3-D, Near-Surface Seismic Reflection Investigation at a Groundwater Contamination Site

We conducted a near-surface, 3-D seismic reflection survey at a groundwater contamination site. Part of a month-long experiment that included tomography and downhole seismic studies, the experiments were designed to image the near surface (<20m) at a high level of detail (< 0.5m laterally and vertically) to enhance ongoing remediation activities at the site. The site, Operable Unit 2 (OU2), located at Hill Air Force Base in Ogden, Utah has been subject to continuing efforts to remove dense, nonaqueous-phase liquids (DNAPLs) released at the site. Subsurface mapping of the near-surface geology, based on more than 200 monitoring wells drilled at the site, has not been sufficient to allow effective remediation. The near-surface geology of OU2 includes unconsolidated sands, silts and gravels overlying a thick clay aquiclude. Incised in the clay is a paleochannel about 15m deep trapping both groundwater and DNAPL. The 3-D reflection survey, covering an area of 95 by 37m centered over the channel, used approximately 630 RefTek "Texans", a single-channel, non-cabled recording system. Receivers were spaced at 0.35m intervals across the channel profile (east - west) with a 2.1m separation along the channel strike (north - south). The source, a .223 caliber rifle, was fired in a rotated brick shooting pattern between the receiver lines producing a data set of over 1.8B traces with .175 x .175m CMP bin size. Over 3,700 useable shot records were taken, forming a data volume of 3.6 Gbytes. After data preprocessing, filtering (90-440Hz passband), velocity analysis and stacking, the 3-D stack shows coherent shallow reflections from the stratigraphy within the channel and the top of the clay layer from 20 to 80ms. Inline sections show the changing channel profile, steeply dipping along the western wall, with a more gentle slope along the east wall. Consistent with depth-to-clay values from the well data, the depth to the base of the channel, the likely collection points for DNAPLS, varies from 60ms to 80ms at different points within the stack. The sections also correlate with our results from the initial 2-D survey at OU2. Time slices show the lateral extent of the channel. The east edge of the channel is visible at 40ms, and both channel banks are visible at 50ms within the stack. From 70-90ms only deeper portions of the channel are visible within the south portion of the survey. These variations match monitoring well locations at which deeper channel depths were measured. Comparison of the seismic data with maps made from well data show excellent correlation between the separate data sets with the seismic data successfully imaging the varying depth-to-clay at the site from 3 to 15m. The reflection images also agree well with the travel time tomography images of the channel made from the complementary tomography experiment described by Azaria et al (also in this session).