Time-lapse 3-D seismic imaging of shallow subsurface contaminant flow

This paper presents a physical modelling study outlining a technique whereby buoyant contaminant flow within water-saturated unconsolidated sand was remotely monitored utilizing the time-lapse 3-D (TL3-D) seismic response. The controlled temperature and pressure conditions, along with the high level of acquisition repeatability attainable using sandbox physical models, allow the TL3-D seismic response to pore fluid movement to be distinguished from all other effects. TL3-D seismic techniques are currently being developed to monitor hydrocarbon reserves within producing reservoirs in an endeavour to improve overall recovery. However, in many ways, sandbox models under atmospheric conditions more accurately simulate the shallow subsurface than petroleum reservoirs. For this reason, perhaps the greatest application for analogue sandbox modelling is to improve our understanding of shallow groundwater and environmental flow mechanisms. Two fluid flow simulations were conducted whereby air and kerosene were injected into separate water-saturated unconsolidated sand models. In both experiments, a base 3-D seismic volume was recorded and compared with six later monitor surveys recorded while the injection program was conducted. Normal incidence amplitude and P-wave velocity information were extracted from the TL3-D seismic data to provide visualization of contaminant migration. Reflection amplitudes displayed qualitative areal distribution of fluids when a suitable impedance contrast existed between pore fluids. TL3-D seismic reflection tomography can potentially monitor the change in areal distribution of fluid contaminants over time, indicating flow patterns. However, other research and this current work have not established a quantifiable relationship between either normal reflection amplitudes and attenuation and fluid saturation. Generally, different pore fluids will have unique seismic velocities due to differences in compressibility and density. The predictable relationships that exist between P-wave velocity and fluid saturation can allow a quantitative assessment of contaminant migration.