Interpolating Qualitative Time-Lapse Seismic Interpretations

Abstract

SummaryTime-lapse seismic surveys, even if repeated at different calendar intervals, can only capture snapshots of a reservoir’s state. To obtain continuous information from these surveys, we propose a method to interpolate qualitative time-lapse seismic interpretations. As inputs, we take the sequence of interpreted boundaries of an expanding time-lapse seismic anomaly. Our solution combines the fast-marching method (FMM) and the Pollock (1988)pathline-tracing algorithm to estimate the intermediate boundary positions at times between acquisitions. For validation, we use a synthetic data set representing the scenario of a waterflooded sandstone reservoir. The results confirm that, despite small localized discrepancies in shape, at each time, the estimated boundary positions are in close agreement with the truth case. When time-lapse anomalies result primarily from changes in water saturations, the estimated boundary-arrival times resemble the actual water-front-arrival times. Hence, combined with classical immiscible displacement theory, the estimated arrival times can support a practical, model-independent, quantitative estimation of water saturations in the reservoir.