Spatial delineation of rock types in the Meramec Formation by integrating core and seismic measurements, Loyal area, Kingfisher County, Anadarko Basin

Abstract

The Mississippian-age Sooner Trend Anadarko Canadian Kingfisher (STACK) area of Oklahoma is one of the more important new resource plays in North America. It consists of multilevel drilling targets, including mudrock, siltstone, and carbonate reservoirs, some of which are self-sourced and all of which require hydraulic fracturing to produce commercial quantities of oil. The efficacy of the fracturing relies on whether the created fracture network drains from producible rock types. We have integrated data at core and seismic resolution scales to map the rock types’ spatially exhaustive lateral and vertical variation for this new play. We measured porosity, permeability, saturated compressional wave (P-wave) and shear wave (S-wave) velocity, and density in the laboratory at 2 ft intervals on the cores. We defined rock types corresponding to engineering flow units based on porosity and permeability measurements. We mapped these rock types against alternative elastic property crossplots, including P-impedance, S-impedance, Poisson’s ratio, lambda-rho, and mu-rho. P- and S- impedance are the only two independent elastic properties measured on the core samples, and other elastic properties are calculated from these. Hence, the rock types showed equal sensitivity to all of the elastic property pairs. We observe a 30% difference between the core and seismic elastic values, attributed to dispersion and effective media effects. We managed this discrepancy between the core and surface seismic elastic values by simple linear scaling. We found that the P-impedance and Poisson’s ratio core measurements were easier to scale linearly to the corresponding seismic frequency measurements. Once scaled, we used Bayesian classification to map the P-impedance and Poisson’s ratio rock type template defined by the core measurements to the same elastic parameters measured by prestack seismic inversion. As we move away from the six cored wells deeper into the basin and more distal from the shelf, we encounter P-impedance/Poisson’s ratio pairs not seen in the core, resulting in areas where the rock type is “unknown.” The seismically predicted rock types show an excellent correlation at the well locations and provided stratigraphically reasonable images away from the wells, suggesting that as more wells are cored, these unknown rock types can be classified.