Reliability Of Sediment Quality Assessment In Puget Sound

To be useful in environmental management, sediment criteria must, at a minimum, reliably identify environmental problems. Two measures of reliability are evaluated using actual field data from 13 urban and nonurban embayments in Puget Sound Sensitivity in detecting environmental problems (i.e., are all biologically impacted sediments identified) Efficiency in screening environmental problems (i.e., are & biologically impacted sediments identified). High overall reliability results from correct prediction of a large percentage of biologically impacted stations (i.e., high sensitivity; few false negatives) and correct prediction of a large percentage of the nonimpacted stations (i.e., high efficiency; few false positives). These measures of reliability were applied to a range of sediment criteria generated by the Equilibrium Partitioning (EP) and Apparent Effects Threshold (AET) approaches. Overall reliability ranged from 44-64 percent for the EP approach and from 42-85 percent for the AET approach, depending on the particular criterion and biological indicator tested. A higher percentage of correct predictions was made using a combination of the two approaches than by either approach alone. Unless cause-effect relationships between contaminants and adverse biological effects can be acceptably established, additional steps are necessary to ensure that remedial actions effectively mitigate predicted problems. Recommended steps based on the principles of sensitivity and efficiency include: 1) screen out sediments for which a lower range of criteria sensitively predict no adverse effects, 2) perform biological testing of sediments for which there is substantial disagreement among predictions in a middle range of available criteria, and 3) efficiently predict problem sediments from chemical data using a higher range of criteria that provides a preponderance of evidence of adverse biological effects. Optional biological testing using a battery of indicators provides a means of verifying sediment predictions for site-specific concerns. Such a management program is currently used in Puget Sound.