|
Geology:
The Savannah River Site covers 300 square miles in South Carolina on the South Carolina-Georgia border. The site is located on Cretaceous-Tertiary age Atlantic Coastal Plain sediments that form a seaward thickening wedge from 850 ft (N) to 1400 ft (S) thick across the site. In the northern part of SRS basement is Paleozoic metavolcanic schist and in the southern part of SRS the Dunbarton Triassic rift basin underlies the Coastal Plain. Post-Cretaceous reactivation of basement faults has caused displacement of the Coastal Plain sediments. The near surface geology at MSB comprises Eocene age Upper Coastal Plain sedimentary units. On the surface is the "Upland unit" which consists of cobbles and coarse sand and ranges in thickness from zero to 50 feet. Underlying the Upland are the Tobacco Road and Dry Branch formations which are composed of fine to medium grain sand, clayey sand, and discontinuous layers of clay. The amount of clay in the formations decreases with depth. Below the Dry Branch Formation is the Warley Hill Formation, which is composed of sand and glauconitic green clay. In the study area the "green clay" interval of the Warley Hill Formation ranges in thickness from 0-10 feet and occurs at a depth of approximately 155 feet below the surface. This clay, when present, is considered the confining unit which separates the overlying surficial aquifer form the semi- to confined aquifer below. The DNAPL in M-Area pools on top of the "green clay".
Seismic Reflection
Research:
The approach taken was three fold consisting of: 1) evaluation of existing geological and geophysical data concerning the amount and distribution of
DNAPL, 2) seismic modeling to determine whether an amplitude anomaly could be expected from DNAPL saturated sediments, 3) acquisition, processing, and interpretation of seismic lines designed to specifically image the
DNAPL.
Seismic Modeling
Very little shallow subsurface velocity and density information was available in the vicinity of M-Area seepage basin. P and S-wave velocity information were taken from suspension logger measurements in a nearby borehole and a P-wave VSP was recorded in well MSB-9A. Density information was taken from a density log in MSB-9A. No cores were available to do
velocity/DNAPL saturation measurements; thus, mathematical models based on Gassmann's theory of fluid saturated materials were input in to the seismic model.
Seismic amplitude versus offset (AVO) models were constructed using Aki-Richards and
Smith-Gidlow approximations to the Zoeppritz equations. The model was a sand wedge saturated with either water or TCE overlying either a clay layer or a water saturated sand. In the models, water or DNAPL saturation was assumed to be 100 percent of the available pore space.
The models indicate that the presence of TCE should cause the reflection coefficient at the contact between the sand and the clay aquitard to be become strongly positive for angles of incidence exceeding 22 degrees. Also, if TCE saturated sand is present, the reflection coefficient from top of the saturated "layer" becomes strongly negative at far offsets exceeding 22 degrees angle of incidence. If no DNAPL is present, the reflection coefficient between the sand and the clay is slightly negative and becomes strongly negative past 60 degrees angle of incidence as the critical angle is approached.
The results of the model study strongly suggested that the presence of high concentrations (free-phase) of DNAPL at the green clay horizon could cause detectable seismic amplitude anomalies. Furthermore, the models indicated that the seismic frequencies greater than 100 Hz would be need to resolve the interfaces.
Seismic Acquisition
 The original project consisted of recording one seismic profile, line M-1, across a known DNAPL plume that has a free-phase component (well MSB-3 & MSB-22). Based on the model results, it was determined that approximately half the receivers would be located under 22 degrees angle of incidence and half over. The depth of the target and the need to ensure adequate spatial and temporal sampling dictated the recording parameters. A 48 channel seismograph, sledge hammer source, and an off-end spread were used. Receiver and shot spacing were 2 ft to yield 24 fold data nominally. After preliminary processing and AVO interpretation of line M-1 showed an amplitude anomaly at the location and depth of known free-phase product in well MSB-22, two additional lines were recorded, M-2 and M-3. For lines M-2 and M-3 a 96 channel recording system was used, which allowed a greater range of offsets to be recorded. The receiver and shot spacing for lines M-2 and M-3 was also 2 feet to yield 48 fold data.
Data Processing & Interpretation
The presence of the Upland unit at M-Area seepage basin causes scattering of seismic energy. In addition ground roll is quite severe to the extent that raw field records show no discernible reflected energy. Extensive detailed data processing was required to obtain useful images. Coherent noise attenuation was achieved through use of F-K filters, spiking
deconvolution, and high-pass filtering. Considerable effort was expended in detailed stacking velocity analysis and residual statics determination. For display, the data were filtered to a 90-275 Hz band. To analyze AVO variations near and far offset stacks were generated as well as
Smith-Gidlow "fluid-factor" stacks.
Line M-2 stack. Every other trace plotted. The "green clay" is interpreted at 110 ms.
The near-far offset stacks proved to be the most useful interpretation tool. The seismic models indicated that the presence of DNAPL would cause the reflection coefficient between sand and the "green clay" to become strongly positive at far offsets. On line M-1, which passes through well MSB-22 where free-phase DNAPL has been bailed, an amplitude anomaly is present at the depth of the known contamination. On line M-2 far offset high amplitudes at the level of the "green clay" prompted the drilling of well MRS-2. In MRS-2 contamination in high ppm concentration was found at the depth predicted by the seismic data. Other amplitude anomalies were found to be associated with high DNAPL concentrations in existing wells. Where DNAPL concentrations were low, ppb or non-detect, amplitude anomalies were not present.
|
|
 |
Line M-1.
Near offset stack top, far offset stack bottom.
Amplitude envelope display.
Note amplitude anomaly on far offset stack.
Free-phase product has been pumped
from well MSB-22 |
Line M-2.
Near offset stack top, far offset stack bottom. Amplitude envelope display.
Note amplitude anomaly on far offset stack.
Well MRS-2 was drilled on the anomaly.
DNAPL was found in high ppm concentrations
indicating the presence of free-phase DNAPL. |
The results appear promising that high DNAPL concentrations can be imaged in this area. However, validation of these assertions and the model requires additional boreholes. |
