Statistical integration of time-lapse seismic and electromagnetic data for reservoir monitoring and management

Abstract

Joint integration of seismic and electromagnetic (EM) data has been studied for better characterizing hydrocarbon reservoir because two surveys can give complementary information on reservoir properties. The advantages of joint integration of different geophysical data can also be gained in reservoir monitoring by integrating time-lapse seismic and EM data. Statistical rock physics modeling, where the joint probability density function (PDF) of geophysical measurements and reservoir properties is modeled, can be used to integrate time-lapse seismic and EM data. In applying this statistical method to joint integration, however, there is a critical issue of the scale differences between well logs, seismic, and EM measurements. In this work, we established a workflow to statistically integrate time-lapse seismic and EM data by developing proper and efficient methods to simulate the joint PDFs considering the scale differences. In the developed workflow, analogous reservoirs are created by unconditional multi-point geostatistical simulation, SNESIM (Strebelle, 2002). Then, well logs in the target reservoir are used to assign reservoir properties to the analogous reservoirs. The field-scale PDFs of the analogous reservoirs are simulated by forward modeling and inversion, or filtering. We tested this approach in classifying facies; oil sand, brine sand, and shale of a two-dimensional synthetic cross section. We found that the significantly improved classification with simulated field-scale joint PDFs.