Ecological Risk Assessment of Diffuse and Local Soil Contamination Using Species Sensitivity Distributions - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 644

We conclude from the scientific validation studies, that

• an HC5 that is estimated from NOEC values (or similar measures of none or

little effects in test systems) is an estimate that for many known cases indicates
the absence of no or very limited adverse effects on policy-valued ecosystem
characteristics (like biodiversity), so that such SSD-based standards can be used
for deriving protective soil quality standards;

• the estimation of Potentially Affected Fractions, especially for SSDs based on

EC50 values, provides insight into the potential loss of species at contaminated
sites that permits the relatively robust ranking of contaminated sites from low to
high probability of toxic impacts, but levels of effects need not be well predicted
because many other stressors contribute to causing species loss.

We explicitly state that we have rather few examples, and especially few for
soils and sediments in comparison to examples for aquatic systems, to support the
generality of these conclusions. Nonetheless, the incidental studies seem to imply
that increases in SSD-predicted risk levels are related to increased impacts in natural
systems. An SSD-based risk prediction may be not robust in the sense that such an
estimate exactly predicts what will happen to local species (neither which one will
be affected, nor to what extent), but that large differences in (ms)PAF certainly
seem to imply large differences in potential for species impacts or species loss. This
phenomenon implies that SSDs are useful at least for risk ranking.

14.6 SSDs and Ranking of Contaminants or Sites

14.6.1 SSDs and Ranking Contaminants

The two ways of using SSDs in Ecological Risk Assessment were already intro-
duced in Fig. 14.4 . When these two approaches are used to assess contaminants or
contaminated sites, and in view of the validation studies, the key interpretation of
the SSD output is: ranking . Note that this is not peculiar to SSDs. Relative ranking
is more reliable than absolute estimation for all assessment techniques.
When adopting standard guidance protocols, one can derive Hazardous
Concentrations for p percent of the species for all possible contaminants being tested
using the Y

X approach. When this is done for two or more contaminants, two or
more SSDs are generated, and two or more HC5 values result. The output of this
conceptually implies an ecotoxicological “ranking of contaminants”, in the sense
that the lower the HCp-value, the more toxic the contaminant is for the tested sets
of soil organisms .
Some issues arise when using SSDs to rank contaminants in soils. First, the
ranking can be based on the HC5 values, showing which contaminant is the most
toxic at this exposure level and how the other compounds rank. But when another
HCp-level is used, rank orders might change as a consequence of SSD-slope differ-
ences. Second, there is an option to trigger less debate on the ecological meaning of

→