3-D Seismic Reflection Surveys For Direct Detection Of DNAPL

Mary-Linda Adams, Brian Herridge (Resolution Resources, Inc., Ionia, MI)
Nate Sinclair (Naval Facilities Engineering Service Center, Port Hueneme, CA)
Tad Fox, Chris Perry (Battelle Memorial Institute, Columbus, OH)

1st International Conference on Remediation of Chlorinated & Recalcitrant Compounds
Monterrey, CA
May 18-21, 1998

Abstract

High resolution three dimensional (3-D) seismic reflection data have been collected at several sites where Dense Non-Aqueous Phase Liquids (DNAPL) are or are believed to be present. Confirmatory drilling and chemical analysis have been performed at some of these sites. The seismic surveys were performed to more completely characterize the structure and stratigraphy to delineate fluid migration pathways and traps. The feasibility of using seismic attribute analysis to directly image DNAPL was also demonstrated. Information from these studies can be used to better characterize sites and to optimally position remedial systems.

Introduction

3-D high resolution seismic reflection surveys have been performed at hazardous waste sites to better map the heterogeneous subsurface structure and stratigraphy. The technique is non-invasive and was originally developed by the oil industry to locate oil reservoirs. Advances in hardware and software have made it possible to adapt the technology for use at shallow depths of 5 to 200 feet (1.5 to 61.0 m).

The seismic reflection method is based on the principle that sound waves or acoustic signals will reflect from interfaces between layers in the subsurface. A sledgehammer impact will generate waves that penetrate the earth and spread out along spherical wavefronts. Most stratigraphic changes produce reflections that make it possible to distinguish structure and stratigraphy. DNAPL can be found in pools within topographic lows on confining layers. However, more often DNAPL occurs as microscopic and macroscopic globules or ganglia. These ganglia appear to attenuate the acoustic signals that pass through them.

Previous Work

The first 3-D seismic reflection survey performed at a hazardous waste site was in 1994, when 2-D and 3-D data were collected, in conjunction with a fracture trace analysis on aerial photographs and a review of background site history, at Site 9 Naval Air Station North Island (NASNI) in San Diego harbor, California.

2-D seismic surveys have routinely been collected over oil prospects since the 1940's. It was not until the 1980's that technology advancements made it possible to perform 3-D surveys. 2-D surveys show a cross section of the earth, however 3-D surveys make it possible to examine a site in three dimensions as a volume. Once processed at a workstation, the data can be analyzed using a personal computer (PC), from north to south, east to west, top to bottom, and along any other selected azimuth.

Figure 1 shows how the data were collected in the field with 144 live channels. A grid of six cables were placed on the ground and connected to the 144 channel seismic system. Twenty-four geophones (receivers) were bonded to pavement or spiked into the ground along each cable. A sledgehammer or power- assisted weight drop (source) was impacted at the surface and the signal was simultaneously received by the 144 geophones and stored by the seismograph. Multiple impacts were used at each source station to vertically stack the data to reduce noise. After the source was impacted at each station along Line 1 and at each station along Line 2, cables 1 and 2 were moved to new positions to become Lines 7 and 8. The source pattern was repeated at Lines 3 and 4. In this "leapfrog" manner, the 3-D seismic array was pushed across the site.

The seismic survey at Site 9 at NASNI (Resolution Resources, Inc., 1995) was performed over a topographically low area called the "fiery marsh", where an estimated 32 million gallons (121.1 million liters) of liquid waste were disposed from 1942 to 1972. The seismic and photographic data showed that the disposal area was at the juncture of several faults. Figure 2a shows a conventional plot of trace 38, north of the source area. The numbers across the top show the line and trace locations in the grid, and the numbers along the left side show the two-way reflection time in seconds. Three fault zones are apparent from line 58 to 62, line 94 to 109, and from line 128 to 138. Figure 2b shows trace 24, which cuts through the center of the source area. The seismic signal has been attenuated below the source area, probably as a result of the presence of DNAPL, since the structure and stratigraphy between traces 24 and 38 are similar.

To further investigate this phenomenon, instantaneous attribute analysis was applied to the data. Complex seismic trace attributes are a mathematical transformation that can be used to quantify attributes, such as instantaneous amplitude, phase, and frequency for each sample of the trace. Complex seismic trace attributes can be separated and quantified with amplitude and angular (phase and frequency) information. Attribute analysis has been used in the oil industry since the 1970's to emphasize important aspects of the geology, which are not as obvious in the conventional sections. The instantaneous amplitude (envelope) attribute has been applied to trace 24 in Figure 2c. It is a polarity insensitive measurement of the signal level for each sample of the seismic waveform. The envelope attribute has been used to delineate fractures, because fractured materials attenuate seismic signals faster than non-fractured materials. The data suggest that additional attenuation may be expected as a result of DNAPL ganglia. The values of the envelope data (Figure 2c) range between about 0 (white) and 200 (black), and the attribute in contaminated areas is as low as 20, while uncontaminated areas are as high as nearly 200, showing an order of magnitude difference. This work was the first field evidence that suggested that organic liquid compounds attenuated the seismic signal, however bench scale studies such as Wang and Nur (1990) and Geller and Myer (1994) have shown that seismic amplitudes are sensitive to concentrations of nonaqueous phase liquids (NAPL) and represent a step in the application of seismic measurements to NAPL detection in the field.

Recent Work

The initial work at NASNI was funded by Navy Clean and under the auspices of the Navy Environmental Leadership Program (NELP) (1996). Based on this work the Environmental Security Technology Certification Program (ESTCP) funded further studies at Department of Defense (DoD) sites to demonstrate the use of 3D seismic imagery to characterize the subsurface and to directly detect DNAPL. Sites were selected with different geologic characteristics. A 3-D seismic survey was performed at the K-1 area at Letterkenny Army Depot (LEAD), located in south central Pennsylvania in the Valley and Ridge physiographic province. The K-1 area was a former lagoon, as large as 200 x 50 feet (61.0 x 15.2 m), that was used from the 1950's to 1970 as a disposal site for liquid wastes. The K-1 area was filled in by 1970 and later capped. Thirty-three wells have been drilled in this area. The bedrock is Ordovician limestone belonging to the St. Paul Group with yields up to 255 gpm (965.3 lpm). The Letterkenny Fault trends northeast across the seismic grid, which is bounded to the north by the Pinola Fault (ESE, 1993). The immediate area underlying the seismic survey has been folded, faulted, and fractured. According to the LEAD well database from 1996-97, the highest concentration of total VOCs was 43,377 m g/l found in Well 95-DA7. Trichlorethylene (TCE) is one of the main chemicals of concern and the solubility in water of TCE is 110,000 m g/l (Pankow and Cherry, 1996). This is the value that was chosen to define the presence of DNAPL for this demonstration.

To date three wells have been completed, based on the seismic image from among ten possible targets that were selected. Preliminary evaluation of the data shows that:

The wells in the immediate K-1 area were sampled during the drilling phase. None of the previously installed wells showed VOC concentrations as high as two of the four wells which were drilled during the validation of the seismic demonstration.

Future plans include drilling wells at two other sites, where seismic data have been collected, which are underlain by bedrock. Also at NAS Alameda, California, the use of cone penetrometer results in saturated sediments will be used to verify anomalies in very shallow seismic data.

Download this paper:   DNAPL paper.pdf    (376k)

Download poster presentation:   DNAPL poster.pdf    (700k)

References:

Environmental Science and Engineering (ESE), Inc., 1993. Remedial Investigation of the Southeastern Area at Letterkenny Army Depot Draft-Final Report, Vol 1, prepared for U.S. Army Environmental Center, IR Division, Aberdeen Proving Ground, Maryland.

Geller and Myer, 1994. Seismic Imaging of Organic Liquid Contaminants in Unconsolidated Porous Media. Manuscript in progress. Lawrence Berkeley Laboratory, Berkeley, California: 22 pgs.

Navy Environmental Leadership Program, 1996. 2D and 3D High Resolution Seismic Reflection Surveys to Image the Subsurface. Success Stories NELP Fact Sheet No. 6.

Pankow, J.R. and Cherry, J.A., 1996. Dense Chlorinated Solvents and Other DNAPLs in Groundwater: History, Behavior, and Remediation. Waterloo Press, Portland, Oregon.

Resolution Resources, Inc., 1995. Final Report 2D and 3D High Resolution Seismic Reflection Surveys at Chemical Waste Disposal Area, Site 9 Naval Air Station North Island, San Diego, California.

Taner, M.T. and Sheriff, R.E., 1977. "Seismic Stratigraphy - Applications to Hydrocarbon Exploration." American Association of Petroleum Geologists Memoir 26: 301-344.

Wang, Z. And Nur, A., 1990. "Wave Velocities in Hydrocarbon-Saturated Rocks: Experimental Results," Geophysics. 55(6): 723-733.