Understanding the compound seismic signature of CO₂ injection in a thin carbonate reservoir during enhanced oil recovery

Abstract

This study focuses on isolating the seismic signature of supercritical CO₂ injected into a thin (below temporal seismic resolution) oomoldic carbonate reservoir, the Plattsburg Formation of the Lansing Kansas City Group, during a pilot enhanced oil recovery program utilizing supercritical CO₂ as the primary enhanced oil recovery agent. The program was monitored by a series of time-lapse seismic surveys in which a baseline survey was conducted prior to CO₂ injection with monitor surveys following in approximately two to three month intervals. A weak time-lapse anomaly between the baseline and monitor surveys allowed for successful tracking of the CO₂ front in the reservoir. However, quantitative interpretation has been challenged by the low compressibility of the reservoir reducing the fluid effect as well as tuning in the seismic signal resulting from the thinness of the reservoir (about 5 m). The data was detuned by employing time thickness and amplitude curves produced by a wedge model utilizing a representative wavelet extracted from the data. Subsequent fluid replacement modeling utilizing Gassmann’s equations predicts a 1.85% decrease in seismic impedance and an ~8% change in horizon amplitude at the reservoir level of the seismic data. This change is of insufficient magnitude to be observable in difference maps and is smaller than an approximate 15% noise component of the seismic data spectrum. This, in essence, means that the background noise in the seismic data likely far exceeds the signal containing the fluid response at the reservoir level and in turn the fluid-related time-lapse seismic signature cannot be interpreted quantitatively.