Formatted Title:

Using Uncertainty Evaluations to Inform Contaminated Sediment Investigations and Delineations

Background/Objectives:

Sediment at the 10-river mile-long Portland Harbor Superfund Site (PHSS) in the lower Willamette River in Portland, Oregon, is contaminated with polychlorinated biphenyls, polycyclic aromatic hydrocarbons, dioxins/furans, pesticides, and metals. The PHSS has been subdivided into approximately 1-river mile-long remedial design project areas for intensive investigation and design. Investigation includes high spatial density sediment coring to delineate the design footprints in three dimensions. To assess the uncertainty in the delineation, indicator methods are employed to predict the likelihood that an area contains contaminant concentrations greater than the active remediation thresholds.

Approach/Activities:

An uncertainty evaluation and semivariogram analysis using indicator methods was conducted to assess environmental contamination at four fund-lead remedial design project areas within the PHSS. Uncertainty evaluations were performed to evaluate surface concentrations (0 to 1 foot below mudline [bml]) and subsurface concentrations (deeper than 1-foot bml) independently to assist with bounding lateral and vertical contamination data gaps. The selected surface and subsurface sediment concentration data were processed to generate a composite indicator value at each sample location representing an exceedance (1.0) or no exceedance (0.0). The composite indicator values and the subsurface only indicator values were interpolated separately using the natural neighbor algorithm to generate a probability of exceedance field. Semivariograms were generated to evaluate the spatial autocorrelation of the data and to optimize future sample spacing.

Results/Lessons Learned:

Areas with a probability of exceedance greater than 20% and less than 80% were targeted for additional sampling to refine the three-dimensional extent of contamination, which will result in an optimization of the design footprints. Semivariogram analysis provided crucial insights into spatial correlation structures, guiding infill sampling strategies with a recommended maximum sample spacing of 100 to 170 feet. These results emphasize the importance of targeted sampling efforts for accurate contamination delineation and informed remedial design decisions at the four fund-lead sites.

Matthew Carlino

Affiliation

Biography

Refine Results

(Group 2, Poster Board #1) Using Uncertainty Evaluations to Inform Contaminated Sediment Investigations and Delineations