Combined use of borehole geophysics and packers to site potable wells in a contaminated area in Montville, Connecticut

A leaking underground gasoline tank contaminated a crystalline bedrock aquifer in Montville, Connecticut, USA with MTBE and benzene. At the original residential bedrock supply wells, the median MTBE concentration was 165 micrograms per liter (mg/L), and the median benzene concentration was 320 mg/L. The maximum concentrations of MTBE and benzene were 4,300 mg/l and 1,700 mg/L, respectively. Because of the unavailability of a public water supply and the long-term expense of point-of-use (on-site) treatment systems, the Connecticut Department of Environmental Protection Leaking Underground Storage Tank Program considered drilling replacement wells for water supply, if suitable drill sites could be located. Borehole geophysical methods were used as part of the investigation to find suitable drill sites. The U.S. Geological Survey performed borehole radar logging in three of the most contaminated wells. Other geophysical logs were run in two of the wells to enhance the hydrogeological characterizations. These data, combined with straddle-packer testing provided by a drilling contractor, formed the basis of a conceptual model used in the search for discrete fractures with better water quality than provided by an open-hole sample.

At Property A, a single transmissive fracture was identified at the bottom of the well. This well, although having historically lower gasoline concentrations than the other two wells, had persistent high iron bacteria fouling of the filtration system. By 2002, concentrations of MTBE and benzene had decreased to 59 and 3 mg/L, respectively, and the water was treatable except for the iron. Because no water-bearing fractures were encountered above the well bottom, an alternate well site was selected based on a set of vertical fractures observed in a nearby outcrop, rather than on the geophysical data. The new well, sited along the strike of these fractures, yielded 9 gallons per minute (gpm) but was found to be more contaminated than the original well. MTBE and benzene were detected at 224 and 7 mg/L, respectively. At Property B, the isolated fractures associated with four radar reflections contained MTBE in concentrations ranging from 460 to 680 mg/L, with concentration increasing with depth. A new well site was selected based upon topography and physical limitations of the property. A target drilling depth was selected to avoid encountering the most contaminated fracture, as projected from the radar data in the contaminated well. A new well, drilled to the target depth, yielded 2 gpm, which was sufficient for domestic supply. No contaminants were detected during 7 years of annual sampling. Over the next 2 years, MTBE was detected twice at 2 and 8 mg/L. At Property C, the isolated fractures associated with 12 radar reflections and acoustic televiewer images yielded MTBE concentrations ranging from 47 to 1,200 mg/L and benzene concentrations from 6 to 1,000 mg/L, with concentrations generally increasing with depth. A new well site was selected based upon physical limitations of the site. A target drilling depth was chosen to avoid encountering the most contaminated fractures, as projected from the radar data in the contaminated well. A new well, drilled to the target depth, yielded 6 gpm. MTBE was detected at concentrations ranging from trace levels to 12 mg/L for 6 years. Benzene was not detected.

These case histories suggest that the combined use of borehole geophysics and discrete-fracture sampling can, in some cases, be used to predict the locations of less contaminated or uncontaminated fractures, at distances of tens of feet from contaminated bedrock wells. This information may be used to improve the chances of successfully siting alternate potable water wells. Likewise, the same data and approach potentially could be used for targeting specific fractures for remediation.