Figure 1. Modeled 1GHz GPR reponse to a thin layer of oil under ice assuming relative permitivities of 6, 2.2, and 88 for sea ice, oil, and sea water respectively. A) Reflected GPR signal with oil layer thickness varying from 0 - 3 cm. B) Maximum instantaneous amplitude for the model data in A). C) Instantaneous phase at peak amplitude for the model data in A). D) Thinbed amplitude vs offset (AVO) response for transverse electric (TE) and transverse magnetic (TM) polarizations. The AVO gradient is computed for incidence angles from 0 to 45 degrees. While amplitude, phase, and AVO data attributes all show a significant response to the trapped oil, the AVO gradient appears to be the most sensitive indicator.

Detecting oil under sea ice using Ground-Penetrating Radar

Ground Penetrating Radar (GPR) has been used in numerous arctic studies to image both internal structures within snow (e.g., Bradford and Harper, 2002), the ice/water contact and subsurface geology below freshwater ice (Best et al., in press; Bradford and McNamara, in review), and the sea ice/sea water contact. In GPR studies, a transmitting antenna generates a downward-directed oscillating electric field that propagates through the subsurface and is reflected back toward a receiving antenna at boundaries separating materials with differing electric properties (dielectric constant, electric permittivity, and conductivity). The reflected wave field is recorded and used to produce a reflector map in travel time, similar to a cross section of the subsurface. The large permittivity contrast between sea ice and sea water (~6:88) and between sea ice and oil (~6:2.2) suggests we can derive an accurate map of subsurface boundaries in these conditions using very high frequency GPR antennas (~1 GHz). Consequently, radar is a tool well suited to imaging both the base of ice, and the sub ice conditions for both fresh and sea ice conditions

The resolving power of the GPR system limits the thickness of sub ice oil that can be measured directly, ie by measuring the travel time difference between wavelets reflected from the top and bottom of a layer. The wavelength of the signal controls the resolution, with a shorter wavelength signal capable of resolving finer features. When a layer is thinner than about ¼ of the dominate wavelength of the GPR signal, it is impossible to clearly differentiate wavelets reflected from the top and bottom of the layer and a simple reflector map is not sufficient to confidently identify the presence of oil under the ice. In this case, rather than relying on a direct measure of traveltime differences, we analyze specific attributes of the reflected wave such as amplitude or phase. Attribute analysis is commonly used in oil and gas exploration to identify relatively thin reservoirs of hydrocarbon in sedimentary rocks. Amplitude and phase measurements can be made from typical fixed antenna GPR data, which is relatively fast and inexpensive to acquire. Another potentially useful measurement is the change in amplitude with increasing offset or AVO analysis. However, this method requires multiple receiver offsets for each source antenna position leading to additional acquisition costs (i.e. time). Although many GPR systems only acquire data at a fixed receiver offset, our system allows the operator to decouple the source and receiver antennas to acquire such data. Bradford (2003, in review) has shown that GPR AVO analysis can be an effective tool for detecting thin layers of hydrocarbons in groundwater studies and we expect that the methods will be more robust in the sea ice/oil system due to decreased stratigraphic complexity. Using an analytical thin bed GPR model, we computed the GPR response, at 1 GHz, to a thin oil layer trapped under sea ice, with the layer varying in thickness from 0 – 3 cm (Figure 1). In fixed antenna mode, we find a 24% increase in amplitude and a 20% change in phase for a 2 cm thick oil layer relative to no oil present (Figures 1b and 1c). Further, the AVO response in TM antenna configuration shows even greater sensitivity with a dramatic increase in the AVO gradient (Figure 1d).

Water strongly attenuates the radar signal, with the rate of attenuation increasing as the dissolved solid concentration (electric conductivity) increases. Thus, pockets of brine trapped in ice may limit signal penetration. It is important to recognize that entrapped brine and sea ice anisotropy may alter the measured GPR attributes. This problem is minimized in field data analysis by computing attributes relative to a background response that is measured from the data.

Given the uncertainties of actual field conditions, we believe it is important to consider all available data analysis tools to include multi-offset acquisition and analysis. Even though it is more expensive, the increase in sensitivity to the presence of oil under the ice may make multi-offset acquisition a necessity to overcome potential problems in field data such as low signal to noise, variable surface conditions, or ice heterogeneity. We propose to acquire fixed antenna data to analyze travel time, amplitude, and phase information, then collect multi-offset GPR data to characterize the AVO response. By performing this set of tests under variable conditions to include smooth or rough bottomed ice and variable oil thickness, we will determine the methods necessary to most effectively detect the presence of oil under ice.