Sediment bioassays

d. Elutriate and suspended-solids tests may artificially alter the availability

of contaminants. For example, clumps of sediment might break apart

during the test that would have remained intact in the field, resulting in

an overestimate of the potential for effects.

Precision of water column bioassay. Because it is easy to quantify

uncertainty due to precision error (Ingersoll et al. 1997) and the utility and

accuracy of the standard elutriate test have been extensively field verified

(USEPA 1996a), the magnitude of uncertainty is likely to be low.

Interpretation of water column bioassay. The magnitude of uncertainty is

ranked "moderate" with at least a factor of 10 due to the extrapolation from acute

to chronic effects (Kenaga 1982). Further uncertainty is introduced due to

possible aeration and disaggregation of dredged material during the test

procedure. Given the way the assay is performed and interpreted, these sources of

uncertainty are more likely to overestimate the potential for effects than

underestimate them.

Extrapolation of bioassay results to species in the field and from acute to

chronic exposures represents the primary quantifiable sources of uncertainty.

Two kinds of models are typically used for effects extrapolation (Suter 1993):

(a) additive and multiplicative factor models (i.e., applying uncertainty factors to

test results), and (b) statistical models (typically linear regression). As in the case

of the factor of 0.01 applied to the water column toxicity test results, the use of

these factors implies a linear relationship between the independent and dependent

variables. Consequently, statistical extrapolation models can also be used in

place of uncertainty factors. These models typically assume that the

toxicological parameter of interest is characterized by a probability density

function, and that all future observations are drawn from this distribution. For

example, this approach has been used to determine USEPA Water Quality

Criteria. The approach assumes that interspecies sensitivity follows a log-

triangular distribution and the lower 5 percent concentration is used as the

concentration affecting the most "sensitive" species (i.e., protection of 95 percent

of the species).

In order to assess causality, toxicity testing must be linked to other models or

toxicity tests specifically designed to identify contaminants or classes of

compounds responsible for observed toxicity (Ingersoll et al. 1997). The water

column toxicity bioassay can be performed and interpreted without knowing

which contaminants are causing the toxic effect. However, it is not possible to

predict in advance of the test which sediments might fail.

Sediment bioassays

Whole-sediment bioassays are designed to determine whether the dredged

material is likely to cause unacceptable adverse effects on benthic organisms.

The limiting permissible concentration (LPC) is defined as "that concentration

which will not cause unreasonable acute or chronic toxicity or sub-lethal adverse

effects based on bioassay results" (40 CFR 227.27). Dredged material does not

meet this condition in the commonly used acute toxicity test if bioassay mortality

(a) is statistically greater than in the reference sediment and (b) exceeds mortality

Chapter 4 Uncertainty in Tiered Evaluation of Dredged Material