Accomplishments

Accomplishments

Progress during the first year of our project has been along two complementary lines: (1) field characterization using hydrological and geophysical approaches, and (2) development of dual domain modeling framework and simulations to determine when dual-domain modeling approaches improve plume transport predictions over conventional approaches. These accomplishments are briefly discussed below.

1) Field Acquisition of Characterization Data

Well Installation, Sampling, and CPT Campaign
Three first-round groundwater observation and testing wells have been installed at the study site by SRNL. These wells (POS1, POS2, and POS3) are being used for various downhole or in-situ experiments and tests including seismic and electrical tomography, borehole flow meter measurements, slug testing, as is partially described below. A CPT campaign was undertaken in September 2006 at the study site. CPT groundwater samples were acquired at 10 locations and over several depths; these data were used to define the vertical and horizontal extent of the P-Area TCE plume.

Green dots represent POS wells; concentrations
are from CPT sampling at blue dot locations

Lithologic logs were recorded at three CPT locations, and dual-level piezometers were installed at the same locations. Headspace gas analyses were performed on sediment samples collected from both permeable and relatively impermeable intervals to define the vertical plume extent, and the distribution of contamination with respect to mobile and immobile water regions in a dual-domain conceptual model. The CPT data indicated peak TCE concentrations in the 20-30,000 ppb range in a transmissive zone that is centered at a depth of approximately 90 ft below ground surface. Comparison of lithologic and concentration data suggests that facies seem to be controlling transport in the vicinity of our study site. Based on this initial dataset, three permanent wells were installed in October along the plume center line.

Correlated geophysical well log data

Seismic
Zero-offset and multi-offset P-wave and S-wave vertical seismic profiles (VSPs) were acquired by ESRI-University of South Carolina in two wells previously constructed at the study site for preliminary hydrogeologic and contaminant distribution characterization. These two wells extend through the first aquitard to approximately 30 m below land surface. Data from the VSP surveys is being used to evaluate the vertical acoustic velocity profile at the study site for optimal design of the 2.5D surface seismic survey. The multi-offset and azimuthal VSP data provide detailed information on lateral and vertical stratigraphic heterogeneity, subsurface velocities, and noise and signal attenuation. These data are necessary for design of acquisition parameters for the 2.5D surface seismic survey. Further, the VSP results are being used to generate acoustic and elastic impedance models to assist in mapping hydrofacies.

Results of vertical seismic profile

A preliminary surface ground penetrating radar (GPR) survey was conducted by the University of South Carolina, Department of Geological Sciences at the study site to evaluate the recoverable data quality from this geophysics technique. Several GPR profiles were surveyed: 4 profiles with 50 MHz antennae, 4 profiles with 100 MHz antennae, 1 profile with 200 MHz antennae, and common mid-point GPR surveys with 50 MHz and 100 MHz antennae. Analysis of the resulting GPR data indicated, due to the high clay content in the unsaturated zone, that surface GPR is unlikely to image the subsurface at the P-Area Reactor study site below the water table. However, the viability of cross-hole GPR in characterizing the subsurface environment at the study site is continuing to be evaluated.

Resistivity
A series of surface electrical resistivity surveys has been conducted at the study site by ESRI-USC and the Department of Geological Sciences. Electrical resistivity variations can often be related to changes in lithofacies, hydrofacies, and/or changes in moisture content and water chemistry. The purpose of the resistivity survey is to map changes in subsurface lithofacies and hydrofacies down to a depth 45 meters below land surface. The resistivity data were collected using a 56 electrodes 8 channel SuperSting^TM instrument with a Swift^TM automatic resistivity acquisition system. Two types of surface resistivity arrays were used on the first two survey lines to determine which array would produce the best results. The first array used was the pole-dipole array and the second array was the Wenner Array. Based on the initial results data from the Wenner Array produced the best results. Both arrays had the same field geometry, i.e., 4 meter electrode spacing and 8 meter line spacing.


*Tomographic Datasets* Seismic and electrical tomographic datasets have been collected between wells POS1, POS2, and POS3 in the saturated section of the P-Area Reactor site by LBNL. Three seismic tomographic datasets were collected along the entire saturated section in each well.
An example of Seismic Tomogram between Wells POS2-POS3	These data were collected using two Geometrics Geode recording systems, a 4” Lawrence Berkeley National Laboratory piezoelectric source, and an ITI 48 channel hydrophone sensor string with 0.25 station spacing. A 215 V DC power source yielding ~kV peak-to-peak signal was used. The waveforms were recorded over 20ms, sampled at 25microsecond intervals, and stacked 25 times. A 60Hz notch filter was used to reduce power line noise, and spectral analysis suggests that usable frequencies are apparent up to ~6000 Hz. Electrical resistance tomographic (ERT) datasets were also collected between the three wells, but only over regions located between the wellbore screened intervals. These data were collected using a Zonge GDP 32II acquisition system operating at 1 Hz (with a 5 and 60 Hz filter), with electrode strings were fabricated at LBNL. Inversion of both datasets is underway at LBNL.

2) Dual Domain Model Development and Preliminary Simulations

A key component of our project entails the development of a dual-domain modeling approach using the TOUGH2 family of codes that can incorporate the key interactions between mobile and immobile transport regions that are expected to play a role in long term plume behavior. The model is intended to be used in two modes: preliminary, to guide the field scale characterization effort, and subsequently, to be parameterized using hydrological parameters obtained from field scale datasets. In the preliminary mode, the model will be used to indicate which components most control transport in the dual domain system, and thus which components need be characterized. Examples of potential characterization targets include: orientation of strata relative to head gradient, fraction of fine to coarse grained units within a given region; or spatial correlation of low hydraulic conductivity lithofacies.

High resolution permeability field and solute transport simulation

During the first year of our project, a version of the LBNL iTOUGH2 code (using the EOS7R module) was modified to enable simulation of dual-domain transport in a sedimentary system. To develop and test the modeling framework, a synthetic high-resolution heterogeneous permeability field in two-dimensions was generated. Spatial variability is represented as a distribution of two discrete facies with median permeabilities differing by 4 orders of magnitude, and continuous variability within facies (Figure above). Solute concentration was assumed to be initially uniform throughout the 10 meter high by 100 meter long region. Solute break-through out to 4000 days was recorded. Using the high-resolution model as the reference, coarse mesh single- and dual-domain simulations were optimized using iTOUGH2. The dual-domain formulation produced excellent agreement with the reference data (Figure below). In contrast, the single-domain model deviated significantly from heavy tail observed for the high-resolution model. Ongoing work is extending this model and using the simulations to determine what are the critical components of the system that need be characterized.

Comparison of coarse-mesh single- and dual-domain break-through curves to high-resolution transport simulation.