Spatial orientation and distribution of reservoir fractures from scattered seismic energy

Abstract

We present the details of a new method for determining the reflection and scattering characteristics of seismic energy from subsurface fractured formations. The method is based upon observations we have made from 3D finite difference modeling of the reflected and scattered seismic energy over discrete systems of vertical fractures.Regularly spaced, discrete vertical fractures impart a ringing coda type signature to any seismic energy which is transmitted through or reflected off of them. This signature varies in amplitude and coherence as a function of several parameters including: 1) the difference in angle between the orientation of the fractures and the acquisition direction, 2) the fracture spacing, 3) the wavelength of the illuminating seismic energy, and 4) the compliance, or stiffness, of the fractures. This coda energy is the most coherent when the acquisition direction is parallel to the strike of the fractures. It has the largest amplitude when the seismic wavelengths are tuned to the fracture spacing, and when the fractures have low stiffness. Our method uses surface seismic reflection traces to derive a transfer function which quantifies the change in an apparent source wavelet before and after propagating through a fractured interval. The transfer function for an interval with no or low amounts of scattering will be more spike-like and temporally compact. The transfer function for an interval with high scattering will ring and be less temporally compact. When a 3D survey is acquired with a full range of azimuths, the variation in the derived transfer functions allows us to identify subsurface areas with high fracturing and determine the strike of those fractures. We calibrated the method with model data and then applied it to the Emilio field with a fractured reservoir giving results which agree with known field measurements and previously published fracture orientations derived from PS anisotropy.

Characterization of in-Situ Stress and Permeability in Fractured Reservoirs

“…Estimates of the aggregated fracture orientations using these volumes from the analysis of scattering indices are shown as rose plots in Figure 8, modified from Willis et al (2004). The rose plots, which estimate the distribution of fractures in the area around well 4 (left) and well 8 (right), agree with the ENE and NNW directions derived in the studies mentioned above and shown in Figure 7.…”

“…The scattering indices were sorted and directions for those with the highest angular contrast in values (in this case for scattering index values greater than 5) were plotted as 'quivers' (i.e., short line segments) giving a map view of the location and direction of possible fractures determined by this method (Figure 20). These plots were also shown in previous reports and papers (e.g., Willis et al, 2003Willis et al, , 2004. The locations of these fracture measurements, however, are taken from stacked data (rather than migrated data), and as such, their locations are in the un-migrated positions.…”

“…Briefly, this process entails for each azimuth stack trace: 1) identifying a target depth range (the fractured layer) to investigate 2) extracting a window of data containing reflections which are above this fractured layer (denoted as the input), 3) extracting a window of data below the fractured layer (denoted as the output) containing reflections which includes the scattered energy, 4) taking the autocorrelation of the extracted windows, 5) windowing the autocorrelations to focus on the source wavelet near zero lag and 6) computing the spectra ratio of the lower and upper windowed autocorrelations from the amplitude spectrum of the time domain transfer function (see Willis et al, 2004). The right panel in Figure 2 shows the time domain transfer functions for the 50m fracture case.…”

“…In this paper we describe our recent work (Willis et al, 2003;Willis et al, 2004a;Willis et al, 2004b;Willis et al, 2004c; to extract fracture distribution and orientation from scattered coda waves where the fracture systems are of a size comparable to the wavelength of the seismic source. We first describe our modeling results of vertically fractured reservoirs, our methods to extract the fracture properties from surface reflection seismic acquisition data, and finally the results on field data.…”

“…For fractures whose lengths are comparable to seismic wavelength, discrete fracture models (DFM) are more realistic . Willis et al (2004) used the DFM and scattered seismic energy to determine spatial orientation and distribution of reservoir fractures. They applied their method to synthetic and field data and showed results consistent with log data.…”