Shear magic

It is becoming easier to get the message across about the value of land multicomponent seismic surveys. In the United States and Canada alone more than 150 projects have been undertaken using multicomponent acquisition systems which offer excellent coverage with survey sensors recording in the region of 18,000 channels.

The breakthrough development was the introduction 5 or so years ago of Micro Electro Mechanical System (MEMS) technology, now available in land seismic acquisition systems from Sercel and Input/Output. This was the vital hardware ingredient that facilitated sensors able to more effectively record shear-wave data, which potentially provide additional information for imaging the structure and lithology of oil and gas reservoirs. In multicomponent studies, it is common to refer to the compressional wave as PP, the pure shear waves as SS and the converted waves as PS. There is little dispute these days that, technically at least, the PS wave can produce extra information of value.

Multicomponent acquisition systems equipped with MEMS-based digital accelerometers have significant technical and operational advantages over methods used previously. The conventional design for multicomponent data acquisition on land depends upon the response of one vertically deployed and two horizontally deployed elements. One of the issues is that vector fidelity can be seriously compromised unless the geophone unit is approximately level, which is not always easy to guarantee during field operations. With the deployment of single sensors (point receivers) at each recording station, made possible by MEMS, a number of factors help to improve vector fidelity, including the capacity to measure and correct for sensor tilt. The digital output through the sensor and integrated field electronics eliminate any problem with pick up noise, cross-talk and sensitivity to leakage. Harmonic distortion is also reduced. In addition, single sensor deployment is operationally more efficient, with lighter and less complicated equipment.

These are some of the reasons why the results of surveys which collect converted-wave data (together with P-wave data) can legitimately challenge traditional P-wave projects as the way forward. It's also why companies are increasingly willing to pay more for the demonstrable added value. Our experience is that clients will upgrade normal P-wave projects into multicomponent if the project costs don't rise by more than 15% to 20% and if they are convinced that the conventional P-wave data result will not be compromised by the requirement to deploy single sensors rather than geophone arrays.

In this context it is important to understand that PP and PS data acquired in a multicomponent survey should be treated as companion datasets combining to produce a better picture of the subsurface and additional information about rock properties. Together they can provide crucial information, for example, in reducing drilling risks and identifying previously unseen accumulations of recoverable hydrocarbons. It is wrong to imagine that converted-wave data on its own will produce a spectacular outcome. That only happens where there is a problem such as a gas cloud. Then the specific insensitivity of the shear wave to gas-saturated sediments above the reservoir provides the imaging solution.

Significantly, there are other motivations besides recording shear waves for oil companies to acquire multicomponent data. For example, 3-C data provide an opportunity to apply polarization filtering for the suppression of ground roll rather than depend upon large arrays of geophones. This helps both P-wave and converted-wave imaging and various P-wave prestack analyses: it's a case where multicomponent recording can also benefit the P-wave data. In some locations, clients may be frustrated in their effort to secure permits for any seismic work on land. In such situations it makes sense to collect as much information as possible while in the field with a permit, i.e., acquire PP and PS data, even if it means archiving the converted-wave information for later analysis.

In North America Veritas has now recorded more than 100 such surveys. Along with improved acquisition methods, the company has had to convince clients that the tools for processing, analysis and interpretation of the data are effective and ultimately rewarding, maximizing output in the exploration, development or production phase. From the many surveys available, we are showcasing three different settings and issues where the use of converted-wave data in combination with PP data has produced tangible, positive results.

One illustration is from a multicomponent (3-C/3-D) survey at North Emerald in the Anadarko Basin gas production province, where natural gas is produced from thin sands (less than 25 ft or 8 m each) in the Mississippian Springer Formation at depths of near 10,500 ft (3,200 m). Production is primarily related to porosity development. Sands with greater than 7 ft (2 m) of 8% to 12% porosity tend to be commercial, while sands with porosities of less than 8% are generally non-productive.

PS and PP data from the survey provide potentially different views of subsurface reflectivity. Figure 1 suggests that changes in PS reflectivity may offer better clues than PP reflectivity changes when it comes to distinguishing dry holes from gas producers. To check this finding, Hampson-Russell's EMERGE software package was used to quantitatively determine whether the PP and PS data together could provide a better empirical estimate of gas production than using PP data alone. For 17 wells in the 3-C/3-D image area with Springer penetrations and tests, a simple log trace was created representing cumulative gas produced. Using cumulative gas as the target, a step-wise regression was performed, and the most significant seismic attributes were identified. For validation, each well was removed from the process and tested based on the relationships determined by the remaining population.

Figure 2 shows the empirical results for predicting cumulative gas production. The combination of PP and PS data correctly identifies the three wells with higher cumulative gas produced. Crossplots of estimated gas versus actual gas production show that using PP and PS data together produce a more accurate linear fit than when using the conventional P-wave data alone.

In another study, we were able to demonstrate that co-locating single-sensor 3-C MEMS receivers with conventional geophone groups as an embedded test can be a cost-effective method of evaluating 3-C technology. In this case we were able to not only establish the basic feasibility of the technology but also its capacity to determine the orientation of horizontal stress through azimuthal anisotropy. The test took place during a 3-C/3-D survey of a fractured carbonate gas play in the United States. It showed strong mode-converted reflections at the depths of interest and demonstrated characteristics typical of shear waves in azimuthally anisotropic media. Detecting the presence of azimuthal velocity anisotropy is important because measurements of "fast" and "slow" S-wave polarizations can be related to open fractures and crack-like pore structures in the subsurface. Another key point was that the azimuth stacks of this embedded test displayed all the expected characteristics of azimuthally anisotropic PS-waves (Figure 3).

Our third example is from the shallow heavy oil sands of Alberta, Canada, where PP and PS survey data are shown to be in excellent agreement with the corresponding well information. The survey was acquired over the McMurray productive formation, which is 265 ft (80 m) thick at a depth of 720 ft (220 m) at the survey location. Production comes from bitumen-saturated, Cretaceous-age sands within a fluvial valley-fill depositional system. Extraction involves drilling pairs of horizontal well bores, vertically separated by 16 ft (5 m), injecting steam into the upper well bore and producing the heated bitumen via the lower well bore. A main source of economic risk is shale plugs and layers interfering with the injection and extraction process by imposing impermeable barriers to fluid flow.

3-C/3-D data were acquired using digital MEMS sensors and integrated with the available well control. Seismic inversions of both the PP and PS data were input into the Hampson-Russell EMERGE software to produce an estimate of shale volume. A cross-section through the shale volume is shown in Figure 4. Four wells are integrated into the profile showing the measured sand/shale volumes from log control. Core data were also extracted at these four well locations to produce an interpreted core cross-section (Figure 4). The seismic estimate, using PP and PS data, is seen to provide good corroboration of the well information.

Hopefully these necessarily abbreviated descriptions provide some further evidence of the potential value of multicomponent data when incorporated into land seismic acquisition. Given the dramatic progress made in the last few years, we anticipate that such surveys will soon become the norm in many regions of the world.