Abstract
A large number of methodologies have been
proposed to support the relative comparison of chemical emissions in
the context of implicit toxicological concern. Nevertheless, the
selection of a methodology typically remains subjective and influenced
by resource availability (time, knowledge and money). To help
practitioners address such issues, the results of three extensive
studies are outlined in this presentation:
- The merits and applicability of five
categories
of methodology are first illustrated with the help of a hierarchical
framework and straightforward case study. The hierarchy is based
on the level of representation of the environmental mechanisms (from
chemical emission, to fate, via exposure to toxicological potency) and
perceived sophistication.
- A comparison is then presented of two of
the more prominent but structurally different methodologies (tiers 4
and 5 in the hierarchy) used in the US. The WMPT facilitates
comparison in terms of key properties using a framework of expert
judgment to reflect levels of concern in terms of Persistence,
Bioaccumulation and Toxicity (PBT). Toxic Equivalency Potentials
(TEPs) account for chemical fate, multi-pathway exposure and toxicity
using a multimedia model structure. Using the same data for 318
organic chemicals and minimizing scenario differences,
a strong relationship and parallel support role is demonstrated to
exist
between these two particular approaches in the context of human health.
- Focusing data collection efforts can
dramatically reduce the time required to compare a large number of
chemicals or emissions using the more sophisticated multimedia
approaches in the hierarchy. Based on model insights and a stated
trade-off between data needs and error, four straightforward guidelines
are presented to help predetermine which
degradation rates (air, water, soil and sediment) are likely to be
pertinent,
hence which can be omitted. The introduced error is generally
less
than an order of magnitude for the 318 chemicals when compared to the
full
human health model predictions, while the data requirements (1272
half-lives)
and associated collection times are drastically reduced.
|
