--> The Balance between Geophysics and Geology in Seismic Interpretation, by Alistair R. Brown, #40615 (2010).
[First Hit]

Datapages, Inc.Print this page

Click to view article in PDF format.

 

GCThe Balance between Geophysics and Geology in Previous HitSeismicNext Hit Interpretation*

 

Alistair R. Brown1

 

Search and Discovery Article #40616 (2010)

Posted October 22, 2010

 

*Adapted from the Geophysical Corner column, prepared by the author and entitled "Balance between G&G is Critical," in AAPG Explorer, December, 2005. Appreciation is expressed to Alistair Brown, editor of Geophysical Corner, and to Larry Nation, AAPG Communications Director, for their support of this online version.

 

1Consulting reservoir geophysicist, Dallas, TX ([email protected])

 

Introduction

Everyone is a product of his/her own experience. Hence geophysicists tend to favor geophysical methods and geologists tend to favor geological methods. It is only natural. However, we all know that the search for and development of oil and gas involves both disciplines. We all need to be geoscientists. We all need to meld geophysics and geology in an effective manner. We need to be integrated. I regularly have cause to admonish a course student of mine for "over-geologizing" the solution. He or she imposes a geological concept or model on the Previous HitseismicNext Hit interpretation rather than "letting the data speak." The majority of Previous HitseismicNext Hit interpreters come from a geological background, so they have to learn the intricacies of geophysics. With modern 3-D data there is a lot more to learn than there used to be. We have to appreciate the value of Previous HitseismicNext Hit amplitude, the resolution of the data, how to recognize data defects, the precision of workstation autotrackers, the complexities of Previous HitseismicNext Hit horizon identification, and the mystifying plethora of Previous HitseismicNext Hit attributes.

 

 

Introduction

Figures

Example 1

Example 2

General comments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Introduction

Figures

Example 1

Example 2

General comments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Introduction

Figures

Example 1

Example 2

General comments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Introduction

Figures

Example 1

Example 2

General comments

Figure Captions

Figure 1. Map of turbidite sand made from Previous HitseismicNext Hit amplitude with some interpretive smoothing. Well is located slightly updip of center of sand body.

Figure 2. Map of turbidite sand made from detailed analysis of Previous HitseismicNext Hit amplitude. Note the four “flows” and that the well is now targeting a gap in the turbidite.

Figure 3. Well-to-Previous HitseismicNext Hit tie for a reservoir less than one quarter of a wavelength thick. Note that the top and base amplitudes are outside the reservoir thickness based on an exact depth-to-time tie.

Figure 4. Previous HitPhaseNext Hit Circle showing Previous HitseismicNext Hit character for the top and base reflections of a low impedance layer, such as a hydrocarbon sand.

Example 1

In a prospect in Latin America, a high amplitude was identified as a turbidite sand that had been recognized in a nearby well. The amplitude was mapped manually to indicate a prospective area of useful size. The geologically oriented interpreter made the map look like a turbidite because he had seen one before (Figure 1). He showed the map to his exploration manager and they agreed to drill a well at the location indicated.

A more detailed geophysically oriented analysis of the high amplitude was then undertaken. All the data were used, the tracking was performed with a high-precision autotracker, and the horizon amplitude was extracted to produce a horizon slice. A simplified version of this is shown in Figure 2. Note that the initially proposed drilling location is now found to be in a local amplitude low, indicating that the turbidite sand here is thin or absent. Clearly the well needs to be relocated, as we believe that amplitude is a measure of porosity-thickness.

Example 2

Many Previous HitseismicNext Hit interpreters take a formation top on a log, measured in depth, convert it to time and pick at that exact position on the Previous HitseismicNext Hit section. A simple synthetic seismogram constructed from a sonic log using a Ricker wavelet may be used as an aid. These conventional approaches neglect errors caused by Previous HitseismicNext Hit velocity, hole conditions, tuning effects, data Previous HitphaseNext Hit and data polarity. All these issues, and others, make the Previous HitseismicNext Hit-to-well tie quite complicated.

The velocity function normally comes from a well velocity survey or a vertical Previous HitseismicNext Hit profile in the well being tied or in a nearby well. Errors of measurement and lateral variations mean that the velocity used is rarely exactly correct. The Previous HitphaseNext Hit and the polarity of the data are significant uncertainties, but first let us consider that the data really are zero Previous HitphaseNext Hit and of known polarity.

The reservoir we wish to study in this example has good contrast at top and base; therefore, tying the well should be simple. If the reservoir is less than quarter-wavelength in thickness (as they so often are), the reflections from top and base cannot get close enough together so that they are both mislocated (Figure 3). The top reflection is early and the base reflection is late. The diagram in Figure 3 is drawn for perfect velocity, but we still do not use a simple depth-to-time tie to pick the top of the reservoir on the Previous HitseismicNext Hit data. An understanding of tuning effects tells us we must pick the reservoir top at point P rather than point A. This is not only geophysically correct, but point P is where the amplitude is to be found for studying the internal fabric of the reservoir. The other relevant amplitude for studying this reservoir is from the base reflection at point T.

Let me give an example of how we could over-geologize the interpretation here. We want an amplitude to characterize this reservoir, and we have accurate picks on the top and base of the sand. We spot the top and base on the Previous HitseismicNext Hit, A and B as shown, and then extract a windowed amplitude such as Average Absolute Amplitude or RMS Amplitude over this interval. Note that the window, which exactly ties the reservoir, completely misses the two relevant amplitudes from top and base! The best amplitude to characterize this reservoir is Composite Amplitude, made by absolute value summation of the horizon amplitudes from P and T.

General Comments

This tie is further complicated by data Previous HitphaseNext Hit and polarity. Interpreters should not rely on nominal zero Previous HitphaseNext Hit data being actually zero Previous HitphaseNext Hit. Nor should they rely on the data polarity being "normal" for the region of the world where they are located (American or European). We should check these out by scrutinizing the character of high amplitude reflections whose geological nature we understand. It is the responsibility of every Previous HitseismicNext Hit interpreter today to assess as thoroughly as possible the Previous HitphaseNext Hit and polarity of his or her data before a final well-to-Previous HitseismicNext Hit tie is established. Suppose the high amplitude reflections being used are a gas bright spot. The sand generating the bright spot must then be a low impedance zone relative to the embedding shale, so the relevant Previous HitphaseNext Hit circle is shown in Figure 4. Note that we are considering top and base reflections together. Note also that there are two versions of zero Previous HitphaseNext Hit (American and European polarity), which are opposite from each other. These are equally useful for interpretation, but it is critical to understand which we have. The issue is further complicated by the all-too-common 90-degree Previous HitphaseNext Hit data. Other Previous HitphaseNext Hit conditions also exist.

Knowledge of Previous HitphaseNext Hit and polarity permits us to predict the character of the reservoir top (for example) and thus to make a meaningful well tie. If the determined Previous HitphaseNext Hit is significantly non-zero, then Previous HitphaseNext Hit rotation of the data is probably a good idea. Previous HitPhaseNext Hit rotation is easy; recognition of Previous HitphaseNext Hit errors by today's Previous HitseismicNext Hit interpreters is more difficult. Zero-Previous HitphaseTop data are the easiest to work with, and the co-location of time and amplitude is a big advantage.

Return to top.

 

Copyright © AAPG. Serial rights given by author. For all other rights contact author directly.