A practical method for estimating effective parameters of anisotropy from reflection seismic data

Abstract

The location of any event imaged by P-wave reflection seismic data beneath a tilted transversely isotropic (TTI) overburden is shifted laterally if isotropic velocities are used during data processing. The magnitude of the shift depends on five independent parameters: overburden thickness, angle of tilt, symmetry-axis velocity, and the Thomsen anisotropy parameters ε and δ. The shift also varies with source-receiver offset.We have developed a procedure to estimate these five parameters when the tilt of the symmetry axis from the vertical is equal to the dip of the TTI layer (except in the special cases transverse isotropy with vertical or horizontal axis of symmetry). We observe three attributes of seismic data processed using isotropic velocities: the zero-offset arrival time of a selected reflection, the difference in arrival time between a near-offset and a far-offset arrival, and the difference in imaged location (smear) of this target event between the same offsets. We then perform a cascaded scan of the five parameters to determine those combinations of the five that result in calculated attributes equivalent to the observed attributes. The multiple solutions are averaged to give the parameter estimates.Application of this method to synthetic and physical model reflection data results in multiple solutions, which are constrained and averaged to obtain the effective imaging parameters. These effective parameters are close estimates of the true model parameters in both cases. For field seismic data this procedure requires that there be a suitable observable event below the TTI overburden and assumes that the measured times and shifts are reasonably accurate.

The wedge model revisited: A physical modeling experiment

“…Historically, physical modeling was particularly useful when computing technology was not adequate for advanced numerical modeling, but to this day it still can lead to a better understanding of wave propagation in complex media, as in the case of 3D elastic and anisotropic models. In addition, because the physical models have known geometries and physical properties, they are ideal for comparisons with numerical models and field data from similar geologic situations, as well as for tests of processing, imaging, and modeling algorithms ͑e.g., Kosloff and Baysal, 1982;Cheadle, 1988;Edwards, 1997;Isaac and Lawton, 2004͒. At the University of Calgary, the Consortium for Research in Elastic Wave Exploration Seismology ͑CREWES͒ maintains a physical modeling facility capable of generating acoustic and elastic seismic data sets over complex physical models, as described by Cheadle ͑1988͒, Cheadle et al ͑1991͒, and Wong et al ͑2009͒.…”

Moveout approximations for P- and SV-waves in dip-constrained transversely isotropic media

“…For more details about DTI media, see, e.g., Audebert et al (2006) (who consider DTI as a special case of tilted transverse isotropy), Sava (2010, 2011), Golikov et al (2012). Models with dips can be a reasonable approximation of the complex geologic formations such as flanks of salt domes and fold-and-thrust belts (Tsvankin, 2001;Isaac and Lawton, 2004).…”

VTI anisotropic corrections and effective parameter estimation after isotropic prestack depth migration

“…Most of the conventional methods for flattening depth-migrated common-image gathers ͑CIGs͒ are based on an isotropic assumption and are inadequate for the anisotropic case. Successful anisotropic-parameter estimation will usually require several iterations of PSDM ͑Martínez and Lee, 2002; Isaac and Lawton, 2004;Yan et al, 2004͒, special treatment of near-and far-offset data in time and depth domains ͑Toldi et al, 1999͒, and use of nonseismic information.…”