Methods for Deriving Pesticide Aquatic Life Criteria - Reviews of Environmental Contamination and Toxicology - page 78

An alternative to the use of AFs is employing statistical extrapolation methods;
such methods use single-species toxicity data to make ecosystem predictions.
These methods, discussed next, still rely on acute-to-chronic extrapolation.
According to Posthuma et al. (2002a), many ecologists and ecotoxicologists have
independently developed methods for predicting ecosystem-level effects using
single-species toxicity data. The USEPA was first to use the SSD method to derive
water quality criteria (Suter 2002). In Europe, Kooijman (1987) developed the concept
of deriving a hazard concentration for sensitive species (HCS; Van Straalen and Van
Leeuwen 2002). According to Van Straalen and Van Leeuwen (2002) Kooijman's
idea was soon refined and modified by Van Straalen and Denneman (1989), Wagner
and Løkke (1991), and Aldenberg and Slob (1993), and Aldenberg and Jaworska
(2000). The most recent version of these SSD techniques appears in the ANZECC
and ARMCANZ criteria derivation methodology (2000). The main difference
among these methodologies is in selection of the shape of the distribution that is
used for extrapolations. Other differences include kinds and quantity of data used in
the extrapolation procedures, level of confidence associated with derived criteria,
and how data are aggregated to construct the distribution. One area in which there is
much agreement among methodologies is in the selection of the 5th percentile as the
cutoff for prediction of no-effect concentrations. Figure 1 provides a general illustra-
tion of the SSD technique, and will be referred to in the following discussion.
1.0
0.75
0.5
0.25
0.05
HC 5 /FAV (5th percentile)
Log LC 50 or NOEC
Fig. 1 Generic illustration of the species sensitivity distribution ( SSD ) technique; HC 5 hazard
concentration potentially harmful to 5% of species, FAV final acute value, NOEC no observable
effect concentration