MULTI-COMPONENT TIME-LAPSE SEISMIC: ON SATURATION-PRESSURE DISCRIMINATION AND STATISTICAL DETECTABILITY OF FLUID FLOW

Shahin, A., Stoffa, P.L., Tatham, R.H. and Sava, D., 2011. Multi-component time-lapse seismic: on saturation-pressure discrimination and statistical detectability of fluid flow. Journal of Seismic Exploration, 20: 357-378. We evaluate the production-induced time-lapse response of three sandstone reservoirs corresponding to three rock physics models spanning a full range of common degrees of consolidation. For a range of water saturation and pore pressure, we compute multi-component (MC) seismic, i.e., conventional P-P, converted P-SV, and pure shear SH-SH, traveltimes through and reflection coefficients (RCs) at top of the sandstone reservoir embedded in a background shale. Then, we compute changes in traveltimes and reflection coefficients with respect to a reference traveltime and RC calculated at reference saturation and pressure conditions. We plot changes in RCs versus changes in traveltimes. The corresponding time-lapse cross-plot shows interesting patterns for saturation and pressure changes and has the potential for quantitatively discriminating pressure and saturation changes. Next, we deploy a statistical method to determine the efficacy of MC seismic in detecting production-induced time-lapse changes. The significant and representative data in time-lapse cross-plot allow one to statistically analyze the detectability of a known scenario of saturation and pressure changes using MC seismic attributes. Applying different thresholds for traveltimes and RCs, we construct single and joint probability detectors that help to compare the likelihood of detection of a known change in dynamic reservoir properties using different component of the seismic data. Our analysis demonstrates that seismically detection of the changes in fluid saturation and pressure is significantly limited for a consolidated sandstone reservoirs. However, the detection is plausible for poorly to medium consolidated reservoirs in the presence of realistic seismic noise levels. In these cases, conventional P-P seismic data is dominant in amplitude change compared to converted P-SV and pure SH-SH seismic data. P-P data reflects changes in both fluid saturation and pore pressure. However, the main effect of P-P data is the saturation component than pressure. SH-SH seismic data capture most the pressure information using traveltimes of pre-stack data. P-SV seismic data is the weakest detector of changes in time-lapse amplitude, but its traveltime shows an intermediate detectability between P-P and SH-SH seismic data for changes in both fluid saturation and pressure.