid excluded
738 52
115.72
59 19
37 48
21 37
14 16
10.08
3 70
Saitnonid required
291 37
84 76
40 55
20 52
12 35
7 69
4.99
3 69
Daphnid and salmonid
16.16
13.12
10 69
7 48
5 53
4.35
3 74
90th Percentile
Any family
1712 54
160 22
68 60
39.44
21.13
13.14
7.10
4 94
Daphnid required
51 01
30 11
22 06
16 29
12 36
8.60
6.24
4 85
Daphnid excluded
2262.17
238 86
115 72
68 79
41.02
25 85
17.35
4 86
Salmonid required
2015 46
175.68
75 26
42 43
21.59
13 06
7 44
4.97
Daphnid and salmonid
25 52
20 50
18.33
12 61
8.75
6 47
4 90
95th Percentile
Any family
8367
1007
190.0
90 1
53.59
33.17
16.94
7 99
Daphnid required
94
58
50 5
41.8
31.02
22 00
13.12
7.26
Daphnid excluded
10024
1680
424 1
175.6
111 17
64 23
39 90
7 51
Salmonid required
20211
1857
216.2
97.9
59 66
33 73
17.18
8 06
Daphnid and saintonid
52
48 2
38.8
29.00
19.90
13 22
8 42
69
-------
Table 21. Effect of Daphnids on Overall Percentiles
N
Table 21a. FAVFS with Daphnid Required
5 4 3 4 5 8 11 13
8 9 8 10 9 8 12
a N = number of chemicals (out of 31) for which the worst-case
FAVF in Table 11 is greater than the 95 th percentile FAVF.
b Daphnids were neither required nor excluded, except they were
required at n=8. Except at n=8, a daphnid might or might not
have been in the subset.
M = number of chemicals (out of 31) for which the worst-case
FAVF in Table 8 is greater than the 95 ” percentile FAVF.
Overall
Percentile
Number of Minimum Data Requirements Satisfied
1 2 _ 4
50
55
60
65
70
75
80
85
90
95
5 6 7 8
4.9
6.4
7.8
10.1
13.8
15 . 7
20.5
31.6
51. 0
93 . 5
3.2
3.7
4.5
5.7
7.2
9.0
13.2
18.6
30.1
57 . 8
2.6
2.9
3.4
3.9
5.0
6.5
8.6
13. 1
22. 1
50.5
2.4
2.6
2.8
3.3
3.9
4.9
6.5
9.1
16.3
41.8
2.2
2.4
2.6
2.8
3.3
3.9
5.0
7.0
12 . 4
31.0
2.0
2.2
2.4
2.6
2.9
3.4
4.0
5.6
8.6
22.0
1.9
2.1
2.3
2.4
2.7
3.0
3.6
4.3
6.2
13.1
1.9
2.0
2.2
2.3
2.5
2.7
3.2
3.7
4.9
7.3
Table 21b. FAVFs
with Daphnid Not Requiredb
Overall Number
Percentile
of Minimum
Data
Recxuiremerits
Satisfied
1
18.0
2 ..L.
7..7
.j.
5
6
7
8
22.6
9.8
34.8
12.6
52.2
17.5
80.0
24.8
3.7
4.5
5.7
7.1
9.3
3.1
3.7
4.4
5.5
6.8
2.6
3.1
3.7
4.4
5.6
50
4.8
2.4
1.9
55
6.].
2.6
2.0
60
7.8
3.1
2.1
65
10.1
3.7
2.3
70
13.3
4.6
2.5
75
134.
41.4
20.1
12.8
8.9
6.9
5.8
2.7
80
242.
64.8
36.2
20.1
12.9
9.2
7.2
3.2
85
739.
116.
59.2
37.5
21.4
14.2
10.1
3.7
90
2262.
239.
116.
68.8
41.0
25.9
17.4
4.9
95
10000.
1680.
424.
176.
111.
MC
5
70
-------
Table 22. FAVFs Based on Log-triangular Distribution
S
PERCENT
1
2
3
SAMPLE
4
SIZE
5
6
7
8
1.0
95
2.63
2.05
1.79
1.65
1.56
1.50
1.44
1.40
1.0
80
2.10
1.69
1.52
1.42
1.35
1.30
1.26
1.23
1.0
50
1.62
1.37
1.26
1.19
1.15
1.11
1.09
1.07
2.5
95
11.2
6.0
4.3
3.5
3.1
2.7
2.5
2.3
2.5
80
6.4
3.7
2.9
2.4
2.].
1.9
1.8
1.7
2.5
50
3.3
2.2
1.8
1.5
1.4
1.3
1.2
1.2
5.0
95
125
36.2
18.5
12.4
9.4
7.5
6.2
5.3
5.0
80
41
13.6
8.2
5.8
4.5
3.7
3.2
2.8
5.0
50
1].
4.9
3.2
2.4
2.0
1.7
1.5
1.4
7.5
95
1411
217
79.6
43.5
28.7
20.6
15.6
12.4
7.5
80
264
50
23.5
14.0
9.6
7.1
5.6
4.6
7.5
50
38
11
5.6
3.7
2.8
2.2
1.9
1.6
10
95
15835
1309
342.6
152.9
88.0
56.4
39.0
28.6
10
80
1692
185
67.2
33.8
20.3
13.7
9.9
7.6
10
50
126
24
10.0
5.8
3.9
2.9
2.3
1.9
FAVFS = Final Acute Value Factors.
S = Slope factor.
71
-------
Table 23.
Final Acute-Chronic Ratios from Source Documents
Chemical Freshwater Saltw er Reference
Aluminum
Arnmoni a
Ainmon ia
Arsenic (III)
Cadmium
1 7
1 8C
3.8
6. 9 d
NAb
13
3.8
9.1
a This value was obtained the lowered FCV
given in Table 3 in the
b NA = Not Available.
C Calculated as a weighted average of the ratios for the range of
pH from 6.5 to 9.0. (See page 95 in the source document.)
d Calculated as two times the geometric mean of the quotients of
the CMCs divided by the CCCs at hardnesses of 50, 100, and 200
mg/L given in the “National Criteria” in the source document.
The purpose of the factor of two is to account for the fact
that the numerator used in this division is the CMC, which is
one-half the FAV.
This value was obtained by dividing the FAV by the value for
brook trout given in Table 3 in the source document.
Chl ordane
Chloride
Chlorine
Chlorpyri fos
Chromium(III)
Chromium (VI)
Copper
Cyanide
Dieldrin
Endosuif an
Endrin
Lead
Lindane
Mercury
Nickel
Parathion
Pentachiorophenol
Selenium
Toxaphene
Zinc
14
7.6
3.3
4.1
1 6 d
2.9
8.6
8.5
3.9
4.0
51
25
3.7
18
10
3.2
8.0
>38e
2.2
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
440/5-86-008
440/5-85-001
44 0/5-88-004
440/5-84-033
440/5-84-032
440/5-80-027
440/5-88-001
440/5-84-030
440/5-86-005
440/5-84-029
440/5-84-029
440/5-84-031
440/5-84-028
440/5-80-019
44 0/5-80-046
44 0/5-80-047
440/5-84-027
44 0/5-80-054
440/5-84-026
440/5-86-004
44 0/5-86-007
440/5-86-009
440/5-87-006
440/5-86-006
440/5-87-003
14
NA
3.3
4.1
NA
43
2.0
2.0
8.5
3.9
4.0
51
NA
3.7
18
NA
3.2
8.0
2.0
2.2
by dividing the FAV by
source document.
72
-------
Table 24. FCVFs Corresponding to Various Percentiles
Percentile
50
55
60
65
70
75
80
85
90
95
96
97
97. 5
98
Freshwater
7.1
8.2
9.4
10 . 8
12 . 6
14 . 9
17 . 9
22.2
29.1
43.4
48.7
56.2
61.3
68.0
Saltwater
4.4
5.2
6.2
7.3
8.8
10. 8
13 . 5
17. 4
24. 2
39.2
45.1
53 . 6
59.5
67.5
73
------- |