Quantitative Interpretation of 4D Seismic data for Thermal EOR Applications

The earliest 4D seismic field studies were designed to monitor thermal recovery operations in heavy oil reservoirs in the USA, Canada and Indonesia. Thermal recovery operations were chosen to demonstrate the viability of 4D seismic because of the large contrasts in the seismic velocity of the heavy crude as it is heated (Nur, 1982), combined with the weak framework of the unconsolidated sand reservoirs and the relatively shallow reservoir depths. All of these factors contribute to the large seismic changes that occur during thermal production.

Early pilot projects successfully demonstrated the viability of the 4D seismic method as documented by Greaves and Fulp (1983), Stang et al (1984), and Pullin et al (1987). The application of thermal recovery processes to heavy oil reservoirs clearly caused differences in amplitude and travel time in the seismic data. However, as the thermal operations continued in mature fields it was often difficult to understand the magnitude and types of seismic changes that were observed in the data. This led to a more rigorous study of the temperature, pressure and saturation changes that occur during thermal recovery and the effect of these properties on the velocity and density of the reservoir rocks (Wang et al, 1996).

Improved rock physics and synthetic seismic modeling methods, which include phase transitions for bitumen and water, are essential to the correct understanding of the complex seismic changes that occur in mature thermal recovery operations. Seismic inversion offers the additional potential of simplifying time-lapse seismic monitoring by converting the observed amplitude changes to acoustic impedance changes, which are more directly related to rock properties. Unfortunately, this benefit frequently isn’t realized due to differences in event times between surveys that are caused by velocity changes within the reservoir interval. These time differences cause a misalignment of geologic events below the reservoir that result in artifacts when the acoustic impedance results for each survey are subtracted. This problem is compounded when the velocity changes in the reservoir contain frequency components that are below the effective seismic bandwidth, and this often occurs in thermal recovery processes. Fortunately, the low frequency velocity change information can be inferred through a careful analysis of the changes in event timing between base and monitor surveys. When this low frequency information is included in the seismic inversion, the results improve significantly over conventional inversions (Hirsche et al, 2006).

As the interpretation of 4D seismic surveys has improved there has been a growing desire to understand the seismic response in terms of underlying changes in reservoir properties such as temperature, pressure and fluid saturations. This is especially true in thermal operations where accurate information about these properties can have significant economic impact. Unfortunately, the seismic response to changes in these properties can be highly non-unique. For example, a small drop in pressure can liberate methane from cold bitumen and this can cause the same magnitude of velocity decrease that occurs at very high reservoir temperature. Detailed modeling and analysis can help to distinguish the reservoir changes that underlie a specific seismic response. The integration of production and injection data together with volumetric information from the seismic monitoring results can further reduce uncertainty and derive a more quantitative interpretation (Hirsche and Ma, 2006).