Methods for Deriving Pesticide Aquatic Life Criteria - Reviews of Environmental Contamination and Toxicology - page 40

The quantity of ecotoxicological effects data required for criteria derivation is
substantially different for different countries, and depends on what derivation
methodology is used, what type of criterion is being developed (i.e., values to be
used in standard setting vs advisory values), and what level of uncertainty is acceptable
in the criterion. Criteria are derived by extrapolating from available effects data to
real-world situations. The two basic methods for doing these extrapolations are (1)
application of AFs and, (2) statistical extrapolation of SSDs. There is not much
debate about what constitutes appropriate levels of data for the AF method. Factors
are applied according to the types and volume of data available, and many method-
ologies allow for derivation of a numerical guideline value (as opposed to an
enforceable criterion). Such guideline values for a contaminant may be based on as
little as one datum and may be an estimated toxicity value (e.g., calculated from a
QSAR) rather than a measured one. In contrast, for methods that utilize statistical
extrapolation, there is little agreement among methodologies concerning how much
data are needed to produce criteria with the appropriate level of certainty.
High reliability TVs, as defined in the Australia/New Zealand methodology
(ANZECC and ARMCANZ 2000), can be determined either directly from at least
three multispecies chronic NOEC values or from statistical extrapolation using
at least five single-species chronic NOEC values (from five different species).
A moderate reliability TV can be derived from at least five single-species acute
toxicity values, and a low reliability TV can be derived from a single acute or
chronic toxicity datum.
The Dutch methodology (RIVM 2001) requires at least four chronic NOEC
values from species of different taxa for a refined effects assessment; for a pre-
liminary effect assessment, an ERL may be derived from a single LC 50 or QSAR
estimate. Toxicity values, estimated by QSARs, may also be used in statistical
extrapolation models.
The OECD guidelines (1995) present several methods for criteria derivation,
and each has its own data requirements. For statistical extrapolations using the
methods of Aldenberg and Slob (1993) or Wagner and Løkke (1991), at least five
chronic NOECs are required. To derive a final chronic value (FCV) using the
USEPA methodology (USEPA 1985) requires chronic NOEC values for at least
eight animal families, including Salmonidae, a second family in the class
Osteichthyes, a family in the phylum Chordata, a planktonic crustacean, a benthic
crustacean, an insect, a family in a phylum other than Arthropoda or Chordata,
and a family in any order of insects or any phylum not yet represented. Unlike the
USEPA method (1985), the OECD does not allow for derivation of a chronic
criterion by application of an acute-to-chronic ratio (ACR) to a final acute value
(FAV). An environmental concern level (ECL) can be determined by the OECD
AF method (1995) from a single LC 50 value. If no toxicity data are available,
QSARs may be used to estimate toxicity for some classes of chemicals. Such
estimated values may be used to derive MTCs.
For derivation of a FAV, the USEPA (1985) requires acute toxicity data for eight
North American species, each of which must represent different families, as previ-
ously described for the OECD methodology. Despite having fewer than the required