Methods for Deriving Pesticide Aquatic Life Criteria - Reviews of Environmental Contamination and Toxicology - page 71

7.2 Basic Methodologies

Two basic criteria derivation methodologies are in use, or proposed for use,
throughout the world. The aim of both methods is to extrapolate values from available
toxicity data to ones that will protect the environment. The first of these two
methods is the AF method. It involves multiplying the lowest value of a set of toxicity
data by a factor to arrive at a criterion. The second method is the statistical extrapo-
lation method. It involves the use of one of several similar SSD techniques to
determine the criterion. Some countries exclusively use one of these methods, and
others use a combination of the methods, depending on data availability. In a 1993
review of the statistical procedure of Van Straalen and Denneman (1989), and the
AF method used by USEPA (1984b; now in Nabholz 1991), Calabrese and Baldwin
(1993) concluded that these two methods produced the same results and no strong
argument exists for selecting one method over the other.
One advantage of the SSD approach over the AF method is that it provides for
deriving a criterion with a known level of confidence. Unfortunately, the SSD
method utilized by the USEPA (1985) does not allow the inclusions of confidence
levels. However, other SSD methods do provide for quantification of confidence
levels (RIVM 2001: ANZECC and ARMCANZ 2000).
France, Germany, Spain, the UK, and Canada utilize only the AF method for
derivation of water quality criteria (Lepper 2002; BMU 2001; Zabel and Cole 1999;
CCME 1999). Others, including Australia/New Zealand, the Netherlands, USEPA,
the EU, Denmark, and OECD utilize a combination of the SSD and AF methods
(ANZECC and ARMCANZ 2000; RIVM 2001; USEPA 1985; ECB 2003; Bro-
Rasmussen et al. 1994; Samsoe-Petersen and Pedersen 1995; OECD 1995).
In France, Spain, Germany, and the UK criteria are derived by multiplying (or
dividing) the lowest toxicity value from a minimum data set by a factor. One criterion
is derived that is supposed to protect against long-term exposures (Lepper 2002;
Irmer et al. 1995; BMU 2001). In France, AFs of 1-1000 are applied to single-
species toxicity values. For derivation of low-level criteria, acute data may be used
with an AF of 1, but high-level criteria are derived by applying an AF of 10 to
chronic NOEC data, or 1000 to acute data (Lepper 2002). In Spain, data corre-
sponding to the most sensitive organism are used in criteria derivation. LC 50 /EC 50
values are multiplied by a safety factor of 0.01 and chronic NOEC values by a factor
of 0.1. Further safety factors are applied to account for lack of relevant species,
persistence or bioaccumulative potential and genotoxic potential (Lepper 2002).
In the UK, the lowest relevant and reliable adverse effect concentration in the data
set is multiplied by a safety factor. An MAC, to protect from acute toxicity, is derived
from acute data, with a factor of 2-10 applied to the lowest available acute toxicity
value. An AA concentration to protect from chronic toxicity may be derived from either
acute or chronic data, or from acceptable field data, with application of appropriate
factors (from 1 to 100) to the lowest available toxicity value (Zabel and Cole 1999).
The Canadian methodology (CCME 1999) uses chronic LOEC values to derive
criteria. If there is an adequate data set, then the lowest LOEC is divided by a factor