You are on page 1of 12

GEOPHYSICS, VOL. 64, NO. 2 (MARCH-APRIL 1999); P. 609620, 15 FIGS.

Tutorial 3-D seismic stratal-surface concepts applied to the interpretation of a uvial channel system deposited in a high-accomodation environment

Bob A. Hardage and Randy L. Remington

ABSTRACT

A fundamental thesis of seismic stratigraphy is that seismic reections follow impedance contrasts that coincide with stratal surfaces, which are surfaces where depositional processes occur at a xed moment in geologic time. This stratal-surface concept is used in this paper to image a narrow (width 300 ft or 90 m), thin, uvial channel system that is embedded within a seismic reection peak that reects from a large (about 2 2 mi or 3.2 3.2 km) area of nonchannel facies that dominate the waveshape of the reection peak. The targeted channel facies are conned to an interval that vertically spans less than 30 ft. According to principles of seismic stratigraphy, four conformable seismic stratal surfaces that pass through the interior of this channel sequence were constructed across a 2 2-mi (3.2 3.2-km) (approximately) area of a 3-D seismic-data volume. The channel images portrayed on these seismic horizons, which were spaced at vertical increments of 2 ms, illustrate the principle that seismic attributes viewed on seismic stratal surfaces provide

valuable images of facies distributions within thin-bed sequences and help seismic interpreters segregate channel facies from nonchannel facies. These stratal-surface interpretations of the uvial system were then integrated to create a thin (8-ms-thick), stratal-bounded seismic-data window that spans the short geologic time period during which the targeted uvial depositional system was active. Seismic-attribute calculations within this stratal-bounded analysis window improve parts of the channel image by integrating the facies information from all stratal surfaces into a unied seismic attribute that vertically spans the total depositional sequence. A comparison is made between channel images on seismic stratal surfaces that are conformable to two different reference surfaces, one reference surface being positioned below the targeted uvial system and the second reference surface being above the thin-bed channels. This comparison supports the premise that seismic interpreters should extrapolate stratal surfaces both upward and downward across a thin-bed target to optimize the image of that target.

INTRODUCTION

The term accommodation space refers to that volume of a depositional basin that lies below base levela surface above which erosion occurs and below which deposition occurs. A high-accommodation depositional environment is a basin setting in which the space available for sediment to accumulate, whether because of basin subsidence or sea-level rise, always balances or exceeds the volume of sediment in-

put. High-accommodation sediments, which are rarely subjected to erosion, therefore, tend to provide excellent records of depositional processes and environments because once sediment layers are deposited in such an environment, they tend to be altered and modied only by normal compaction and diagenetic processes. Holocene- to Miocene-age rocks of the Gulf of Mexico are excellent examples of high-accommodation depositional environments. The 3-D seismic-data examples illustrated in this paper come from this very basin setting, the

Manuscript received by the Editor July 24, 1998; revised manuscript received October 22, 1998. Bureau of Economic Geology, The University of Texas at Austin, University Station, Box X, Austin, Texas 78713-7508. E-mail: hardageb@begv.beg. utexas.edu, remingtonr@begv.beg.utexas.edu. c 1999 Society of Exploration Geophysicists. All rights reserved. 609

610

Hardage and Remington HORIZON VIEW OF TARGETED FLUVIAL SYSTEM

3-D survey area being located only a short distance northwest of Corpus Christi, Texas. A stratal surface is a depositional bedding plane, that is, a depositional surface that denes a xed geologic time. Any siliciclastic rock deposited in a high-accommodation environment contains numerous vertically stacked stratal surfaces. A fundamental thesis of seismic stratigraphy is that a seismic reection event follows the impedance contrast associated with a stratal surface, that is, a surface that represents a xed point in geologic time (Mitchum et al., 1977; Vail and Mitchum, 1977). Because lithology varies across the area spanned by a large depositional surface, the implication is that an areally pervasive seismic reection event does not necessarily mark an impedance contrast boundary between two xed rock types as the reection traverses a prospect area. The application of this fundamental concept about the genetic origin of seismic reections to seismic interpretation is referred to as stratal-surface seismic interpretation. Tipper (1993) illustrated and discussed situations in which a thin-bed seismic reection can be either chronostratigraphic or diachronous, depending on (1) the vertical spacings between beds, (2) the lateral discontinuity between diachronous beds, and (3) bed thickness. The conclusion that a thin-bed seismic reection is chronostratigraphic or diachronous thus needs to be reached with caution because the answer depends on the local stratigraphy, the seismic bandwidth, and the horizontal and vertical resolution of the seismic data. If two seismic reection events, A and B, are separated by an appreciable seismic-time interval (say a few hundred milliseconds) yet are conformable to each other, then the uniform seismic-time thickness between these two events represents a constant and xed period of geologic time throughout the seismic image space spanned by reectors A and B. An implication of seismic stratigraphy that can be invoked in such an instance is that any seismic surface intermediate to A and B, which is also conformable to A and B, is also a stratal surface. The purpose of this paper is to illustrate that this stratal-surface seismic-interpretation concept becomes valid when the principle is applied to the interpretation of a complex channel system deposited in a high-accommodation environment. The channel system used to illustrate the stratal-surface concept of seismic interpretation is a shallow Miocene uvial system at a depth of approximately 2200 ft (670 m). The 3-D seismic acquisition geometry was designed to produce 55 55-ft (17 17-m) stacking bins and a stacking fold of 12 at this target depth. The signal-to-noise character of these rather low-fold 3-D data is quite good, and the stratal-surface seismic-interpretation approach described in this paper produces excellent images of the channel complex.
COLOR BAR USED FOR DATA DISPLAYS

A display of seismic reection-amplitude strength across the targeted stratal surface is shown in Figure 1. This image corresponds to the geologic time during which the targeted Miocene uvial system was active, and the meander patterns revealed in the lower-right and upper-right quadrants of the image dene the geometric shapes and spatial positions of the channel complexes that are to be analyzed. The procedure used to construct this attribute display will be explained later. For the time being, this seismic image will be used only to dene the coordinates of seismic proles that cross selected features of the channel systems. These proles can then, in turn, be used to illustrate the subtle changes in reection waveshape that mark the transition from nonchannel to channel facies. Crossline proles 174, 200, and 222 (Figure 1) are the key vertical slices through the 3-D seismic volume that will be used to illustrate the principles of stratal-surface interpretation. The east-west-trending ripplelike pattern in the south half of the image is not caused by a depositional process. Rather, it is the result of the seismic analysis window cutting through lowamplitude, zero-crossing regions of the reection waveforms in a way that causes the reection amplitude to shift systematically from a small positive value to a small negative value. The ripple effect can be altered by adjusting the four-tone color bar used to display the data.
VERTICAL VIEWS OF CHANNEL AND NONCHANNEL SEISMIC FACIES

Crosslines 174, 200, and 222 (Figure 1) will be used to show the subtle changes in reection waveform that distinguish

A four-tone (black, dark gray, light gray, and white) color bar is used to display seismic reection amplitudes in this paper. The best way to dene the relationships between seismic reection amplitude values and the color tones is to refer to the wiggle-trace displays in Figures 2 through 9. In these displays, large positive reection amplitudes are black, small positive amplitudes are dark gray, small negative amplitudes are light gray, and large negative amplitudes are white. This same fourtone color bar is used to display seismic reection amplitudes in map view in Figures 1 and 10 through 14.

FIG. 1. 3-D seismic image of the targeted thin-bed Miocene uvial channel systems. This image is a display of reection amplitudes on a stratal surface 26 ms below, and conformable to, reference surface 2 dened in Figure 15.

Stratal-Surface Interpretation

611

seismic channel facies from seismic nonchannel facies in this particular prospect area. These seismic facies will be viewed by examining the reection-amplitude response across a series of closely stacked stratal surfaces that collectively span the depositional time during which the channel system was active.

Crossline 174 Crossline 174 crosses two separate Miocene channel systems, a narrow feature spanning inline coordinates 240245 and a wider feature between inline coordinates 6480 (Figure 1). Because the stacking bins created by this 3-D recording geometry measure 55 55 ft (17 17 m), these coordinate ranges mean that the smaller channel system is about 300 ft (90 m) wide and the larger channel complex is about 900 ft (275 m) wide at the point where crossline 174 crosses the feature. That part of crossline 174 that images the narrow channel is displayed in Figure 2; the part that traverses the wider channel is shown in Figure 3.

Only 200 ms of data centered about the targeted channel system are displayed in these gures, as will be the case for crosslines 200 and 222 that follow. These wiggle-trace displays are shown in a highly magnied format to allow subtle variations in reection waveshape to be seen easily. Every second trace is plotted to avoid trace overlap so that individual reection waveshapes can be inspected independently. Because of this trace decimation, the data appear to have a trace spacing of 110 ft (34 m), whereas the actual trace spacing in the 3-D data volume was 55 ft (17 m). The wiggle-trace data are shown with a high display gain to ensure that low-amplitude features can be seen; some high-amplitude troughs are consequently clipped. A good-quality reection peak, the reference surface, is labeled in each display. This particular reection peak satises the fundamental criteria required of a reference stratal surface used to study thin-bed sequences such as this uvial channel system, namely (1) the event extends over the total 3-D image space and has a high signal-to-noise character, (2) the event is reasonably close to the targeted thin-bed sequences (within

FIG. 2. Data window from crossline 174 centered on the northern, smaller channel system. Refer to Figure 1 for location of this seismic prole. The targeted uvial system is embedded in the reection peak that occurs at 0.70 s at inline coordinate 270. The horizon labeled reference surface is the reference seismic stratal surface that is used to construct additional stratal surfaces that pass through the targeted thin-bed interval.

612

Hardage and Remington

100 ms in this example), and (3) the event is conformable to the targeted thin-bed sequence. Criterion 3 is the most important requirement for any seismic stratal surface that is to be used as a reference surfaceone from which constant-depositional-time surfaces are to be made that span targeted thin-bed sequences. Because the surface labeled reference surface follows the apex of an areally continuous reection peak, the basic premise of seismic stratigraphy is that this reference surface follows an impedance contrast that coincides with a stratal surface. The displays of crossline 174 are repeated in Figures 4 and 5, with four conformable surfaces, A, B, C, and D, that pass through the targeted thin-bed interval added to the proles. These four surfaces are, respectively, 92, 90, 88, and 86 ms aboveand conformable tothe reference surface. Visual inspection of the reection events above and below surfaces A, B, C, and D shows that all of these reection peaks and/or troughs are reasonably conformable to the reference surface event. Surfaces A, B, C, and D can thus be assumed to be stratal surfaces, or constant-depositional-time surfaces, because they are conformable to a known stratal surface (the reference surface) and are embedded in a seismic window in which all re-

ection events are approximately conformable to the selected reference surface. The circled features in Figures 4 and 5 identify locations where stratal surfaces A, B, C, and D intersect obvious variations in reection waveform. These waveshape changes are the critical seismic attributes that distinguish channel facies from nonchannel facies, as can be veried by comparing the inline coordinates spanned by the circled features with the inline coordinates where crossline 174 intersects channel features in Figure 1. Crossline 200 Crossline 200 traverses the northern, smaller channel system at a rather oblique angle and then transects at least three loops of the southern, larger channel system (Figure 1). A 200ms-wide window extending along this crossline prole is displayed in Figure 6. The event labeled reference surface is the same reference surface labeled on crossline 174 (Figures 2 and 3) when that surface is extended to crossline 200. Ideally a reference surface should be smooth. The ripples on this particular surface, such as the three peaklike distortions between inline

FIG. 3. Data window from crossline 174 centered on the southern, larger channel system. Refer to Figure 1 for location of this prole. The targeted uvial system is embedded in the reection peak that occurs at 0.73 s at inline coordinate 120. The horizon labeled reference surface is the reference seismic stratal surface that is used to construct additional stratal surfaces that pass through the targeted thin-bed interval.

Stratal-Surface Interpretation

613

coordinates 125 and 145, are vertical displacements of only one time sample point (2 ms). The expanded time scale used in this wiggle-trace display magnies these small irregularities. The same prole is shown in Figure 7, with stratal surfaces A, B, C, and D included in the display. Again these surfaces are, respectively, 92, 90, 88, and 86 ms above, and conformable to, the reference surface. The circled data windows highlight the changes in reection waveform that correspond to the channel facies that are traversed in the southern channel loops. (Refer to Figure 1 for the channel positions on crossline 200.) Two of these channel-facies intervals are included in the circular data window that spans inline coordinates 110130; only one channel facies occurs in the window centered on inline coordinate 80. As stated in the discussion of crossline prole 174, surfaces A, B, C, and D are stratal surfaces that show areal distributions of channel and nonchannel facies at four distinct, constant depositional times during the genesis of the channel system.

Crossline 222 Crossline 222 extends across two point-bar segments of the southern uvial channel system (Figure 1). Part of the prole that is centered on these two uvial features is displayed in Figure 8 to show the position of the reference surface when that seismic surface is extended to crossline 222. Following the previous format used for crosslines 174 and 200, the subsequent gure (Figure 9) shows the positions of stratal surfaces A, B, C, and D that are conformable to, and 92, 90, 88, and 86 ms above, respectively, this reference surface. The two highlighted data windows in this display emphasize the variations in reection waveform that occur when the prole crosses the two point-bar complexes. The inline coordinates on which these circular data windows are centered correspond to the coordinate positions of the point-bar meanders in Figure 1.

FIG. 4. Data window from crossline 174 showing stratal surfaces that traverse the northern, smaller channel system. This data window is the same as the one shown in Figure 2. Surfaces A, B, C, and D are reasonable approximations of surfaces of constant depositional time because (1) they are conformable to the reference surface that follows a seismic stratal surface and (2) they are embedded in a 200-ms data window in which all seismic reection events are approximately conformable. The highlighted data window encircles the subtle changes in reection waveform that identify the seismic channel facies. The inline coordinates at the center of this circular data window correspond to the position of the channel image in Figure 1.

614

Hardage and Remington

STRATAL-SURFACE IMAGES OF CHANNEL SYSTEMS

The reection-amplitude behavior across stratal surfaces A, B, C, and D is shown as Figures 10, 11, 12, and 13, respectively. The vertical time separation between these successive stratal surfaces is 2 ms, which corresponds to a vertical depth separation of only 67 ft (1.82.1 m) for the low-velocity rocks where this uvial deposition is preserved. As these amplitude-based images are viewed in sequence, starting with surface A and ending with surface D, the effect is approximately the equivalent of stripping off successive layers of deposition, each layer 67 ft (1.82.1 m) thick. The total isopach interval dened by the four stratal surfaces is approximately 2530 ft (7.59 m) thick. That the two channel systems (the northern one and the southern one) are not completely imaged on any of these four horizons implies that none of the four surfaces is a perfect stratal surface. That considerable portions of the channels appear on each surface, however, is evidence that each horizon is a reasonably good approximation of a stratal surface. Surface B (Figure 11) is probably the best approximation of a stratal surface that coincides with the deposition time during which both channel systems were active. Surface A (Figure 10)

loses some of the detail that is imaged on surface B. Although surface C (Figure 12) fails to image parts of the prominent meander loops of the southern system, it shows new channels coming into the complex from the southeast corner of the image space. Surface D is the best image of the downstream part of the southern channel (and splays?) that extends southwest from the meander-loop area. In summary, each surfaceA, B, C, and Dshows a slightly different geometric shape and spatial position of selected parts of the channel complex, in effect showing the genetic growth and lateral movement of the uvial environment as approximately 2530 ft (7.59 m) of sediment was deposited.
STRATAL-BOUNDED SEISMIC ANALYSIS WINDOWS

The channel-system images shown in Figures 1013 are surface-based images; that is, the seismic attribute that is displayed is limited to a data window that vertically spans only one data sample. When a 1-point-thick data window is a good approximation of a stratal surface that passes through the interior of a targeted thin-bed sequence, then seismic attributes dened on that surface can be important depictions

FIG. 5. Data window from crossline 174 showing stratal surfaces that traverse the southern, larger channel system. This data window is the same as the one shown in Figure 3. Surfaces A, B, C, and D are reasonable approximations of surfaces of constant depositional time, as explained in Figure 4. The inline coordinates at the center of the circular data window correspond to the position of the channel in Figure 1.

Stratal-Surface Interpretation

615

FIG. 6. Data window from crossline 200. Refer to Figure 1 for location of this prole. The targeted uvial systems are embedded in the reection peak that occurs at 0.72 s at inline coordinate 200. The horizon labeled reference surface is the reference seismic stratal surface that is used to construct additional stratal surfaces that pass through the targeted thin-bed interval.

FIG. 7. Data window from crossline 200 showing stratal surfaces that traverse channel facies. This prole is the same as the one shown in Figure 6. Surfaces A, B, C, and D are reasonable approximations of surfaces of constant depositional time because (1) they are conformable to the reference surface that follows a seismic stratal surface and (2) they are embedded in a 200-ms data window in which all seismic reection events are approximately conformable. The highlighted data windows encircle the subtle changes in reection waveform that identify the seismic channel facies. The inline coordinates inside these circular data windows correspond to the positions of the meander loops in Figure 1.

616

Hardage and Remington

of facies distributions within the sequence, as the images in Figures 1013 conrm. An alternate, and usually more rigorous, way of determining facies distributions within a thin-bed sequence is to calculate seismic attributes in a data window that spans several data points vertically yet is still conned (approximately) to only the thin-bed interval. The bottom stratal surface of this data window must reasonably coincide with the onset depositional time of the sequence, and the top stratal surface must be a good approximation of the shutoff depositional time for the sequence. Such a data window is a stratal-bounded seismic analysis window. Stratal surfaces A and D shown in section view in Figures 4, 5, 7, and 9 and then in map view in Figures 10 and 13 are examples of surfaces that dene a stratal-bounded data analysis window that spans a targeted thin-bed sequence, specically the thinbed uvial channel system that is the interpretation objective of this study. In this instance, the analysis window is 4 data points (8 ms) thick. As previously stated in the discussions of Figures 1013, surfaces A, B, C, and D are good-quality stratal surfaces because each horizon images a signicant part of the thin-bed uvial system. Because each of these four seismic horizons is a good approximation of a constant-depositional-time surface, collectively the four surfaces are a good representation

of the facies distribution within the total thin-bed sequence that they span. One way to evaluate facies-sensitive seismic information spanned by surfaces A and D is to calculate some type of averaged reection-amplitude attribute in each stacking bin of the 4-point-thick data analysis window bounded by horizons A and D. For example, the average peak amplitude between A and D is displayed in Figure 14 and shows an alternate image of the total channel system. Specically, the image shows more detail about the meandering channel in the extreme southeast corner of the image space, identifying how that channel trends to the northwest and blends into the larger uvial system. This southeast part of the channel system is not obvious on stratal surfaces A or B (Figures 10 and 11). Although the system in the southeast corner of the image is detectable on stratal surfaces C and D (Figures 12 and 13), it is not seen in the same degree of detail as in Figure 14. An improved image of this particular part of the thin-bed system appears in Figure 14 because the stratal-bounded data window dened by surfaces A and D is a good approximation of the depositional-time interval when that part of the uvial system was active, whereas no individual stratal surfaceA, B, C, or Dis a good approximation of a depositional-time surface that passes through this particular part of the thin-bed sequence.

FIG. 8. Data window from crossline 222 centered on selected point-bar facies. Refer to Figure 1 for location of this prole. The targeted uvial system is embedded in the reection peak that occurs at 0.72 s at inline coordinate 150. The horizon labeled reference surface is the reference seismic stratal surface that is used to construct additional stratal surfaces that pass through the targeted thin-bed interval.

Stratal-Surface Interpretation

617

In contrast, the average peak reection amplitude in the stratal-bounded window (Figure 14) does not image the northern channel system as well as do stratal surfaces A, B, and C (Figures 1012). Other attributes, such as the average instantaneous frequency, could be calculated in this stratal-bounded window and evaluated as options for imaging the channel system. These alternate imaging attributes will not be used here because our purpose is simply to illustrate that the stratalsurface concept and the stratal-bounded window technique have relative advantages and disadvantages and should always be used in a thin-bed interpretation, their results being integrated into a unied model of the depositional system.
COMBINING UPWARD AND DOWNWARD EXTRAPOLATIONS OF SEISMIC STRATAL SURFACES TO IMPROVE THIN-BED INTERPRETATION

The channel images shown up to this point were created by interpreting a reference stratal surface below the targeted thinbed sequence and then extending conformable stratal surfaces upward until they passed through the targeted channel system. In this instance, the reference surface is approximately 90 ms below the channel sequence that is to be interpreted (Figures 4, 5, 7, 9). In challenging thin-bed interpretations such as the uvial channel system considered here, it is important to dene two

seismic reference surfaces that bracket the thin-bed system that is to be interpreted, one reference surface being below the interpretation target and the second reference surface being above the target. By creating conformable reference stratal surfaces above and below a thin-bed system, we can extrapolate conformable seismic stratal surfaces from two directions to sweep across a thin-bed target. One set of seismic stratal surfaces is commonly a better approximation of constant-depositional-time surfaces within the targeted thinbed sequence than the other set is and produces more accurate images of facies patterns within the thin-bed unit. To illustrate the advantage of this opposite-direction convergence of seismic stratal surfaces onto a thin-bed target, a second reference surface was interpreted above (and, in this case, closer to) the targeted uvial system. Specically, this second stratal reference surface followed the apex of the reection troughs immediately above the thin-bed channels (Figure 15). The uvial system is about 2430 ms below this second reference surface. The reection-amplitude response across the channel systems observed on a stratal surface 26 ms below and conformable to this overlying reference surface is displayed in Figure 1. The improved channel image in this case occurs because stratal surfaces that are conformable to the overlying seismic stratal surface are better approximations of constantdepositional-time surfaces for this channel system than are

FIG. 9. Data from crossline 222 showing stratal surfaces that traverse point-bar facies. This data window is the same as the one shown in Figure 8. Surfaces A, B, C, and D are reasonable approximations of surfaces of constant depositional time, as explained in Figures 4 and 7. The inline coordinates at the centers of these circular windows correspond to the position of point bars in Figure 1.

618

Hardage and Remington

FIG. 10. Reection-amplitude behavior on stratal surface A, which is 92 ms above, and conformable to, the selected reference surface. A four-tone color bar is used to display reection amplitude and polarity (see text). Locations of crosslines 174, 200, and 222 are indicated. That a large part of the two uvial systems is depicted on this horizon is evidence that surface A is a reasonably good approximation of a constant-depositional-time surface.

FIG. 12. Reection-amplitude behavior on stratal surface C, which is 88 ms above, and conformable to, the selected reference surface. This surface, approximately 6 ft (2 m) below, and conformable to, stratal surface B (Figure 11), shows new channel systems in the extreme lower-right quadrant of the image space.

FIG. 11. Reection-amplitude behavior on stratal surface B, which is 90 ms above, and conformable to, the selected reference surface. This surface, approximately 6 ft (2 m) below, and conformable to, stratal surface A (Figure 10), shows a slightly improved image of the channel systems. A four-tone color bar is used to display reection amplitude and polarity (see text).

FIG. 13. Reection-amplitude behavior on stratal surface D, which is 86 ms above, and conformable to, the selected reference surface. This surface, approximately 6 ft (2 m) below, and conformable to, stratal surface C (Figure 12), provides a better image of the uvial system extending toward the lower-left corner of the image space.

stratal surfaces that are conformable to the deeper reference surface. This result illustrates that upward and downward extrapolations of conformable stratal surfaces across a thin-bed target are a recommended interpretation procedure, especially in those instances when valid stratal reference surfaces can be interpreted both above and below the targeted thin-bed
CONCLUSIONS

FIG. 14. Average amplitude of reection peaks in a thin, stratal-bounded window spanning the targeted Miocene uvial depositional sequence. The top boundary of the window is stratal surface A; the bottom boundary is stratal surface D.

The interpretation of thin-bed reservoirs in 3-D seismic data volumes can be achieved by (1) interpreting a reference surface that is conformable to the areal geometry of the thin-bed sequence and (2) creating seismic stratal surfaces conformable to this reference surface that pass through the thin-bed target. If the seismic stratal surfaces constructed according to this logic are satisfactory approximations of constant-depositionaltime surfaces that existed during the deposition of the thin-bed sequence, then seismic attributes across these stratal surfaces can be valuable indicators of facies distributions within the sequence. Evidence that this stratal-surface approach to 3-D seismic thin-bed analysis is a robust interpretation technique is provided by the uvial channel images described in this paper. A previous publication supports the same conclusion (Hardage et al., 1994). An expanded application of this stratal-surface concept is to calculate seismic attributes inside a thin, stratal-bounded analysis window that is centered vertically on the thin-bed target. Facies-sensitive attributes extracted from carefully constructed stratal-bounded windows are commonly better indicators of facies distributions within a thin-bed target than are attributes that are restricted to a 1-point-thick stratal surface that passes through the target. This fact implies that the geologic time interval during which a thin-bed sequence is deposited can sometimes be more accurately portrayed by a stratal-bounded

FIG. 15. Location of reference surface 2 on crossline 200. Reference surface 1 is the horizon labeled reference surface in the displays of crossline 200 in Figures 6 and 7. Reference surface 2 is an alternate seismic stratal surface positioned above the thin-bed target. The positions of the thin-bed channels on this prole are dened by the circular data windows in Figure 7.

620

Hardage and Remington

data window than a xed geologic time during the thin-bed deposition can be approximated by a 1-point-thick seismic stratal surface. The uvial channel images included in this paper demonstrate that a rigorous 3-D seismic thin-bed interpretation should be based on attributes extracted from both stratal surfaces and stratal-bounded windows. This two-pronged approach to thin-bed interpretation can be further strengthened by extrapolating seismic stratal surfaces and stratal-bounded windows onto the thin-bed target from opposite directions, that is, from both below and above the thinbed target. The logic in this dual-direction extrapolation is that one of the seismic reference surfaces is generally more conformable to the thin-bed sequence than is the other reference surface, and this improved conformability leads to improved attribute imaging of facies distributions within the thin bed.
ACKNOWLEDGMENTS

acterization of Complex Reservoirs, to illustrate seismicinterpretation principles that need to be implemented in order to characterize complex thin-bed reservoirs. The seismic interpretation was done using software technology provided by Landmark Graphics Corporation under its University Grant Program.
REFERENCES Hardage, B. A., Levey, R. A., Pendleton, V., Simmons, J., and Edson, R., 1994, A 3-D seismic case history evaluating uvially deposited thin-bed reservoirs in a gas-producing property: Geophysics, 59, 16501665. Mitchum, R. M., Jr., Vail, P. R., and Thompson, S., III, 1977, Seismic stratigraphy and global changes in sea level, part 2, the depositional sequence as a basic unit for stratigraphic analysis, in Payton, C. E., ed., Seismic stratigraphyApplications to hydrocarbon exploration: Am. Assn. Petr. Geol. Memoir 26, 5362. Tipper, J. C., 1993, Do seismic reections necessarily have chronostratigraphic signicance?: Geology, 130, 4755. Vail, P. R., and Mitchum, R. M., Jr., 1977, Seismic stratigraphy and global changes of sea level, part 1, overview, in Payton, C. E., ed., Seismic stratigraphyApplications to hydrocarbon exploration: Am. Assn. Petr. Geol. Memoir 26, 5152.

The current investigation was done as a part of the industry funded research project, Seismic Vector-Waveeld Char-