Geosciences –
Geology and Hydrogeology

From 1953 until 1993, rail tank cars unloaded TCE at the southeast corner of the C-400 Cleaning Facility where it was stored, transferred and used for degreasing in heated liquid baths, high-pressure sprayers and vapor degreasing units.

Losses of TCE DNAPL to the shallow subsurface included: 1) leakage along TCE transfer system piping; 2) discharge through process wastewater; 3) spills; and 4) cleaning process vapor releases. TCE-laden water from the C-400 cleaning baths was discharged through wastewater piping for transfer to the PGDP wastewater treatment facility. The discharged TCE dissolved the oakum joints used to seal the pipes of the wastewater transfer system and allowed TCE-laden water to enter the subsurface directly.

TCE is a dense non-aqueous phase liquid (DNAPL) and is denser than water. TCE’s density causes it to sink through porous soil, aquifer materials, and groundwater. As TCE sinks through soil and aquifer media it leaves TCE DNAPL in the interstitial pore spaces where it remains as it is slowly dissolved. When sufficient quantities of DNAPL are released, their downward movement will continue until they encounter impermeable materials such as clay. When impermeable material is encountered TCE DNAPL will pool. Once pooled, the DNAPL will remain a long-term source to groundwater contamination.

From 1953 to 1976, the PGDP feed plant reprocessed spent nuclear fuel rods containing uranium. Reprocessing resulted in the introduction of technetium-99 (⁹⁹Tc or Tc-99), a man-made fission radioisotope, and other radioactive materials not associated with naturally occurring uranium. Deposition in process equipment, piping and transfer equipment as well as transfer and storage of ⁹⁹Tc-bearing liquids introduced ⁹⁹Tc to PGDP waste and wastewater streams.

Technetium-99 is a unique radionuclide in environmental settings because it easily dissolves in water where it forms the pertechnetate ion,TcO₄. The pertechnetate ion is relatively unreactive with aquifer materials and very mobile in groundwater.

The southeast and northwest corners of the C-400 Cleaning Building have been identified as major groundwater contamination sources related to the PGDP Northwest Plume which contains TCE and ⁹⁹Tc. Sources of the Southwest Plume (TCE and ⁹⁹Tc) include uranium burial grounds, a RCRA-closure drummed uranium above-ground burial facility, an oil landfarm, and the C-720 maintenance facility. The Southwest Plume is limited to DOE property. On-site sources for the Northeast Plume have not been clearly defined.

The Northeast and Northwest TCE Plumes and the Northwest ⁹⁹Tc Plume are the largest in the DOE complex and amongst the largest documented groundwater plumes of their kind in the world. The depth of the contaminated aquifer, the geochemistry of the aquifer and contaminants, and aquifer materials that range from flowing fine sands to cobbles all pose challenges to geoscientists and engineers in their efforts to characterize and remediate PGDP groundwater contamination.

Initial project activities included extensive field reconnaissance in the vicinity of the PGDP, consultation with Kentucky and Illinois State Geological Surveys to correlate the distribution of geologic units in southern Illinois with geologic units in the vicinity of the PGDP, and carbon dating analysis of (Upper) Continental Deposit material overlying the Regional Gravel Aquifer. Approximately 400 lithologic logs were gathered, compiled in Rockware software and used to produce stratigraphic and hydro-stratigraphic surfaces and cross-sections for the PGDP and vicinity.

Aquifer surfaces interpolated from database records were utilized in the 2008 PGDP Groundwater Flow Model Update, 2010/2012 PGDP pump and treat optimization activities, and the 2016 PGDP Flow Model Update. Newly acquired lithologic log data is included in semi-annual updates to the database. In CY 2020 a 28,000-record 10th update to the database was submitted to the PGDP.

Lithologic log data has been acquired from PGDP site & CERCLA project documents, the Kentucky Geological Survey water well data repository, U.S. Army Corps of Engineers, TVA Shawnee Steam Plant engineering documents, geologic quadrangle maps and the Illinois Geological Survey. Approximately 1700 digital lithologic log files were transferred from KRCEE to PEGASIS in 2014.

Stratigraphic surface and hydrogeologic unit surface maps, isopach maps and 3D renderings have been produced from Lithostratigraphic Database records utilizing a variety of geotechnical software packages including Surfer, ARCMap, ArcScene, Earth Vision Software (EVS), Leapfrog, Groundwater Modeling System (GMS) and Rockware.

Lithostratigraphic Database Project Team participants included PGDP personnel, University of Kentucky faculty & staff, contractor subject matter experts and Earth and Environmental Science graduate students. The Kentucky Geological Survey enlisted graduate students to conduct a significant data input during the summer of 2009.Approximately 1,000 lithologic logs were updated and/or digitally entered by Biosystems and Agricultural Engineering students during summer 2012.

The first Project Team task undertaken by Department of Civil Engineering faculty and post-doctoral staff was an independent review of historical PGDP modeling efforts. The review encompassed the initial 1994 PGDP Groundwater Flow Model (MODFLOW) and model updates through calendar year 2000 which included development and implementation of a flow and transport model (MODFLOWT). (1998 PGDP Groundwater Flow Model)

Next, the Project Team ran the 1998 MODFLOW and MODFLOWT model simulations and evaluated results. Sensitivity analyses were then conducted on key model parameters including hydraulic conductivity and TCE degradation rates. Sensitivity analysis of TCE plume extents over a range of TCE intrinsic biodegradation half-lives, from 5 years to 26.5 years are shown in accompanying figures.

A summary of Project Team findings is provided in the GW Modeling Efforts FFA Summary Presentation below.

The purpose of an ANN model for the PGDP is to forecast TCE concentrations as accurately as the MODFLOWT model so that it can be incorporated with an optimization technique and integrated into a management model.

An optimization model requires numerous evaluations of the objective function, and this is not feasible with a MODFLOWT model that can take hours (to days) for one simulation. A properly trained ANN model could give results of the objective function in seconds.

Three distinct groundwater modeling tasks were accomplished for the Project. Results are discussed in the reports below.

“Within the funds provided the Department shall undertake a study of the potential purchase of property or options to purchase property that is located above the plume of contaminated groundwater near the facility site. The study shall evaluate the adequate protection of human health and environment from exposure to contaminated groundwater and consider whether such purchase, when taking into account the cost of remediation, long-term surveillance, and maintenance, is in the best interest of taxpayers.”

The Land Study required a groundwater modeling evaluation for a complete range of groundwater remedial alternatives identified in PGDP decision (regulatory program) documents. Five alternatives were identified, No Action and four remedial scenarios (Table 3.4.1 below). For each alternative, the temporal as well as maximum extent of plume impacts was modeled over a 100-year period.

A KWRRI & UK-Civil Engineering Project Team applied experience in site groundwater modeling (see Groundwater Modeling 1) to conduct MODFLOW and MODFLOWT simulations of each remedial action scenario using the 1998 PGDP Flow and Transport Models (Bechtel-Jacobs, DOE, 1998*).

Further project groundwater modeling information is available in the text and Appendix A of the Property Acquisition Study Final Report.**

*DOE (U.S. Department of Energy) 1998. Ground Water Flow Model Recalibration and Transport Model Construction at the PGDP, Paducah, Kentucky, DOE/OR/07-1742&DO, United States Department of Energy, Paducah, KY, June 1998.

**The findings of the PGDP Property Acquisition Study are presented in greater detail with PGDP Future Use, Public Outreach & Education Projects

Preliminary site and seep characterization and assessment work was conducted independently (journal article: Fryar et al., 2000) and as part of thesis (Mukherjee, 2003) and dissertation work (LaSage, 2004) with funding from the University of Kentucky Water Resources Research Institute (KWRRI) Federal Facilities Oversight Unit (FFOU).

Subsequent publications, thesis and dissertation work built upon earlier works.

ITM outputs for several time frames through 2018 were compared to 1) transport outputs from the 2008 update of PGDP groundwater flow model and 2) published interpretations of PGDP trichloroethene (TCE) plume extents based on field measurements

Recommendations regarding the development and refinement of a transport model to supplement the 2016 update of the PGDP sitewide groundwater flow model are included in the report and presentation below.

The scenario selection process, developed by SRNL, was utilized to evaluate PGDP groundwater conditions in order to identify the physical, chemical and biological fate and transport processes actively occurring and impacting TCE concentrations in PGDP groundwater.

The attenuation scenario selection process considered PGDP aquifer physical properties, groundwater chemistry, TCE concentrations, TCE degradation products, and TCE source/plume geometry.

The results of the scenario selection process indicated that aerobic co-metabolic TCE degradation was likely to be occurring in the Regional Gravel Aquifer.