Methods for Deriving Pesticide Aquatic Life Criteria - Reviews of Environmental Contamination and Toxicology - page 76

NOECs are available from at least three species representing three different trophic
levels. Factors of 1-5 are applied to results of SSD extrapolations. For field or
model ecosystem data, the size of the factor is determined on a case-by-case basis.
Acute and chronic safety factors used in the South African methodology (Roux
et al. 1996) are intended to compensate for missing information. They are applied
to compensate for insufficient data being available to assess inter- and intraspecies
variability, or if chronic data are absent. Acute safety factors range from 1 to 100
and depend on completeness of the data set. For example, if the minimum data set
is available, and includes results from more than one test in at least three taxonomic
groups, then a factor of 1 is applied. At the other extreme, if only one acute result
is available, a factor of 100 is applied to the FAV. The chronic safety factors range
from 1 to100 and also depend on completeness of the data set. For example, if the
chronic database contains ACRs or chronic exposure data for at least one species
from three different taxa (including at least one fish), then the factor is 1. If only
acute data are available, then a factor of 1000 is applied to the FAV or the FPV to
arrive at the CEV. In the South African methodology, ACRs are derived by dividing
the geometric mean of available acute values by the geometric mean of chronic
values, where acute and chronic values were obtained in the same test, or in tests
run in similar water dilutions. The same concerns discussed for the USEPA meth-
odology, regarding covariance of SMAVs and ACRs, and for intraspecies ACR
variability, apply in the South African methodology.
Canadian factors range from 10 to 100, but the total factor applied could be
higher if, for example, a measured ACR is higher than 10. A factor of 10 is applied
to chronic data to account for variability in species sensitivity, extrapolation from
laboratory to field, and differences in test endpoints. Higher level factors are
applied to acute data when no chronic data are available, and are used to extrapolate
from acute-to-chronic exposures, or to derive criteria directly, as described earlier.
There is no theoretical basis for any of the AFs used by the various criteria
derivation methodologies. They are all empirically derived numbers. The origin of
generic factors of 10, for each step of uncertainty, is not clear; those methodologies
that explain the reason for the selection of a value of 10 simply state that it is widely
accepted. Measured ACRs seem to have a firmer basis in empirical evidence, but
they are usually derived for a particular chemical and for a particular species and
are then applied to other species or groups of species, which may lead to further
uncertainty in final criteria values.
Different default ACRs are used in different methodologies, when no measured
ACR is available. The Great Lakes guidance uses a value of 18 (USEPA 2003a),
Canada uses either 2 or 10 (CCME 1999), the OECD, the USEPA's OPPT, and
Australia/New Zealand use 10 (OECD 1995; Nabholz 1991; ANZECC and ARMCANZ
2000). Kenaga (1982) reports that ACRs were less than 25 for 86% of 84 chemicals
tested. However, for pesticides, 70% of ACRs were more than 25, with the largest
at 18,100 for propanil. The large percentage of chemicals with ACRs less than 25
resulted from the fact that 93% of industrial organic chemicals fell into that category.
Based on Kenaga's results, the USEPA “Guidelines for Deriving Ambient Aquatic
Advisory Concentrations” (USEPA 1986) use a default ACR of 25 for calculation