next up previous [pdf]

Next: Results Up: Decker et al.: Carbonate Previous: Permo Triassic Khuff Model

Method

Seismic diffraction events carry much less energy than reflection events, requiring that they be separated to be utilized. Several methods for seismic diffraction extraction exist (Landa et al., 1987; Kanasewich and Phadke, 1988; Harlan et al., 1984; Khaidukov et al., 2004; Klokov and Fomel, 2012; Landa et al., 2008), including plane-wave destruction applied to common-offset data (Fomel et al., 2007).

Plane-wave destruction (PWD) filters (Claerbout, 1992; Fomel, 2002), determine the dominant slope of seismic events as they attempt to map data to adjacent traces. Data not conforming to the local slope field is iteratively minimized. Because reflection events appear planar in common-offset data while diffraction events appear hyperbolic, this residual will contain the set of diffractions along with random noise present in the data (Harlan et al., 1984).

Zero offset data are modeled using methods described in the subsequent section. We use PWD to determine the dominant slope field of our modeled zero-offset data and remove the reflections that conform with local slope, providing us with zero-offset diffraction data. Zero-offset ``conventional'' data containing diffractions and reflections as well as zero-offset diffraction data are migrated, providing our conventional and diffraction images respectively. A workflow for the diffraction extraction and imaging process starting from common-offset data is displayed in Figure 3. Plane-wave destruction of common-offset data may face difficulties extracting diffractions in regions with complex geometry or velocity structure (Decker et al., 2013). The synthetic models we use in this paper have small enough lateral velocity variations for this to not be an issue. If the wavefield is sufficiently complicated to prevent common-offset data plane-wave destruction from functioning properly other methods of separating diffractions exist, including plane-wave destruction of angle-migrated partial images (Decker and Klokov, 2014).

DD-PWD
DD-PWD
Figure 3.
Our diffraction imaging workflow
[pdf] [png]

Although we employ the same method of diffraction extraction on both the Orodovician and Khuff synthetic models, we adopt different methods of modeling and migration that are best suited for each model's scale and subsurface position. Reverse-time migration (Fomel et al., 2013b; Zhang and Sun, 2009) is used on the Ordovician model for greater accuracy while one-way wave-equation migration (Kessinger, 1992; Gazdag and Sguazzero, 1984) is utilized on the reservoir-scale Khuff model to allow for upward continuation of modeled data through an overburden to model the response of the interval at a geologically plausible depth.


next up previous [pdf]

Next: Results Up: Decker et al.: Carbonate Previous: Permo Triassic Khuff Model

2015-03-25