Computational Modeling of Seismic Wave Scattering by High-Porosity Geologic Inclusions in the Central Basin Platform of the Permian Basin

Significant untapped domestic petroleum resources exist within oil-saturated high-porosity zones of the Central Basin Platform (CBP) geologic province of the Permian Basin in Texas and New Mexico. However, high-porosity inclusions are relatively small and deeply-buried, rendering detection and characterization via conventional surface seismic reflection methods difficult. A quantitative measure of "detectability" of an isolated porosity anomaly embedded within a typical CBP geologic section is obtained by numerical modeling of the 3D seismic wavefield scattered by the inclusion. The energy of scattered arrivals observed on a surface recording spread is determined, and is subsequently normalized by the energy of a prominent primary reflection event from the background 1D CBP section. Synthetic seismic data are calculated with an explicit, time-domain, O(2,4) finite-difference algorithm that solves the velocity-stress partial differential equations of isotropic elastodynamics. All arrivals (reflections, refractions, primaries, multiples, diffractions, surface waves) are generated with fidelity, provided spatial and temporal grid intervals are chosen appropriately. In order to accommodate the large size earth model (>27 million gridpoints), long trace durations (>8000 timesteps), and broadband wave propagation (to 100 Hz) in reasonable execution times, a parallel domain-decomposition algorithm is used. Incorporation of anelastic attenuation into the modeling study, via the memory-variables approach, is currently underway. Scattered energy from a sub-wavelength-size porosity inclusion consists of extremely low-amplitude P and S events observed on both horizontal and vertical motion components. These arrivals are effectively masked by larger amplitude (x100) reflections from interfaces of the CBP geologic model, and must be isolated by a subtraction technique. Repetitive parametric modeling allows the range of inclusion characteristics (size, shape, depth, impedance contrast) leading to detectable seismic responses to be inferred. Sandia National Laboratories is operated by Sandia Corporation, a Lockheed-Martin company, for the USDOE under contract DE-AC04-94AL85000.