Environmental Engineering Reference
In-Depth Information
14.4.3 The Conceptual Interpretation of SSDs: PAF and PES
As shown in Fig.
14.4
, SSDs can be presented with two equivalent Y-axes. The
statistical interpretation is the same, but the conceptual interpretations differ.
The interpretation of distribution-based Risk Assessments is best explained by
a hypothetical experiment. Assume one has two contaminated samples, for which
the Hazard Potential is estimated by Potentially Affected Fractions of 20 and 80%,
calculated with SSDs constructed from EC50-values. This would imply that expo-
sure of all tested species together in such soils would imply an effect level of 50%
for probably 20 and 80% of these tested species. When assuming that the distribu-
tion of contaminated site species sensitivities resembles the calculated distribution,
the same PAF-values (20 and 80%) are expected for the local species. Note that an
ecologist would suggest in the latter case that 20% and especially 80% are underes-
timates of actual effects, due to possible indirect effects (such as in Fig.
14.2
). For
example, the loss of prey species would indirectly affect predator species, even if
they could withstand the intoxication fully.
We do not know
which
species would be affected nor whether there are additional
indirect effects. However, we have certainly
learned something about the potency of
the contaminated soil to affect test species
and hence
we can interpret the “hazard”
of this soil by extrapolation as a “risk level” for local species too, in the framework
of practical soil management.
In other words: soils can be ranked according to their
potential to affect test species (and possibly local species).
SSDs re-calculate contaminant concentrations in soils (from the realm of chemistry,
expressed in mg/kg
dw
of a contaminant) into a toxic potency (in the realm of biology, in a
dimensionless but conceptually meaningful unit, fraction of species), focusing on the frac-
tions of species likely to be affected. PAF is, in fact, a characteristic of a soil to be hazardous
to its inhabitants. It is
not
an ecological entity.
In the same hypothetical experiment one can also interpret the SSD according
to the concept of
Probability of Effects on a Species
(Fig.
14.4
). In this approach,
the hypothetical experiment proceeds as follows. Imagine the same soils, but now
the assessors are concerned with risks to a particular species, and that species may
be considered to be a random draw from the set of all species, which is estimated
by the distribution of the set of tested species. In this case, the probability that this
species would be affected by the contamination would be 0.2 and 0.8, on a probabil-
ity scale of 0 to 1. Again,
this interpretation tells us something about the potential
of a contaminated soil to affect a randomly chosen species that could reside in that
particular soil
.
The
PES
and
PAF
interpretations can be used interchangeably, but PAF is
generally used in practice as the unit of toxic pressure.
14.4.4 Discussions of SSDs, Assumptions and Interpretation
Since SSDs are statistical models based on minimal ecological knowledge, they
have been discussed intensely since their invention, amongst scientists and with
Search WWH ::
Custom Search