Establishing Spatial Pattern CorrelationsBetween Water Saturation Time-Lapse andSeismic Amplitude Time-Lapse

摘要

The extensive spatial coverage of seismic data provides aunique source of information widely used for surface and faciesmodeling, faults detection, more generally static heterogeneitiesmodeling. The advent of 4D time-lapse seismic surveys opens anew dimension – the possibility of monitoring fluid flow henceproduction. Discounting the favorable cases in which fluiddisplacement can be seen directly from time-lapse seismic data,the preliminary challenge is to establish a correlation betweensaturation changes and seismic data changes. Because of thelow resolution of seismic data, one should not expect any usefulpoint-to-point correlation; instead, in the general case onemight expect correlations between spatial patterns of seismictime differences and corresponding spatial patterns ofsaturation changes. Spatial patterns involve multiple-locationswithin a fixed template window, and are summarized by theprincipal/canonical components (PC/CC) of the within-templatevariability of each variable.This preliminary study aims at developing a generalmethodology to establish such pattern correlations, hencepreparing the way for a systematic utilization of time-lapseseismic surveys beyond trivial visual observation. Theexhaustively known synthetic Stanford V 3D clastic reservoir,over which both flow simulation results and repeated syntheticseismic surveys are available, is used to explore suchcorrelations. It is found that the first few PC’s or CC’s of the4D seismic and water saturation variables do capture thespatial patterns, and they account for most of the total variance.Through the principal/canonical analysis, significantcorrelation is found between multiple-point spatial patterns of4D seismic variable and spatial patterns of water saturationtime-lapse variable. The 4D seismic patterns are then put to useto predict changes in saturation.