Abstract

In most mature hydrocarbon provinces it is estimated that 40 per cent of the oil is found in stratigraphic traps, which are often very difficult to identify on seismic data directly. I suggest that the difficulty is caused by three related problems; deconvolution, multiple removal - especially in marine data and well ties. Each of these problems demands accurate knowledge of the wavelet for its solution. Flaws in the current approach to the problem of wavelet estimation are discussed briefly.

Since seismic waves are causal, the peak energy in the reflected signal arrives later than the reflection time. To identify the event correctly from well logs, it is convenient to process the data to zero-phase, for then the peak in the wavelet occurs at the reflection time. This correction to zero phase can be made only if the wavelet is known.

A related problem is to ensure that the wavelet remains constant across the seismic section. If the shot-to-shot variations in the wavelet are known and removed in processing by wavelet deconvolution, all lateral variations in the seismic data can be attributed to the earth. This process requires the wavelet to be known. If the data have also been processed such that most of the multiple energy has been removed, the reflections are essentially primaries only. With accurate well ties, made using the known zero-phase wavelet, individual stratigraphic units can be identified and followed laterally, and stratigraphic traps can be identified.

The goal is to determine the wavelet independent of the well logs, for then we may check our method using the wells. The synthetic seismogram, calculated using this known wavelet and the logs, should match the seismic data without any time-shifting or frequency-dependent phase shifting. If this can be done on a regular basis, the method can be relied upon to reveal the stratigraphy away from well control.

The first step in achieving this goal is to measure the source signature. Published methods for marine sources, dynamite, and Vibroseis are presented using examples from real data. Normally the source signature is not measured. However, existing data can be calibrated, and the source signature determined, by resurveying part of the existing survey, only this time measuring the source signature. The wavelet can then be computed from the source signature by including source and receiver ghosts and a filter to account for absorption. As a first step in processing, this wavelet should be compressed, by wavelet deconvolution, to a shorter signal with approximately the same amplitude spectrum, only smoother. The second step is to remove the multiples.

The principal source of multiple energy is the free surface of the earth. If the free surface were absent, the upgoing waves would not be reflected back down again. A wave theoretical method now exists to remove all the effects of the free surface in marine seismic data. It requires the source signature to be known. The method is illustrated with an example from real data.

Since the wavelet must be known throughout the processing sequence, no processes should be applied to the data that apply unpredictable changes to the wavelet, such as short gap predictive deconvolution. If seismic data processing has succeeded in achieving an accurately migrated section of primary reflections, a good tie can be expected with a synthetic seismogram calculated simply by convolving the wavelet with the reflectivity. Equally, the raw shot records at the well should match the offset-dependent synthetic seismogram calculated with the known source signature and the earth layer parameters known from the well logs. This is a much more rigorous test of the whole process and is illustrated with an example.

These methods can be applied to the deliberate search for stratigraphic traps. However, the interpretation of the geophysical data can proceed only if there is already in place a rigorous stratigraphic framework within which to work. An example from the Jurassic of the North Sea is used to illustrate this.

End

     

About the Author(s)

Anton Ziolkowski was educated at the University of Cambridge, England, where he received a BA in Engineering and Natural Sciences in 1968, and a PhD in Geophysics in 1971 for his thesis on the theory of air guns. In 1971-1973 he was with the Seismic Discrimination Group of Lincoln Laboratory, Massachusetts Institute of Technology (MIT), where his colleagues taught him the fundamentals of time series analysis and global seismology. In 1973 he briefly changed direction and studied at the London School of Economics (LSE), earning a MSc (Econ.). LSE introduced him to the philosophy of science, particularly that of Karl Popper.

In 1976 he returned to geophysics when he joined the U.K. National Coal Board as Headquarters Geophysicist, responsible for the technical control, development and administration of a large seismic exploration program for new and existing coal mines. Significant developments in high resolution reflection seismology resulted from this program.

In 1980-82 he was an independent research consultant to the British National Oil Corporation and worked primarily on marine seismic sources and deconvolution; the work resulting in a major patent for measurement and control of marine seismic sources.

In 1982 he became Professor of Applied Geophysics at Delft University of Technology, Delft, The Netherlands. He also became a leading critic of the statistical approach to deconvolution in the processing of seismic data, and has worked to provide an alternative approach, based on measurements of the seismic source signature. He conducted several large experiments to establish methods for the determination of the source signature of the most important commercial seismic sources: marine air guns and water guns, land and marine vibrators, and land dynamite.

He jointly set up a cross-well experiment in the Groningen gas field in 1990, demonstrating that Q is at least 1,000 in certain North Sea sedimentary rocks. He initiated the real-time monitoring of laboratory high-pressure hydraulic fracturing demonstration experiments using ultrasonic piezo-electric transducers.

In 1992 he left the Netherlands to take up his present appointment as the Petroleum Science and Technology Institute (PSTI) Professor of Petroleum Geoscience at the University of Edinburgh where he is Head of the Predictive Geoscience Research Unit, with responsibility to lead research into methods to aid exploration and production of hydrocarbons. In this capacity he initiated a joint project with SSL/GECO-Prakla to acquire and process 500 km of seismic data intersecting over 20 wells in the Inner Moray Firth, using continuous measurements of the source signature as the input for deconvolution, well ties and stratigraphic correlations. He is also the leader of a major European Community-funded project on the joint inversion of seismic and transient EM data for direct hydrocarbon detection and monitoring of fluid flow in reservoirs.

In 1982 he received the Conrad Schlumberger Award of the EAEG; from 1987-1990 he was a visiting scientist at the Department of Earth, Atmospheric and Planetary Sciences, MIT; in 1990 he was recipient of a $25,000 Research Grant from Schlumberger Doll Research; in 1991 he was elected a member of Hollandse Maatschaapij der Wetenschappen (Dutch Society of Scientists); and in 1993 he was the Keynote Speaker at the CSEG Workshop. He has been invited to give training courses on seismic sources and deconvolution by oil companies in Europe, Asia and North America, sometimes as an IHRDC instructor, and for the last few years has taught at Institut Francais du Petrole. He has five U.S. patents. He is a member of SEG, EAEG, ASEG, and is a Fellow of the Royal Astronomical Society.

He is author of the IHRDC book "Deconvolution" and co-author with Guido Baeten of the Elsevier book "The Vibroseis Source".

References

Appendices

Join the Conversation

Interested in starting, or contributing to a conversation about an article or issue of the RECORDER? Join our CSEG LinkedIn Group.

Share This Article

Related Reading

Sorry, no results were found.