Integrating Ensemble Data Assimilation and Indicator Geostatistics to Delineate Hydrofacies Spatial Distribution

Abstract:

We present a new framework for delineating spatial distributions of hydrofacies from indirect data by linking ensemble-based data assimilation method (e.g., Ensemble Kalman filter, EnKF) with indicator geostatistics based on transition probability. The nature of ensemble data assimilation makes the framework efficient and flexible to integrate various types of observation data. We leveraged the level set concept to establish transformations between discrete hydrofacies and continuous variables, which is a critical element to implement ensemble data assimilation methods for hydrofacies delineation. T-PROGS is used to generate realizations of hydrofacies fields given conditioning points. An additional quenching step of T-PROGS is taken to preserve spatial structure of updated hydrofacies after each data assimilation step. This new method is illustrated by a two-dimensional (2-D) synthetic study in which transient hydraulic head data resulting from pumping is assimilated to delineate hydrofacies distribution. Our results showed that the proposed framework was able to characterize hydrofacies distribution and their associated permeability with adequate accuracy even with limited direct hydrofacies data. This method may find broader applications in facies delineation using other types of indirect measurements, such as tracer tests and geophysical surveys.