Rapid sediment assessment: Indicator analysis and screening analysis approaches

Efficiently characterizing the distribution of contaminated and toxic sediments in rivers and harbors is usually limited by the expense of conventional chemical and toxicological analyses. Two approaches were developed to address this problem; the indicator analysis approach, used in the ARCS project, and the screening analysis approach, here applied to a sediment assessment project on the Ottawa River (Toledo, Ohio). The indicator analysis approach utilized two suites of analyses; 23 conventional toxicological and chemical analyses performed on a subset of samples, and 11 rapid, inexpensive chemical and toxicological assays performed on many samples, including those analyzed using the conventional analyses. Predictive correlation equations were generated using step-wise linear regression, and these equations were used to calculate values for the conventional analyses for samples on which they were not performed. This approach generated statistically strong predictive equations, as well as a ''weight of evidence'' data set useful for evaluating relative sediment contamination. The equations, however, were very site-specific, and sometimes contained terms which were counter-intuitive, and the approach failed if the data sets contained too many ''non-detect'' or 100% mortality values. The screening analysis approach measured total PCBs by enzyme immunoassay (EIA) and 18 elements by x-ray fluorescence spectroscopy (XRF). These analyses correlated very strongly with gas chromatography (GC) and atomic absorption spectroscopy (AA), respectively, and their production rates and costs were far superior. A low bias was observed in the EIA data, compared to the GC data, possibly due to inefficient EIA extraction of the oily sediments, or to a mismatch between the PCB mixtures in the sediment and used as a calibrator for the EIA. XRF data for Cr, Cu, Mn, Ni, Sb, and Zn exhibited a positive bias compared to AA, while Cd and Pb did not. This was probably due to metal-specific variations in the contribution of mineral matrix-associated metal to the acid-digestible metal quantified by AA. Both EIA and XRF can be performed in the field, to produce near-real time data to guide sampling. Detection limits of the PCB EIA (0.12 mu g/g DW) and of XRF (typically 5 to 15 mu g/g DW) are adequate for most sediment assessment projects. Of the two approaches, screening analyses are recommended for the rapid, cost-effective characterization of contaminated sediments.