ERA practice has tended to emphasize the risk assessment process in the abstract without a
grounding in a decision-making process. For example, due to the peculiarities of Superfund
regulations, the guidance for baseline ERA at contaminated sites in the United States does not
address the consequences of remedial decisions (Sprenger and Charters 1997). This situation is
changing with the realization that the highest-quality assessment is worthless if it does not
address the needs of the decision maker (National Research Council 1994; The Presidential=
Congressional Commission on Risk Assessment and Risk Management 1997). Risk-based
environmental decisions generally fall into three categories: should we permitx(e.g., use of a
new chemical, release of an effluent, or increased harvest of a resource); what should we do
aboutx(e.g., remediate, treat, or restore); should we dox,y,orz(e.g., which pest management
method poses the least risk)?
Probability, the other core concept in ERA, has also been surprisingly neglected. The
probabilities that characterize risks may result from variability or uncertainty (Chapter 5).
Although quantitative methods for analyzing uncertainty and variability in terms of prob-
ability have existed for centuries, most ERAs treat them qualitatively or in nonprobabilistic
terms. This does not mean that uncertainty and variability are ignored or that, as some have
contended, most current risk assessments are not truly assessments of risk. Rather, they are
often dealt with by semiquantitative precautionary practices. That is, conservative assump-
tions and safety factors have been assumed to provide sufficient safety to avoid the need for
a formal probabilistic analysis. However, formal probabilistic analysis of uncertainty is
increasingly common. This is because the semiquantitative practices are subject to criticism
that they are insufficiently precautionary, excessively precautionary, or precautionary to an
undefined degree.
Risk assessment uses science, but is not science in the conventional sense, i.e., it does not
seek to develop new theories or general knowledge. It rather uses scientific knowledge and
tools to generate information that is useful for a specific purpose. In this sense, risk assessors
are like engineers, and in fact much of the practice of ERA has been developed by engineers
(see, e.g., Haimes 1998). However, contrary to some critics, risk assessment is based predom-
inantly on factual information and scientific theory, and is not simply a scientific smoke
screen for policy. Typically, risk assessments and their components are intensely and publicly
reviewed and are often challenged in court. As a result, the use of bad science to justify a
preordained decision is likely to be detected in contentious cases.
1.1 PREDICTIVE VS. RETROSPECTIVE RISK ASSESSMENT
The EPA’s framework and guidelines for ecological risk assessment and the previous edition of
this text distinguish retrospective from predictive risk assessment. This distinction has created
some confusion, because it is nonsensical to speak of risks of events in the past. This text
eliminates that distinction and focuses instead on the decision-supporting function of risk
assessment. Hence, when assessing risks from spills or other past events, we are assessing risks
associated with future consequences of those events. They include ongoing toxic effects, the
spread of toxic levels of contaminants to other areas, loss of habitat due to failed restoration, and
other sequela. Even when performing assessments to set monetary damages for past actions, we
are not assessing risks of past events. For example, during the Exxon Valdez oil spill, a certain
number of sea otters, bald eagles, and other wildlife were killed. The natural resource damage
assessment (Section 1.3.8) did not assess risks to those organisms. Rather, to the extent that the
uncertainties in the damage assessment could be interpreted as risks, they should be interpreted as
risks that the level of monetary damages assessed will be either insufficient or excessive with
respect to the cost of restoration of the populations of those species, making good lost services
of nature, and otherwise remediating ecological and economic injuries.
Glenn Suter/Ecological Risk Assessment L1634_C001 Final Proof page 4 6.10.2006 7:53pm
4 Ecological Risk Assessment