Near-Surface Complexity Could Masquerade As Anisotropy

Summary

Near-surface velocities could vary with azimuth, impacting seismic data processing and interpretation. In this study, we developed a methodology to investigate the variations of near-surface velocities with azimuth, using 3D turning-ray tomography. The input data are the first arrivals selected from pre-defined azimuth sectors in terms of shot-receiverpair directions. The output velocities from tomography correspond to the selected azimuth sectors. A near-surface tomography study based on seismic data from a shallow heavy-oil reservoir in Canada has suggested that the observed azimuthal traveltime variations are not necessarily related to azimuthal (HTI) anisotropy induced by the stress field or fractures. It could also be caused by the nearsurface heterogeneity or acquisition footprint. Near-surface complexity could masquerade as anisotropy. Potentially this can influence statics and prestack imaging.

Introduction

Recent advances aimed at improving resolution and data quality from land seismic surveys have significantly improved the structural imaging and reservoir characterization capabilities. In particular, point-source, point receiver and high fold acquisition designs together with 3D noise attenuation processing algorithms are being increasingly adopted in exploration and exploitation areas for mapping stratigraphic channels and fractured reservoirs. Static corrections become more robust with single source and high-density seismic surveys (Diggins et al., 2005) and, when adequately sampled, surface-waves and other types of coherent noise can be effectively attenuated from data collected using cross-spread acquisition geometry.

However, seismic data collected from multi-azimuth surveys can present both opportunities and challenges for data processing and interpretation. The observed "apparent azimuthal anisotropy" on the Common-Offset-Common- Azimuth (COCA) cubes (Cheadle et al., 2001) may not be necessarily related to the anisotropy induced by the stress field or fractures. It could also be caused by the nearsurface heterogeneity and acquisition footprint.

To investigate the presence of apparent azimuthal anisotropy, we conducted a 3D azimuthal near-surface tomography study using a field dataset from the Surmont Project, Northern Alberta in Canada. The input data are the first arrivals selected from pre-defined azimuth sectors in terms of shot-receiver-pair directions (Figure 1). Crossspread acquisition was employed in this survey and source and receiver lines were orthogonal (Figure 2). The distances between the inlines and crosslines are 12 m and 9 m, respectively. Although the surface elevation has only about 200 m of relief within the 76 km2 survey area, this is significant because the target reservoirs are less than 500 m below the topography. Previous studies based on regional geology and stratigraphy have indicated that the reservoir rocks in the study area are primarily unconsolidated sands containing heavy oils overlain by a relatively soft overburden. The lack of any extensive set of fractures or faults in these rocks and overburden raised the question regarding the source of the azimuthal velocity variations as seen on the COCA display in Figure 2.

Results from 3D azimuthal turning-ray tomography indicate that the near-surface velocities estimated from tomography are raypath dependent. They are a function of azimuth, the maximum offset and the first-break pick fold. The observed azimuthal traveltime variations (lower right in Figure 2) could be caused by the near-surface heterogeneity or acquisition footprint.