Ecological Risk Assessment of Diffuse and Local Soil Contamination Using Species Sensitivity Distributions - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 632

Denneman ( 1989 ) for Europe (especially the Netherlands). Glenn Suter coined the
term Species Sensitivity Distributions for this type of modeling in OECD ( 1992 ).
The SSD method was based on the insight:

We can see that the species sensitivity (LC50 or LD50) [that is: measures of sensitivity]
distributes itself in a rather consistent way for most chemicals. The distribution resembles a
lognormal one. Thus, each species we test is not representative of any other species, but is
one estimate of the general species sensitivity (Mount 1982, as cited in Suter ( 2002 )).

The SSD-modeling concept is based on the fact that species appear to differ in
sensitivity to the same contaminant (like the species in Fig. 14.2 ). SSDs do not
explain these differences, and do not attempt to simulate soil ecology . They just
represent sensitivity differences to toxicants in a useful way for environmental pro-
tection and appraisal (see below). The same holds for the functional traits, in regard
to Ecosystem Services, resulting from the activities of all the species. For these dis-
tributions the term Functional Sensitivity Distributions has been coined (FSDs) by
Posthuma et al. ( 2002b ). In the remainder of this chapter, we use “SSD” to refer to
both SSDs and FSDs.

14.3 SSD Modeling and Practical Needs

14.3.1 Basics of Distribution Modeling as an Assessment Approach

SSDs are distributions of data, in this case of laboratory data from ecotoxicity stud-
ies. These are studies in which chosen species are tested in controlled exposures to a
suite of concentrations of the chosen compound. Such studies yield estimates of the
sensitivity of different species to that compound, and the compilation of such data
provides the input for SSD modeling. SSD models are thus descriptive, statistical
models of species sensitivity data which were collected in controlled exposures .
Distributions of data, such as sensitivity data, are at the heart of statistics. Anyone
with a basic scientific training knows concepts like the mean and the variance of a
data set. SSDs and FSDs thus describe the variability of sensitivities across species
or functions using extremely basic statistical concepts .
A very useful and illustrative introduction to the statistics and interpretation
of species sensitivities in relation to contaminant management is provided by
EUFRAM ( 2006 ). This report provides a basic explanation and visualization of dis-
tribution statistics in terms of the distribution of the variation in height of people
attending a meeting (see Fig. 14.3 ), as follows (adapted for this chapter):

• panel A shows the raw height data summarized as a bar diagram and as a Normal

probability density distribution derived from these data;

• panel B and Panel C show the same data, now as Cumulative Distribution

Function and an Exceedance Function, illustrating how to estimate the fraction of
the visitors' population with heights below (B) or above (C) for any given value.