Ecological Risk Assessment of Diffuse and Local Soil Contamination Using Species Sensitivity Distributions - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 648

Species X Species Y Species A Species B

NOEC EC50 NOEC EC50

Test Concentration

Fig. 14.7 Plotted laboratory toxicity data, collected for two species ( a and b ), show the prin-
ciples of designing toxicity tests and collecting summary data on species sensitivities. ( a )isa
relatively insensitive species (position on the X axis) with a low increment of response at increases
of exposure (slope), and ( b ) is more sensitive species with a high increment of effects at increasing
exposure

Input data consist of sensitivity data, collected usually in ecotoxicity tests of
various species. To obtain such data, species are usually exposed to multiple concen-
trations of the contaminant of interest, under artificial test conditions (e.g., artificial
soil, or a well-known test soil under controlled conditions). After a period of expo-
sure, either the survival of the exposed organisms is scored, or the performance of
vital characteristics is quantified (e.g., with parameters describing reproduction or
growth performance).
Summary sensitivity data are derived from such tests, per species. Collected data
may be critical (chosen) concentrations, like No Observed Effects Concentrations
(NOECs, the highest concentration without significant effects), Median Effective
Concentrations (e.g., EC50, the test concentration that affects 50% of the test pop-
ulation), or other test endpoints. The principles of designing toxicity tests and
collecting summary data on species sensitivities are shown in Fig. 14.7 .
Original tests may be necessary, for example, when data are lacking for contam-
inants for which regulatory action is deemed to be needed (e.g., due to frequent use
or occurrence of the contaminant at contaminated sites).

14.8.1.2 Pre-Treatment of Input Data
Pre-treatment of input data can include practices such as sub-selections based
on quality assurance criteria (e.g., sub-selection of data collected under Good
Laboratory Practice), selection of data with consistent exposure durations, or con-
version of data to standard units, or calculating the median sensitivity when more
than one effect data value is present for a specific species and endpoint. When
SSDs are formally adopted in regulations, it is common that there are Guidance
Documents stating how data pre-treatment should proceed. For example, in the
U.S. EPA's guidance for water quality criteria, multiple data for a species are com-
bined by taking the geometric mean. When there is no prescribed guidance, data
pre-treatment should match the context of the case.