|
A record of decision (ROD) was signed in April 1999 for Operable Unit (OU) 1 at Ross Metals. Required activities under the ROD included removal, treatment, and off-site disposal of contaminated soil, slag, and other materials.
In 2002, a portion of the Site (approximately 2 acres) was sampled and remedial activities were completed following a "conventional approach", which involved multiple mobilizations where samples were collected and sent for off-site analyses. Later, site soil was excavated based on those results, and additional samples were collected to confirm the effectiveness of the remediation.
In July 2003, EPA Region 4 requested the assistance of the EPA Brownfields Technical Support Center (BTSC) (http://www.brownfieldstsc.org/) in developing a strategy for site characterization and remediation for the remaining portion of the Site (approximately 3 acres). The BTSC assisted Region 4 in preparing a work plan that incorporated the Triad Approach.
Highlights of the project from a Triad perspective include:
| Site Name | Ross Metals Superfund Site |
| Location | Rossville, TN |
| Site Type |
Former Lead Smelting Facility |
| Project Lead Organization | EPA Region 4 |
| Project Lead Type |
EPA Lead |
| Regulatory Lead Program |
Superfund Remedial |
| Triad Project Status | Field Program Completed |
| Reuse Objective Identified |
Yes |
| Proposed Reuse: |
Recreational, Commercial/Industrial |
From 1978 until June 20, 1992, Ross Metals operated as a secondary lead smelter. Prior to 1978, the property was undeveloped. Comprising approximately 13.7 total acres on a 200 acre parcel, the Site produced specification-alloyed lead that was sold for use in vehicle batteries, lead shot pellets, and sheet lead (radiation shields). The primary material in the recycling process included spent lead acid batteries, with automotive and industrial batteries accounting for approximately 80 percent of the raw material processed. The remaining 20 percent consisted of other lead-bearing materials, such as recycled dross, dust slag, and factory scrap.
From 1979 until December 1988, blast slag that had accumulated as a part of the smelting process was disposed of in an on-site landfill. From 1988 until the facility closed in 1992, Ross Metals was subject to a number of Resource Conservation and Recovery Act (RCRA) investigations. Since that time a number of removal actions have been conducted, and the Site was added to the National Priorities List (NPL) on March 31, 1997.
In 2007, land use restrictions were instituted on the Ross Metals property that restricted the use of groundwater and limited future use of the Site as industrial. No redevelopment plans for the Site have been made.
The DWS field program at Ross Metals allowed for the rapid high-resolution delineation and excavation of lead-contaminated soil to meet the facility residential cleanup standard of 400 milligrams per kilogram (mg/kg). Through a DMA that derived a safety factor based on a 95% confidence level, this cleanup standard was statistically correlated to a field-based action level of 325 mg/kg used for decision-making purposes.
Lead concentrations measured by the FP-XRF method during the high-resolution delineation sampling ranged from 19 mg/kg to over 20,000 mg/kg. Sixty-one of the 181 30 foot x 30 foot grid elements sampled across the Site contained lead concentrations above the field-based 325 mg/kg action level. These grid elements required some level of excavation and removal as delineated through additional focused sampling. Excavation activities were performed to depths of 36 inches below ground surface (bgs), and were verified clean through field-based and off-site sample analyses, as documented in the Remedial Action Report.
A total of 70,621 cubic yards of soil and waste were excavated and disposed during the field program. Final confirmation sampling using FP-XRF and off-site analyses by EPA SW-846 method 6010 established that project goals of removing site soil with lead concentrations above 400 mg/kg (or 325 mg/kg based on the FP-XRF data) were accomplished.
Project objectives included use of HRSC sampling performed under a Triad-based approach to increase data density, limit decision uncertainty, and in a single mobilization, sufficiently characterize the nature and extent of lead to allow simultaneous excavation and disposal of soil with lead concentrations exceeding the facility cleanup action level of 400 mg/kg.
The Triad Approach used at the Ross Metals Site resulted in sound decision logic and verified the ability of FP-XRF to successfully guide and complete a characterization and excavation program for surface soil in a single mobilization.
Although no quantitative estimates of cost or time savings have been developed by the project team, there is general concurrence among the stakeholders that the Triad produced a cheaper, faster, and better remedial action than the conventional approach could have. Cost and time benefits were assessed even after accounting for the additional upfront systematic planning, Work Plan preparation, and the DMA that were performed under the Triad. The increased costs and longer remediation timeframes for the conventional approach were projected to include additional mobilizations to the Site with accompanying sampling and analysis plans. Moreover, because the exclusive use of fixed-base laboratory methods under a conventional approach would have produced higher per sample costs, the project team might have had to base remedial decisions on lower data densities. Lower data densities in combination with longer laboratory turnaround times would have resulted in higher disposal costs, given that contaminated areas of the Site would have been defined at a lower level of resolution, and that downtime or additional mobilizations would have been incurred in waiting for results. The conventional approach would have resulted in higher site uncertainty, therefore, more likely leading to overestimates of excavation areas to reduce the risk of an incorrect disposal decision (that is, leaving contaminated material in place).
An initial work plan called for a conventional approach to site characterization, involving limited grid sampling along with fixed-base laboratory analysis. With BTSC's assistance, the project team revised the sampling approach into a more comprehensive HRSC grid and judgmental sampling program implemented using the Triad Approach.
Systematic project planning (SPP) activities focused on framing the principal study questions and decisions based on data from previous investigations and the DMA. Review of the existing conceptual site model (CSM) for the Site indicated that prior investigations had adequately established contaminants and media of concern, and project objectives for remediation were already established and agreed upon by the governing regulatory agencies in the ROD.
The strategy for uncertainty management included HRSC using FP-XRF as a real-time analytical method to increase the data density applied across the Site, thus improving confidence in excavation decisions. A DWS was also adopted to further increase and refine data density over certain areas of the Site as needed. Key ranges of concentrations, with safety factors, were identified and refined to guide data interpretation and decision-making, and these ranges also became the focus of collaborative data collection using off-site laboratory methods to further increase decision confidence.
BTSC also conducted classical statistical data analysis and used the free Visual Sampling Plan (VSP) software package as a statistical calculator to assess multiple populations, design and refine the sampling approach (based on estimates of possible excavation areas), and identify field-based action levels based on field analytical results.
The principal decision maker for the Site was the Region 4 remedial project manager (RPM). Other stakeholders with input to Site decisions included the BTSC and Region 4's technical support contractor, Tetra Tech EMI. BTSC's support was provided in the areas of chemistry, statistics, geoscience, and engineering. Tetra Tech provided the field investigation team and overall project management and coordination. Additional technical support and training was provided by the FP-XRF vendor, Niton Corporation.
A cost analysis performed in conjunction with the DMA was used to evaluate the effect of various grid sizes on overall project costs. The costs of soil disposal and sampling and analysis activities were plotted against various grid sizes and the analysis determined that a grid size of 30 foot by 30 foot provided the optimal cost-benefit relationship (see link to Case Study below for the cost-benefit plot). Smaller grids tended to have much higher labor and analytical costs, while larger grids had higher excavation and disposal costs based on use of a limited data set to characterize large volumes of soil.
One hundred and fifty-two grids were determined to be present in the remaining 3 acres requiring delineation and excavation at the Site. Each grid was further defined by progressive depth layers of 3 inches or "lifts". Within each grid, a total of 10 short duration (30 second) in situ FP-XRF readings were collected to assess the distribution and concentrations of lead in the grid. Interpretation of the data was supported using a delineation decision logic diagram (see Case Study). The refined field-based action level (325 mg/kg) for lead developed from the DMA and subsequent sampling was used to guide the process.
Based on the results of the initial 10 short duration FP-XRF readings, the grid could be considered clean and moved to the confirmation sampling program, or additional in situ readings were collected to determine the necessity of removal of the entire grid, a portion of the grid, or a hotspot. Any areas where the lead concentration exceeded the field-based action level of 325 mg/kg were then excavated in 3-inch lifts.
Following excavation activities, soil in each grid area was confirmed to be below the field-based action level by subdividing the grid into four 15 foot by 15 foot sub-grids and collecting one long duration (2 minute) in situ FP-XRF reading at a random location within each sub-grid. An aliquot was then collected where the highest lead value was observed. This sample was homogenized and an ex situ FP-XRF analysis conducted and compared to the action level (325 mg/kg). Samples from the location within the grid displaying the highest FP-XRF concentration of lead were collected at a frequency of 20 percent and sent for off-site analysis by EPA SW-846 method 6010 (acid digestion/inductively coupled plasma spectrometry). If a grid was not chosen for the collection of a grab sample for fixed-base laboratory confirmation, the highest 2-minute duration in situ FP-XRF concentration reading from the grid served as the confirmation result.
Decision logic diagrams were incorporated into the revised work plan to allow maximum use of real-time FP-XRF results. The decision logic assisted the project team in assessing "clean" and "dirty" areas using a field-base action level developed and refined during the DMA and subsequent sampling program. The field-base action level was developed using correlations between various FP-XRF readings and corresponding off-site laboratory analyses.
The decision logic was developed to assist the field team in differentiating "no action" areas from areas where excavation was required, and in assessing the dimensions of the excavation. The decision approach focused on prioritizing critical path excavation activities so that the field program could proceed without unnecessary downtime.
FP-XRF analyses were performed in real-time using in situ short (30 second) and long (2 minute) duration readings where the instrument window is placed on the ground surface in direct contact with soil. As part of the confirmation program, aliquots were collected and homogenized for ex situ FP-XRF analysis to compare with results from the highest in situ reading. The ex situ readings were used to provide an additional check on in situ concentrations and confirm that remedial and compliance decisions were being made correctly. Twenty percent of samples collected during the confirmation program were also analyzed for lead using off-site analysis by SW-846 Method 6010.
The DMA study provided crucial information on the variability of results based on sample preparation methods as well as providing the data set necessary to develop preliminary field action levels. The DMA demonstrated that in situ FP-XRF measurements provided sufficient data quality for decision making during characterization, and that ex situ measurements on prepared samples were only necessary during confirmation sampling after excavation, thus saving additional time and costs during the field program. Data collected from the DMA established an initial field-based action level of 314 mg/kg. Through the use of the VSP statistical calculator on a real-time basis, this action level was refined to 325 mg/kg during the sampling and excavation program as additional data were collected.
TQRS pending
BTSC provided support in developing correlations between FP-XRF and fixed-base laboratory SW-846 Method 6010 results. The field-based and fixed-base laboratory methods correlated strongly and together produced a high quality data set that managed sampling as well as analytical uncertainty and was suitable for decision making at the Site.
The project team noted that while blanks and calibrations were performed as quality control (QC) measures for the field test kits and indicated no problems, other QC checks such as field duplicates and spikes were not performed due to limited test kit materials.
BTSC used the sampling design/statistical calculator tool Visual Sampling Plan (VSP) in a novel manner to assist in the sampling design and in developing preliminary field action levels. Statistical evaluations and VSP work products were generated and applied in real-time during the sampling and excavation program; these evaluations and work products are discussed in the Case Study and in the Remedial Action Report.
Fall 2002 (2 acre parcel; traditional approach)
November 2003 (3 acre parcel; Triad Approach)
Not available
404-562-8814To update this profile, contact Cheryl T. Johnson at Johnson.Cheryl@epa.gov or (703) 603-9045.