How wide-azimuth, long-offset seismic reduces field development costs in the Bakken

During the 2013–2014 winter season, Global Geophysical Services acquired a 250-mi2, wide-azimuth long-offset 3D multi-client seismic program within the Fort Peck Indian reservation. The crew was equipped with approximately 12,000 recording units and 10 vibrators.

A recent 3D seismic survey combined the enabling acquisition technology of the AutoSeis HDR, with patented HTI/VTI processing, for the inclusion of far-angle offset amplitudes in pre-stack inversions, ultimately providing a better understanding of the Bakken formation. Including information from the far-angle offset amplitudes of 3D seismic data in pre-stack inversions provides high-resolution attribute information that cannot otherwise be obtained. 

REGIONAL BACKGROUND

The Williston basin underlies large areas of northwestern North Dakota, northeastern Montana and southern Saskatchewan. A key unconventional reservoir within the Williston basin is the Bakken formation, which is comprised of a lower shale member, a middle dolomite member and an upper shale member. Despite being one of the largest continuous deposits of oil and natural gas in the U.S., until the late 1990s, the Bakken was considered only a marginal repository, due in large part to the low permeability of the reservoir. However, the advent of horizontal drilling and hydraulic fracturing has transformed the Bakken into a prolific oil and natural gas producer. Furthermore, the Bakken is expected to be productive for decades to come. The USGS has estimated undiscovered, technically recoverable volumes of 3.65 Bbbl of oil, with estimates of gas exceeding 3 Tcf in the U.S. portion of the formation.

One industry key to unlocking the Bakken has been geosteering a drill bit along geologic zones using high-resolution, 3D seismic data. The challenge, however, is that the overlying strata of the Williston basin contain shallow evaporates, such as salt lenses and dirty salts, which impact travel times, and ultimately, defocus seismic reflections. Furthermore, as a shale play, conventional reservoir characterization methods, such as porosity alone, do not characterize the Bakken’s subsurface adequately enough to make development and production decisions. Additional information, specifically regarding fractures, is crucial to unlocking the productivity of unconventional shale plays, such as the Bakken.

Useful amplitude information for unlocking the productivity of unconventional shale plays resides in the higher-incidence angles (e.g., 30°–50°) of 3D seismic data. To utilize the information contained in the far-angular offsets, processing must solve for seismic anisotropy. Once seismic anisotropy corrections have been made, including these far-angle offset amplitudes in the pre-stack inversion reduces uncertainty in acoustic impedance, elastic impedance and density estimates of various geologic layers. Further, characterizing how far-offset amplitudes (AVAz—Amplitude Variation with Azimuth) and velocities (VVAz—Velocity Variation with Azimuth) vary with shot-receiver azimuth can provide proxy attributes for reservoir properties, such as intensity and direction of fracturing, and even fluid type within the fractures. These properties are important, as they describe information about reservoir “connectivity” and “confinement” of fractures. When combined with more traditional rock property attributes in a multi-variable statistical approach, these additional proxy attributes improve significantly the predictions of sweet spots and areas to avoid for well placement and completion strategies, which in turn, can significantly lower field development costs.

WIDE-AZIMUTH DATA ACQUISITION

Because azimuthal anisotropy can be a significant source of fracture information, wide-azimuth data, leveraged by full offset- and azimuthal-friendly processing, is extremely useful for unconventional development planning. Historically, the wide-azimuth, long-offset data required for this type of advanced processing have been expensive to acquire, due to the limited channel count availability typical of cabled recording systems. In recent years, however, 3D land seismic innovations, such as slip-sweep, simultaneous recording techniques and autonomous nodal recording systems, have increased crew productivity and raised available channel count, enabling the cost-effective acquisition of wide-azimuth, long-offset 3D land seismic data.

During the 2013–2014 winter acquisition season in Montana, Global Geophysical Services acquired a 250-mi², wide-azimuth, long-offset, 3D multi-client seismic program. The Fort Peck 3D survey was conducted in Roosevelt County, Montana, entirely within the borders of the Fort Peck Indian reservation. The program utilized the AutoSeis autonomous nodal recording system. The high-definition recording units proved extremely reliable throughout the project, and, on multiple occasions, allowed the crew to record around large, no-access areas with relative ease and without interruption to data acquisition.

Because the project area was located within the jurisdiction of the Bureau of Land Management, it was subject to an archeological survey, otherwise called a Cultural Resource Study. The archeological survey for Fort Peck identified more than 2,100 historic and prehistoric sites, each of which required a radius of avoidance of 100 ft. As a result, the locations of individual source points were adjusted, and vibrator and vehicle routes were carefully defined. To accommodate the significant number of sites identified, the project was permitted in three phases, allowing archeologists and surveyors enough time to re-route vibrator access.

The crew was equipped with approximately 12,000 channels of AutoSeis and 10 vibrators. The program was designed as 20 lines of 96 stations per line. The receiver interval was 220 ft inline and 1,100 ft cross-line, with a triangular geophone array and 3 ft between elements. Source specifications were designed at intervals of 220 ft in the inline direction and 880 ft in the cross-line direction, with a 311-ft diagonal. Vibrators were divided into three fleets of three vibrators, and performed one 36-sec sweep at each point.

Crews began laying out receivers in November 2013, and production began shortly thereafter. The crew averaged nearly 450 vibe points per day over the course of the project and surveyed more than 65,000 combined source and receiver points.

‘INVERSION-READY GATHERS’ PROCESSING

Affordable, timely acquisition of wide-azimuth, long-offset 3D data, like that of the Fort Peck 3D program, is the first step in accessing the valuable information that resides in the 30°–50° far-angular offsets. To fully leverage the information contained in this more complete type of data acquisition, seismic data processors must confront the presence of seismic anisotropy and its effect on the far offsets. Two types of anisotropy are typically found in wide-azimuth, far-offset PP seismic data: vertical transverse isotropy (VTI) and horizontal transverse isotropy (HTI).

VTI anisotropy is best seen in far-angle offset gathers, and the result of its influence is often referred to as a “hockey stick” effect. VTI anisotropy describes velocity variations in the vertical plane containing the trace shot and receiver locations. Eta is the attribute that characterizes VTI anisotropy and allows removal of this hockey stick effect. A common source of VTI anisotropy is fine layering found in dewatered shales, which have higher velocity parallel to bedding and which are sampled by higher angles of incidence. The VTI correction removes the reflector-offset variation in reflected energy arrival times and is important for flatness of offset gathers at higher angular offsets.

Fig. 1. Offset amplitude inversion of the Fort Peck 3D seismic data showing the rock property attributes density and shear-wave velocity, extracted at and around the Bakken interval.

Azimuthal velocity variations in the earth manifest as small shot-receiver, azimuth-related timing distortions in 3D seismic data, especially on the higher angular offsets. Detailed analysis of 3D surface seismic data shows a subtle velocity (or reflector arrival time) dependency on the shot-receiver azimuth of each seismic trace. This velocity vs. azimuth variation (sometimes referred to as VVAz) is a form of seismic anisotropy known as HTI. The HTI correction removes the reflector azimuthal variation in reflected energy arrival times and is important for flatness of azimuthal gathers, especially at higher angular offsets. If not handled properly during imaging, these timing distortions will adversely affect both bandwidth and amplitudes during Kirchhoff summation into final offset traces. This far-angle bandwidth and amplitude distortion of the final migrated gathers negatively impacts pre-stack inversion and the accuracy of rock property attributes, especially shear-wave velocity and density.

Figure 1 shows the rock property attributes—density and shear-wave velocity—estimated from offset (elastic) impedance inversion of the seismic data. The comparison of impedance inversions at the Bakken level conveys the difference in attributes, when HTI anisotropy is incorporated into the pre-stack migration step. As discussed, having accurate amplitude information at the far-offset angles is critical in accurately estimating rock properties using elastic inversion techniques. Global Geophysical has named offset-azimuth gathers that are flat in both offset and azimuth domain, all while preserving true amplitude, “inversion-ready gathers.”

A popular geologic interpretation of the source of HTI anisotropy is vertical fracturing, which is true for the Bakken formation. In this case, a seismic trace recorded on the earth’s surface with a shot and receiver pair at an azimuth, such that the seismic energy crosses the fractures, will see a slightly lower velocity, and hence experience a small increase in reflector two-way time at this azimuth. Quite often, accompanying this delay in reflection arrival time is a noticeable decrease in amplitude and frequency content of the reflected energy, which is attributed to increased inelastic attenuation of the energy moving across fractures. Conversely, seismic energy that moves in the same direction as (or parallel to) the fractures will exhibit no delay in reflection arrival times, and little or no attenuation in amplitude and frequency content. Characterizing both the amplitude and velocity variation vs. azimuth produces useful attributes, potentially relating to fracture density, fracture direction and even fracture fluid type.

Velocity anisotropy can derive from natural rock fractures, neostress (present-day stress), and rock fabrics, such as mineral-grain alignments and cross bedding. Fluid pressure and rock mechanical properties also affect seismic velocities. Fractures are a dominant cause of both velocity variations and velocity anisotropy.

Global Geophysical has developed a patented approach for characterizing RMS (root mean square) HTI anisotropy, using far-angle, wide-azimuth PP surface-seismic data. The approach seeks to quantify, at each CMP/time point within the seismic volume, the two attributes that describe HTI anisotropy: Vfast azimuth (direction of anisotropy) and the Ellipticity Factor (magnitude of anisotropy.) The algorithm measures HTI anisotropy by systematically imaging the data, using different combinations of the HTI parameters and then determining which azimuth/factor pair maximizes stack power at each output image point. Since ranges of likely HTI parameters are scanned systematically, the approach has become known as Migration Scanning Analysis and incorporates, uniquely, the anisotropy into the HTI analysis, itself.

High-resolution parameter volumes of Vfast Azimuth and Ellipticity Factors, as a result of plotting stack power vs. scanned Vfast azimuth and Ellipticity Factor, are considered RMS HTI values, as they are measured from the earth’s surface and show cumulative overburden effects from the surface down to each impedance boundary. These RMS HTI values are required to flatten the azimuthal gathers in all space and time. In fact, the basic quality control (QC) of the RMS HTI values is how well they flatten the azimuthal gathers. In this manner, QC of these parameters, using azimuthal gathers, is analogous to using offset gathers to QC velocity and eta fields: flatness of gathers.

Global’s method incorporates both the RMS VTI and HTI parameters into a proprietary and patented pre-stack migration algorithm, which, in turn, produces more accurately positioned steep-dip and fault planes, and helps preserve accurate offset amplitude information, especially at the crucial far-angle offsets. Once the RMS HTI parameters are deemed of high quality, the RMS parameters are inverted, using vector equations, to produce interval attributes more closely associated with individual geologic layers.

Fig. 2. Interval HTI magnitude co-rendered with fault probability at the Bakken.

The HTI attributes rendered with fault probability maps, often correlate to well productivity. To better predict well performance using a 3D multivariate production prediction model, the azimuthal attributes should be integrated with a full suite of pre-stack and post-stack seismic attributes, and engineering and production data. Reliable well productivity predictions can lead to better well placement and completion strategies, helping to reduce risk and overall field development costs, while optimizing production rates and recovery.

The company’s derived HTI attributes from hydrocarbon basins around the world have been validated with borehole image logs, cross-dipole sonic logs and well production data. Real seismic data examples show the improved uplift of incorporating anisotropy corrections into the workflow, both in the improvements to the processed seismic data, and also in the subsequent seismic attribute work. Preliminary results of the application of this methodology to 3D data from the Fort Peck 3D have highlighted some intriguing features of the Bakken interval.

Figure 2 shows interval HTI magnitude co-rendered with fault probability at the Bakken level. The magnitude attribute describes the degree or amount of anisotropy, and is the difference between Vfast and Vslow interval velocities measured in the seismic data. It is a proxy for the dominant fracture intensity. Here, the background color scale portrays interval HTI magnitude, in which warmer colors, such as yellow, orange and red, represent greater magnitudes. It is interpreted as a proxy for fracture intensity. The white vectors in Fig. 2 represent fault probability from the stacked seismic data, and measure the likelihood of the presence of faults, fractures and their locations. Altogether, the plot demonstrates that, generally, large-magnitude values correspond to increased fault probability, aligning along the same regions. Inclusion of the far-angle, wide-azimuth information provides additional fault and fracture information, and more accurate rock-property attributes.

Fig. 3. Interval HTI magnitude co-rendered with interval HTI Vfast azimuth vectors at the Bakken. The white needles are oriented by Vfast azimuth and the needle length is scaled by magnitude of anisotropy.

In Fig. 3, interval HTI magnitude is co-rendered with interval HTI Vfast azimuth vectors at the Bakken level. The white needles are oriented by Vfast azimuth, describing the direction of anisotropy in each part of the formation. The needle length is scaled by magnitude of anisotropy, and similar to Fig. 2, the background color property represents interval anisotropy magnitude, as well. The high resolution of both HTI azimuth and magnitude paint an intriguing description of the Bakken. Upon examination, it becomes clear that the orientation of the direction of principal stress is actually perpendicular to the main fault. This suggests compressional or perhaps wrenching forces at work here, and perhaps, unexpected micro-crack and fracture directions.

With decades of production expected, exploration of the Bakken has much further to go. Advanced processing methods, and inclusion of far-angular offset data in the processing work flows, are likely to become increasingly important and useful, as new areas of the Williston basin are explored and increased efficiency and innovation are demanded by the low oil-price environment.