Geostatistical shale models for a deltaic reservoir analog: from 3D GPR data to 3D flow modeling

The effects of shales on fluid flow in marine-influenced lower delta-plain distributary channel deposits are investigated using a three-dimensional ground-penetrating radar (GPR) data volume from the Cretaceous-age Ferron sandstone at Corbula Gulch in central Utah, USA. Using interpreted GPR data, we formulate a geostatistical model of the dimensions, orientations, and geometries of the internal structure from the subaerial exposure surface down to about 12 m depth. The correlation function between GPR instantaneous amplitude and shale index is built after statistical calibration of the GPR attributes (amplitude) with well data (gamma ray logs). Shale statistics are computed from this correlation function. Semivariograms of shale occurrence for ten accretion surfaces indicate only slight anisotropy in shale dimensions. Sequential Gaussian Simulation stochastically maps shales on variably dipping stratigraphic surfaces. Experimental design and flow simulations examine the effects of semivariogram range and shale fraction on breakthrough time, sweep efficiency and upscaled permeability. Approximately 150 flow simulations examine two different geologic models, flow in all three coordinate directions, 8 geostatistical parameter combinations, and 5 realizations for each combination of parameters. Analysis of the flow simulations demonstrates that shales decrease the sweep, recovery and permeability, especially in the vertical direction