Surface Seismic Data Acquisition: Current Technology, Trends and Issues

Introduction

We currently find ourselves armed with a dazzling array of technologies with which to acquire seismic data. These technologies, as well as their attendant trends and issues, have developed in response to both market and technical forces, powerful forces at work for several decades.

The market forces demand lower costs for finding and producing hydrocarbons. The technical forces, as if to test our collective sense of humour, have led us to realize that high-quality 3-D data represent surface seismic's most effective contribution to that goal. The rub, of course, is that acquiring and processing high-quality 3-D seismic data is, itself, very expensive.

This apparent dichotomy has resulted in pressure to reduce the cost of acquiring and processing 3-D data. Seismic contractors and equipment manufacturers have responded with methods and systems to increase acquisition and processing productivity. In the marine environment, the trend has been toward towing more streamers and sources in an effort to illuminate a larger subsurface area with each pass of the vessel over a prospect. Unfortunately, while providing higher productivity, these wide-tow configurations have also· introduced significant amplitude and phase anomalies into the final imaged data, reducing their quality.

In land operations, the quest has led to methods yielding higher production, lighter and more reliable equipment, and shooting geometries that require fewer shots or receiver stations, whichever are more expensive in a particular survey area. For every survey, the chosen recording geometry must be balanced against the terrain and available equipment to yield a smooth time and motion demand.

A second backdrop against which we geophysicists find ourselves laboring is an increasing demand to reduce our seismic crew "footprints" on the environment, as well as a growing demand to acquire data in areas that include transition zones. As luck would have it, many of these transition zones represent some of the world's more ecologically fragile areas. We've developed equipment and methods to satisfy these conflicting demands and continue to refine them.

Finally, and thankfully, some "spit and polish" is being applied to, or at least contemplated for, some existing seismic tools to improve our ability to meet today's market demands. These include reevaluating vibrator data processing to improve its quality, and infusing new technology into the ocean-bottom cable method to improve the resolution of marine data. It also includes a return of serious interest in recording mode-converted shear waves, as well as time lapse 3-D surveys for monitoring reservoir fluid movements.

3-D Geometry Design

The recording geometry with which 3-D data are acquired has been, and remains, among the most contentious issues facing geophysicists. It has a big impact upon the cost of a 3-D survey. Whereas modem wave-equation imaging algorithms such as DMO and migration have drastically improved the accuracy and resolution of seismic data, the amplitude and phase of their outputs have also proven much more sensitive to the acquisition geometry than are simple NMO and stacking. This sensitivity can be appreciated using Huygen's Principle which leads one to understand that every reflecting horizon, dipping or flat, can be represented as an infinite number of point diffractors. Each of these point diffractors contains dips ranging from -90 to +90 degrees.

We do not record the continuous wave field in time or space. In both dimensions, we sample the wave field in such a manner that we avoid spatially aliasing data from the point diffractors, as well as any source-generated noise. Any 3-D recording geometry being contemplated for a survey must be analyzed on the basis of not only midpoint fold, offset, and azimuth distributions, but also on the range and uniformity of dips that the wave equation operators contribute to each cell's output trace. Otherwise, the recording geometry may introduce amplitude and phase anomalies that could be misinterpreted as subsurface geologic changes.

Seismic Energy Sources

The air gun remains the dominant energy source in marine seismic acquisition. Until recently, the sleeve air gun demonstrated clear performance advantages over other available designs; recent improvements have narrowed the performance gap. There have been innovative developments with generator-injector gun clusters to produce quite high pulse-to-bubble ratios. And further refinements in methods of computing far-field air gun array signatures from near-field measurements have been targeted at time-lapse 3-D surveys, also referred to as 4-D.

Potential environmental concerns have resulted in some reengineering of the marine vibrator source in an attempt to improve low-frequency output. However, there has also been concern expressed that impulses may disturb marine creatures less than long duration oscillatory signals that are similar to, or could mask, their communication signals.

The vibrator remains the dominant land energy source, in part because of its lower operating expense. Units capable of generating peak forces approaching 65,000 pounds have been developed and have allowed sweep times to be significantly reduced, achieving higher productivity.

Recording Instrumentation

The 24-bit sigma-delta analog-to-digital converter has almost completely replaced the instantaneous floating point amplifier in new land, marine and transition zone recording systems. To-date, the primary benefits of this technology have been smaller and lighter instruments, reduced power requirements and improved reliability. These new systems provide increased linearity and instantaneous dynamic resolution, and write higher quality data to tape. However, our ability to capitalize upon these improvements by shrinking spatial arrays, recording more noise and attenuating that noise in data processing has been slower to develop.

The increasing numbers of surveys that include open water, transition zone and land topographies have spawned the development of more flexible recording systems. These systems are being designed to accommodate a receiver spread that contains ocean-bottom cable remote units, as well as land cable and radio linked remote units, all of which are controlled and recorded by a single central unit.

Positioning and navigation

Both land and marine navigation and positioning have been completely upgraded with Global Positioning Satellite receivers and systems. Although the system's signals that are available for civilian access are less accurate than those used by the military, the inclusion of receivers at known locations provide submeter accuracies in x, y, and z virtually everywhere in the world.

Field Processing

Compact, powerful workstations with full seismic processing software systems are appearing in many land crew base camps. The vast majority of land crew field processing is performed to verify the accuracy of the shooting geometry information associated with each shot record. The ability to leave the field with completely correct geometry information reduces the time required to process land 3-D data by between 20 and 30 percent. However, on some particularly remote prospects, much more extensive processing is being implemented.

On marine vessels, parallel-processing super computers have been installed. In one known case, a processing flow culminating in a final, pre-stack time migrated cube was implemented and was able to keep up with the data acquisition.

The future of powerful onboard computers, or at least that of having the processing geophysicists onboard, has been clouded somewhat by the advent of wavelet-transform based data compression algorithms. A major oil company recently implemented such a data compression algorithm and compared the processing results from a 3-D survey's uncompressed data with those from the same data that had been compressed by a factor of 60, sent to a processing center via a satellite link, and uncompressed using the same algorithm. The data telemetry rate was able to keep up with the acquisition and the processing comparison results were impressive. The satellite link costs were nominal.

Improved Technologies

In 1988, a need was expressed to acquire 3-D data over producing marine reservoirs where production and drilling platforms made towing streamers dangerous or impossible. The ocean-bottom cable method, essentially land data acquisition at sea, represented a promising solution. However, data from the method's ocean-bottom hydrophones suffered from ghost reflections from the water surface that accompanied each legitimate reflection wavelet. In water depths greater than about 10 meters, deconvolution is unable to eliminate these ghost reflections. A method was developed that successfully eliminated these ghost reflections. The solution was to record data from both hydrophones and geophones at each receiver station and then to properly combine them in processing. This enhancement eliminated the receiver ghost spectral notch and, therefore, increased the data's bandwidth. With numerous other advantages inherent in the method including surface consistent geometry and elimination of the need for cell flexing, the method has produced imaged data with greater resolution than those produced with towed streamers. Its use has been extended to waters clear of obstacles.

Experiments with time-lapse 3-D, or 4D, surveys began more than 10 years ago. Interest in this technology has recently risen dramatically. Mapping fluid front movements, as well as estimating reservoir properties to refine reservoir simulator models, are aimed at increasing the efficiency with which hydrocarbons are produced. Estimating reservoir properties has also led to a recurrent interest in recording and processing mode-converted shear waves. For reservoirs in marine settings, this interest has also expanded the industry's focus on the ocean-bottom cable method with three-component geophones in addition to hydrophones. Shear waves can't propagate through water.

Finally, a recently published technical article eloquently explains why, in some cases, land vibrators output more energy in the form of harmonics and subharmonics of the pilot sweep than in the form of the sweep signal itself. In the latter part of that article, the author presents the result of using a near trace as a measure of what was actually output by the vibrator and of performing signature deconvolution of the other traces in order to collapse constructively the harmonics into each reflection wavelet. The result was encouraging.