Joint distributional analysis

Joint distributional analysis

In this approach, exposure concentrations and effects concentrations are

expressed as distributions. Predicted exposure values are compared to a

distribution of toxicity effects obtained from the literature or from site-specific

information. The overlap in the two distributions can be used to estimate risk

(Suter 1993; USAEWES 1992).

Developing such a model includes the following steps:

a. Define the assessment end point (e.g., the probability of reproductive

failure in a particular fish species) in terms of a test end point (e.g., the

probability of exceeding a relevant effects end point).

b. Obtain the relevant data (i.e., effects data and exposure concentration

data).

c. Calculate the risk that the expected distribution of environmental

concentrations will exceed the distribution of effects concentrations.

This kind of model can be combined with a logistic or binomial model to

evaluate the potential for risk at the population level. However, the appropriate

level of protection must be selected, which is a risk management decision.

Temporal considerations: Potential for recovery of populations

In ecological risk assessments it is desirable to consider how long an

ecosystem may remain degraded after some assault (Suter 1993). Population

levels of some species may recover almost immediately after exposure ceases.

Other species might require longer recovery periods before pre-exposure

population levels are restored. For example, Kennelly (1987) demonstrated that

the removal of adult kelp plants has little effect if it occurs during the time of

year when the plants are reproducing. If, however, plants are removed at other

times of the year, other species may occupy the space and substantially delay the

return of the kelp. Currently, it is not possible to predict the duration of effects

accurately, since the timing, magnitude and order of stresses can cause

unpredictable effects on populations and complicate interpretations of patterns of

competition and predation (Underwood 1989).

Chapter 5 Uncertainty in Tier IV Risk Assessments