Cross-borehole radar methods were used to monitor a field-scale biostimulation pilot project at the Anoka County Riverfront Park (ACP), located downgradient of the Naval Industrial Reserve Ordnance Plant, in Fridley, Minnesota. The goal of the pilot project is to evaluate biostimulation using emulsified vegetable oil to treat ground water contaminated with chlorinated hydrocarbons. Vegetable oil is intended to serve as substrate to naturally occurring microbes, which ultimately break down chlorinated hydrocarbons into chloride, carbon dioxide, and water through oxidation-reduction reactions. In support of this effort, cross-borehole radar data were acquired by the U.S Geological Survey in five site visits over 1.5 years. This paper presents level-run (zero-offset profile) and time-lapse radar tomography data collected in multiple planes. Comparison of pre- and post-injection data sets provides valuable insights into the spatial and temporal distribution of both emulsified vegetable oil and also the extent of ground water with altered chemistry resulting from injections — information important for understanding microbial degradation of chlorinated hydrocarbons at the site.
In order to facilitate data interpretation and test the effectiveness of radar for monitoring oil-emulsion placement and movement, three injection mixtures with different radar signatures were used: (1) vegetable oil emulsion, (2) vegetable oil emulsion with a colloidal iron tracer, and (3) vegetable oil emulsion with a magnetite tracer. Based on petrophysical modeling, mixture (1) is expected to increase radar velocity and decrease radar attenuation relative to background — a water-saturated porous medium; mixtures (2) and (3) are expected to increase radar velocity and also increase radar attenuation due to their greater electrical conductivity compared to native ground water.
Radar slowness (inverse radar velocity) tomograms and level-run profiles show decreases in slowness in the vicinity of injection wells. Slowness anomalies are observed only in planes connected to injection wells, indicating that the emplaced emulsified vegetable oil does not migrate far after injection. In contrast to the localization of slowness anomalies, attenuation anomalies are observed in all level-run profiles, particularly those downgradient of the injection wells. Despite the expected signatures of different tracers, increases in attenuation are observed downgradient of all three injections; thus, we infer that the attenuation changes do not result from the iron tracers. One viable explanation for the observed attenuation changes is that products of oil-enhanced biodegradation (for example, ferrous iron) increase electrical conductivity of ground water and thus radar attenuation.
Application of radar methods to data from the ACP demonstrated the utility of radar for monitoring biostimulation. Results of level-run and tomographic surveys identified (1) the distribution of emulsified vegetable oil, and (2) the distribution of ground water with oil-affected chemistry. Ongoing research efforts include simultaneous tomographic inversion of radar data from multiple planes, petrophysical modeling, geostatistical interpolation, and development of an integrated interpretation considering conventional borehole logs and surface-to-borehole radar data.