Abstract:
A series of significant hydrocarbon releases occurred at a former Ingersoll Rand
facility in Phillipsburg, New Jersey between the 1940’s and 1960’s causing an extensive
light non-aqueous phase liquid (LNAPL) contamination presence in an aquifer which
supplies water for human use. LNAPL migration is difficult to monitor, as its non-aqueous
behavior causes migration to occur in the direction of groundwater flow or along less
confining features such as fractures or faults. Non-invasive geophysical techniques hold
promise for LNAPL characterization at fractured bedrock sites such as this one, given their
ability to provide information on a variety of physical earth properties. The primary
objective of this research is to test the hypothesis that a thrust fault has influenced LNAPL
migration at the site. Geophysical field models were generated from dipole-dipole
Electrical Resistivity Imaging (ERI) data which displayed a roughly linear, significantly
low-resistivity zone that has been interpreted as a shallow, near vertical expression of a
thrust fault which had an apparent influence on LNAPL migration. To complement the ERI
survey, optical televiewer, caliper, formation resistivity, temperature, gamma, fluid
resistivity, and heat pulse flow meter data were collected in two bedrock wells which
provided insight as to the general strike and dip of fractures and bedding features, minor
variations in lithology, and vertical flow conditions as they relate to LNAPL migration.
This research also tests the hypothesis that direct detection of LNAPL contamination
through the use of ERI measurements is not likely based on predictive forward models
constructed prior to the field study. A forward model was generated in the lab using
predetermined parameters of known and anticipated geologic conditions. Both the forward
model prediction and actual field models revealed that direct detection was not possible as
a function of loss of resolution of ERI measurements with increasing depth and the lack of
contrast between the measured ERI responses of the LNAPL plume and surrounding
geologic material. The geophysical techniques applied in this research show a promising
outcome for use at other sites because structural features contributing to contaminant
transport were readily identified in the modeled ERI data sets, and the forward model and
field models both suggest the direct detection of LNAPL contamination would not be
possible under similar conditions.
