Environmental Engineering Reference
In-Depth Information
The first track is a phenomenological study on soil quality standards in relation
to observed occurrence or absence of effects in contaminated field sites. When con-
sidering Fig.
14.2
, the onset of responses might be compared to the position of the
SSD-predicted HC5, to check whether the HC5 is indeed a “safe environmental
concentration”. In various studies, the HC5 based on NOECs (and often assumed
to represent a concentration at which ecosystem structure is protected) appeared to
be positioned in the “non-response left tail” of biotic response curves. Note that, by
this criterion, any way to obtain a low estimate of a soil quality standard would also
yield a valid result (a soil quality standard that is higher than the ecosystem no-effect
level and thus sufficiently protective). For example, dividing the lowest NOEC by
an uncertainty factor of 10 or 100 (often an alternative approach to derive standards)
is also valid in this sense. As noted below (Section
14.13.3
) it is of practical impor-
tance to derive standards that are
not
unnecessarily low (e.g., by choosing a very
high safety factor or the lower confidence limit of a chosen HCp-level), since that
would imply (public) concerns where they are not realistic, and possibly extremely
low cost effectiveness of Risk Management.
In some legislation, there is also a trigger value for considering remediation, and
that trigger is derived from an SSD. In the Netherlands, for example, the HC50 of
an SSD-NOEC model is used to select cases requiring further study, as a step-up
to potential remediation. Similar to the HC5-validation approach, predicted HC50s
(the concentration at which 50% of the species would be exposed beyond their
NOEC) appear to be positioned either in the neutral-response part of the response
curve of field assemblages, or in the steep down-going part of this curve, but (so far)
never in the right tail where all biota suffer (Posthuma et al.
1998
). In other words:
such HC50s appear to signal the presence (or near-presence) of adverse ecological
impacts in the field.
Examples of this kind of study are summarized in Posthuma et al. (
2002b
). More
recent studies have signaled some observed field responses at the level of an HC5
(see DEFRA (
2005
) and Frampton et al. (
2006
)). In particular, effects have been
found in field tests of pesticides. But note here the possibly specific impact of the
clear exposure scenarios (short, peak exposures after deliberate use) and the possible
focus and intensity of the effect studies that can be made at such moments.
The second track of validation studies has been recently developed, and considers
the association between predicted impacts and observed species loss over the whole
concentration range of contaminants in the field, based on very large biomonitoring
data sets. Because of the occurrence of mixtures in such databases, the SSD-based
predictions also incorporate probable mixture impacts (the toxic pressure now con-
cerns the multi-substance PAF, msPAF; see Section
14.10.6
). Figure
14.5
presents
two studied cases on the degree of association between predicted (X, acute toxic
pressure, based on SSD-EC50s) and observed species loss for fish species in Ohio
(U.S.) rivers (De Zwart et al.
2006
; Posthuma and De Zwart
2006
), and for fresh-
water invertebrate species in England and Wales (De Zwart et al.
2008a
). Given the
results of both studies, five things are noteworthy:
Search WWH ::
Custom Search