You are on page 1of 18

GEOPHYSICS, VOL. 68, NO. 1 (JANUARY-FEBRUARY 2003); P. 4057, 25 FIGS. 10.1190/1.

1543193

Tutorial Converted-wave seismic exploration: Applications

Robert R. Stewart , James E. Gaiser , R. James Brown , and Don C. Lawton


P-S data that include imaging reectors within and beneath gas-bearing sediments; detailed shallow imaging and enhanced fault mapping; imaging shale diapirs, mud volcanoes, and targets beneath salt bodies and basalt layers; providing another section with potentially different reectivity, multiples, and tuning; imaging interfaces with low P-wave impedance contrast but signicant S-wave impedance change; using P-S attributes and internal V P /VS analysis for lithology discrimination (e.g., sand/shale, dolomite/anhydrite); augmenting conventional amplitude variation with offset (AVO) analysis to determine density and velocity; calibrating P-wave bright spots; investigating anisotropy for improved processing as well as for revealing fracture density and orientation; and monitoring reservoir changes (time-lapse or 4D analysis). The 2000 SEG-EAGE Summer Research Workshop (MacBeth et al., 2001) conducted a poll to evaluate its attendees assessment of the above applications (see http://www.seg.org/ meetings/past/srwboise2000/poll.html). Conference participants considered most to be proven or possible. The September, 2001 issues of the SEGs Leading Edge and the Canadian SEGs Recorder had special sections on multicomponent seismic activities and their application. In addition, the SEG and EAGE sponsored an excellent Distinguished Instructor Short Course (Thomsen, 2002) that considered a number of the anisotropic aspects of converted waves and how they could be applied. Let us look in more detail at a sample of these applications. First, we provide a selection of imaging cases, then lithology estimates, followed by anisotropy analyses, then uid description examples, and nally examples of reservoir monitoring surveys.
IMAGING

ABSTRACT

Converted seismic waves (specically, downgoing P-waves that convert on reection to upcoming S-waves) are increasingly being used to explore for subsurface targets. Rapid advancements in both land and marine multicomponent acquisition and processing techniques have led to numerous applications for P-S surveys. Uses that have arisen include structural imaging (e.g., seeing through gas-bearing sediments, improved fault denition, enhanced near-surface resolution), lithologic estimation (e.g., sand versus shale content, porosity), anisotropy analysis (e.g., fracture density and orientation), subsurface uid description, and reservoir monitoring. Further applications of P-S data and analysis of other more complicated converted modes are developing.

INTRODUCTION

A great richness of wave types propagate in an exploration seismic survey; hopefully, they encounter a similar wealth of resources. Measurement of the full ground motion (by multicomponent seismic sensors) excited by a seismic source, combined with analysis to unravel the various wave types, is providing some remarkable new images. The converted-wave (P-S) method uses P energy propagating downward, converting upon reection to an upcoming S-wave. P-S analysis can provide improved subsurface images as well as give a measure of S-wave properties relating to rock type and saturation. P-S surveys are a relatively inexpensive (compared to S-S measurement), broadly applicable, and effective way of obtaining S-wave information. If we do have P-S reectivity, what can it be used for? Various authors (e.g., Kristiansen, 2000; Yilmaz, 2001) have suggested or shown a number of applications of

Seeing through gas-charged sediments P-wave energy is delayed, scattered, and attenuated when passing through a gas-bearing sediment. Typically, gas

Manuscript received by the Editor March 29, 2001; revised manuscript received July 3, 2002. University of Calgary, Department of Geology and Geophysics, 2500 University Drive N.W., Calgary, Alberta, T2N 1N4 Canada. E-mail: stewart@geo.ucalgary.ca; brown@geo.ucalgary.ca; donl@geo.ucalgary.ca. WesternGeco, 1625 Broadway, Denver, Colorado 80202. E-mail: james.gaiser@westgeo.com. c 2003 Society of Exploration Geophysicists. All rights reserved. 40

Converted-wave Seismic Exploration

41

saturation will affect compressibility strongly, but rigidity to a lesser degree; thus, the oft-observed P-wave sensitivity and S-wave insensitivity to gas saturation. Whether in the near surface or just above the reservoir, gas-charged sediments can seriously degrade P-wave imaging of deeper features. Some areas in the North Sea have near-surface channels apparently containing biogenic gas that saturates poorly consolidated sediments (Figure 1). Although these near-surface hydrocarbon deposits are not economic, they are of interest because of their impact on conventional P-wave images

and sediment stability. Reections from interfaces within and below the gas-charged channels are poor on the P-wave section (Figure 1a), exhibiting reverberations and attenuation of high frequencies. In contrast, the P-S section (Figure 1b) delineates the channel bases, interfaces below the channels, and even sediment boundaries within the channels. Leaky gas reservoirs can create a gas plume or chimney that makes conventional P-wave imaging and characterization of the reservoir very difcult. S-waves, being generally less sensitive to rock saturants, can be used to penetrate gassaturated sediments. The sub-sea seismic (SUMIC) technique used three-component (3C) geophones planted on the ocean bottom (Berg et al., 1994) that recorded data from which highquality P-S images were constructed. Examples of this imaging through a gas chimney (Granli et al., 1995) led to considerable excitement about marine shooting with ocean-bottom cables (OBC) and sea-bottom seismometers. Rodriguez (2000) analyzed a four-component (4C) case from Valhall eld, Norway (originally conducted by Amoco Norway Oil Co. and the Valhall Licence partners). He used prestack equivalent-offset migration for converted waves (Bancroft, 2000) to image through a gas cloud. The results provided a more interpretable image of the chalk reservoir beneath the gas cloud, especially near its anticlinal top at about 2.8 s on the P-wave section (Figure 2). Li et al. (2001) also analyzed the Valhall data set. They noted more focussed and continuous reectors by using prestack versus poststack P-S migrations. They concluded that converted-wave images provided a better image under gas clouds than P-P methods. Barkved et al. (1999) discussed a subsequent 4C-3D survey that was conducted over Valhall eld. They too found that the P-S images were considerably less degraded than the P-P images below the gas cloud. Structural imaging Resolution of steeply dipping features can be improved using converted waves in certain circumstances. Purnell (1992) showed examples from physical-modeling data where highdip anomalies were more visible on migrated P-S data than on migrated P-P data. We note that structurally complicated areas may have high-velocity layers in the near surface that allow both P and S energy to propagate at signicant angles away from the vertical. This results in P-P and P-S energy being

FIG. 1. Comparison between (a) P-wave and (b) P-S 2D sections over North Sea shallow-gas channels. The P-wave section exhibits reverberations and high-frequency attenuation, but the P-S section delineates the channel base and interfaces below the channel. (Courtesy WesternGeco.)

FIG. 2. P-P and P-S sections from Valhall eld, Norway, showing improved imaging across the anticlinal structure (after Rodriguez, 2000). The white ellipses outline the upper region of the chalk reservoir.

42

Stewart et al.

recorded on both vertical and radial channels, which must be carefully processed. Cary and Couzens (2000) gave examples from Mahogany eld in the Gulf of Mexico, where P-S images show excellent denition of faults associated with salt intrusion (Figure 3). It is not obvious, at this point, why faults should appear to be better dened on the P-S than the P-P sections. Explanations include more prominent P-S scattering from nonwelded or uid-saturated contacts (Chaisri and Krebes, 2000), larger lateral S-wave changes across the faults, or P-S ray paths from steeply dipping features that are more conducive to capture with given receiver apertures. Kendall et al. (1998) also processed 4C data from Mahogany eld using an anisotropic prestack depth migration. They compared their results to streamer data shot in the same area, and illustrated improved imaging of targets underlying a salt structure (Figure 4). We note that Herrenschmidt et al. (2001) also preferred a depth imaging approach for the Mahogany data. Jin and Michelena (2000) described a prestack inversion technique that uses automatic velocity building, a rayBorn prestack depth migration, and AVO inversion from the common-angle migrated gathers. They tested this approach on the Mahogany 4C data set to image a salt body (Figure 5). The S impedance from the P-S data may have offered somewhat higher resolution than the P impedance above the salt body. Le Stunff et al. (2000) applied traveltime reection tomography to build both P- and S-wave velocity models in depth. They then used these velocities to create prestack depth-migration images from the East Natuna Basin in Indonesia. They also used common-image gathers in depth to check the validity of the velocity model. The P-S image showed an anticlinal structure (Figure 6), indicating further promise for converted-wave imaging in complex structural areas. Near-surface imaging We often see more highly resolved reectors in the near surface on P-S sections than on colocated P-P sections. This

may be the result of a number of factors, including greater relative changes in S versus P velocity, a greater impact of density changes on the P-S reectivity than on P-P reectivity, or a shorter S wavelength. For example, a 3C seismic line was acquired over the Steen River impact structure, Alberta, Canada, by Gulf Canada Resources Ltd. (now ConocoPhillips Canada Ltd.) in partnership with the CREWES Project at the University of Calgary (Mazur et al., 2002). The resultant P-P and P-S sections are shown in Figure 7, where the P-P data are stretched by a factor of two (V P /VS = 3) to match the P-S data. The sections are spliced together at a central point on the line. In Figure 7a, the P-P data are on the left and P-S data the right. Note the greater detail evident on the P-S sections. In Figure 7b, the display shows the P-S data on the left and the P-P section on the right. The P-P data are generally more continuous beneath the sub-Cretaceous unconformity (at about 480 ms). Berteussen et al. (1999) showed some striking results from a multicomponent survey conducted in the Norwegian Sea, where the nature of gas hydrates was under investigation. They observed a marked increase in the resolution of the P-S section over that of the P-P section. They concluded that P-S data may provide a signicant contribution to the interpretation of areas where gas hydrates and associated free gas exist.
LITHOLOGY ESTIMATION

Sand/shale P-wave imaging has proven particularly adept at making structural pictures of the subsurface; that is, providing an image of strata interfaces in reection time. However, beyond the conguration of interfaces, we would like to know what kind of rock and uids are in the section. P-wave images may be limited or ambiguous in these regards. S-wave measurements provide additional constraint on the rock properties (especially on density and rigidity contrasts). Much P-S analysis is targeted at nding an S-wave velocity or determining a V P /VS value (e.g., Li et al., 1999). Both VS and V P /VS can be

FIG. 3. (a) Poststack time migration of the vertical geophone component from the Mahogany 4C survey. (b) Poststack time migration of the depth-variant common-conversion-point (CCP) stack of the in-line component data. (From Cary and Couzens, 2000.)

Converted-wave Seismic Exploration

43

good indicators of rock type, especially in combination with V P (Tatham, 1982). Garotta et al. (1985) showed P-S and P-P data for a Viking sand channel reservoir in the Wineld oil eld, Alberta. They found that amplitude anomalies on the P-S data correlate with the known boundaries of the reservoir. They also used P-P and P-S isochron ratios around the area of interest to determine V P /VS and Poissons ratios. Since Poissons ratio increases monotonically with increasingV P /VS , they interpret low Poissons ratios as differentiating sand from the neighboring shales. A series of seismic experiments in the Blackfoot oil eld, Alberta, was conducted to identify sand reservoir facies from

FIG. 4. P-P and P-S migrated sections from Mahogany eld, Gulf of Mexico: (a) towed streamer data (courtesy of Diamond Geophysical), (b) w-z (hydrophone and vertical geophone) processed data from a 4C-2D line, and (c) P-S section from Kendall et al., (1998), who note improved imaging of subsalt features on the 4C data.

nonreservoir rocks (Stewart et al., 1996; Dufour et al., 2002). The surveys included broad-band 3C-2D data, 3C-3D data, and 2D and 3D vertical seismic prole (VSP) surveys. The eld was originally discovered and developed using P-wave amplitude anomalies (Figure 8a); however, there are also amplitude anomalies not associated with sand channels (that is, false positives). P-P isochron maps are also indicative of the channel, but again with some ambiguity. As in many multicomponent projects, a 2D line over Blackfoot eld was rst acquired and analyzed to assess the merit of the method. The processed 2D P-P and P-S lines tied reasonably well. Using this tie and time-thickness maps that included the channel reservoir, an interval V P /VS value was calculated. A good correlation of V P /VS anomalies and known oil production was observed. With this promise, a 3C-3D survey was conducted. The resultant P-S amplitude seems to give a more denitive (but lower-resolution) image of the sand channel (Figure 8b). A P-S isochron map that included the channel is perhaps more compelling (Figure 9a). The V P /VS maps, calculated from P-P and P-S isochron map ratios, are another strong indicator of the reservoir sand channel trend (Figure 9b). A further 3C-3D survey (Goodway and Tessman, 2000) was conducted over Blackfoot eld in 1999 with I/O VectorSeis digital 3C geophones. This survey produced similar results to the previous 1995 3C-3D survey (Figure 10). This independent survey indicates robust and repeatable results, and is being used for time-lapse analysis. MacLeod et al. (1999a) showed a case (now a classic!) of converted waves successfully delineating sand channels encased in shale at Alba eld in the North Sea. A strong contrast in S-wave velocity (from shale to sand) is associated with the top of the reservoir. On the other hand, there is relatively little Pwave velocity change across this lithologic boundary. Thus, the reservoir top generates strong converted waves, but weaker reected P-waves. The P-wave velocity, however, is sensitive to changes in uid at the oil-water contact in the reservoir, while there is only a modest S-wave velocity change at this uid boundary. The impact of the 4C OBC survey on the development of Alba has been positive (MacLeod et al., 1999b). To date, a number of successful wells have been drilled based primarily on the interpretation of the new converted-wave data. There are excellent ties between lithologies encountered in these wells and pseudoelastic impedance computed from the convertedwave seismic response. Also, the P-S data have provided new insights into the complex geometry of the turbidite channel in Alba eld, suggesting signicant postdepositional deformation of the channel. Michelena et al. (2001) discussed a 3C-3D survey conducted over the Zuata heavy-oil eld in eastern Venezuela. They indicated that there is little difference in acoustic impedance between the overlying shales and sand reservoir rocks. However, the S-wave velocity varies signicantly from shale to sand (Figure 11). They used these rock properties, seismic inversion, and neural-net classication to create sections that are indicative of lithology type (Figure 12). Van Dok and Gaiser (2001) described three 3C surveys over the Morrow formation in the southern United States. The Eva South 3C-3D survey, in the panhandle of Oklahoma, showed anomalous P-S amplitudes that correlated with net pay

44

Stewart et al.

thickness (Figure 13). Van Dok and Gaiser also used an automatic event-correlation technique (Gaiser, 1996) on 3D P-P and P-S volumes from the Cave West survey in southwestern Kansas. They found that V P /VS values (contours) from the automatic technique correlated reasonably with sand thickness values from well information (Figure 14). There is also a good correlation of high V P /VS values with a shale zone in the northwest corner of the survey (conrmed by well results). Note that the VP /VS values provide a slightly different and possibly more detailed interpretation of the distributions of shale and sand than the well data. Further down in the section, the St. Genevieve formation is encountered. A coherency plot on the P-P volume (Figure 15a) indicates a complex channel system. Low V P /VS values from automatic analysis around this horizon also give indications of channel sands and also a paleoshoreline in the southwest corner of Figure 15b.

Conglomerate Nazar and Lawton (1993) used AVO stacks and P-P and P-S sections to analyze the productive regions of Carrot Creek eld, Alberta (Figure 16). The oil-saturated conglomerate in this region is a challenge to interpret on conventional data, but is quite apparent on P-S sections as a brightening in amplitude. This relative brightening is partially due to differences in tuning between P-P and P-S reections and the relatively high S-wave velocity of the Cardium conglomerate. Anhydrite/dolomite Miller (1996) found signicant variations in V P /VS in a carbonate play at Lousana, Alberta. She estimated interval V P /VS values from the ratio of P-P and P-S isochrons that include

FIG. 5. P-P and P-S migration and inversion results from the Mahogany 4C-2D survey in the Gulf of Mexico (Jin and Michelena, 2000). The P-wave impedance section from the P-wave data (a) and S-wave impedance values derived from P-S data (b) show the general outline of an encased salt body.

FIG. 6. (a) P-S image from the East Natuna Basin in Indonesia (Le Stunff et al. 2000). (b) An image gather shows attened events and thus an appropriate velocity eld.

Converted-wave Seismic Exploration

45

the region of interest [as described in a companion article by Stewart et al. (2002)]. The V P /VS values in the Cretaceous section (ranging from 2.2 to 2.5) are indicative of a clastic section, whereas those in the Paleozoic (1.52.0) are characteristic of carbonate rocks. The lowest V P /VS values in the Paleozoic section are coincident with an oil-bearing reef. She concluded that this anomaly was associated with dolomitization as opposed to the surrounding higher V P /VS values that were coincident with anhydrite.
ANISOTROPY ANALYSIS

Many hydrocarbon reservoirs are fractured. The volume of oil or gas in place and the reservoirs ability to produce it are dependent on the fracture state of the reservoir. Determining fracture density and orientation from seismic data has thus been a subject of considerable effort (e.g., Probert et al., 2000; Crampin, 2001). Ata and Michelena (1995) showed an

example of three 3C seismic lines arranged in a star pattern in Venezuela. After processing and analyzing the data, they found indications of fracture orientation from their calculated anisotropy. Gaiser (2000) showed the results of applying an Alford (1986) anisotropic rotation procedure and layer stripping (where off-diagonal components are minimized) to 1999 data from the Teal South, Gulf of Mexico 4C-4D surveys conducted by the Energy Research Clearing House (ERCH) of Houston. A multiazimuth receiver gather (Figure 17) indicates the amount and direction of the S-wave azimuthal anisotropy. Van Dok et al. (1997) and Gaiser (1999) showed analyses of a full 3C-3D seismic survey from Madden eld in the Wind River Basin, Wyoming. This analysis used 4C Alford (1986) rotations and layer stripping (Winterstein and Meadows, 1991a) to calculate the fast shear-wave (P-S1 ) polarization direction and the associated percent anisotropy. Figure 18 is a portion of the data from an east-west line: the radial and transverse

FIG. 7. Seismic data over the Steen River structure, Alberta, Canada. The sub-Cretaceous unconformity is evident at about 480 ms on the P-P section and 1000 ms on the P-S section. (a) The left half of the P-P section spliced to the right half of the P-S section. (b) The left half of the P-S section spliced to the right half of the P-P section. Note that some events are more clearly dened in the shallow P-S sections. The data were acquired by Gulf Canada Resources Ltd. (now ConocoPhillips Canada Ltd.) and the CREWES Project at the University of Calgary.

46

Stewart et al.

polarization for east-west propagation. There are equivalent sections for north-south propagation. To compensate for the effects of depth-varying properties (Winterstein and Meadows, 1991b), the reection at 1.5 s is analyzed rst to remove shear-wave splitting effects from the overburden. Although the energy on the transverse component for this event is weak in places, removing overburden effects from the target reections below is an important step in unraveling shear-wave birefrin-

gence. Figure 19 shows the fast S-wave direction and its associated percent anisotropy, corresponding to possible fracture orientation and fracture density of the target horizons between 2.2 and 3.3 s. Regions of stronger anisotropy (9% or more) correlate well with the known east-west trending faults superposed on the maps. As in many situations, resolution is an issue. These fracture-property estimates are averages over 1.1 s of dataan interval clearly larger than the reservoirs of interest. Fracture

FIG. 8. (a) P-P and (b) P-S time slices at the interpreted sand channel level from the Blackfoot 3C-3D survey. The grid lines indicate a section (1 mile 1 mile or approximately 1.6 km 1.6 km).

FIG. 9. (a) P-S isochron map between the Mannville and Mississippian horizons (white/yellow indicates a time thickness of 140 ms through purple with a value of 90 ms) and (b) the V P /VS value as determined from the P-P and P-S isochron maps between the interpreted top of the channel and Wabamun horizons. White/yellow represent a VP /VS value of 1.5 through purple indicating a value of 2.8.

Converted-wave Seismic Exploration

47

detection at ner intervals can be attempted by careful survey design to provide optimal fold, offset, and azimuth distributions. Numerical modeling (Li et al., 1996) suggests that gassaturated and oriented fractures may have an effect on

anisotropic P-S reectivity. This is in contrast to the isotropic case, where uid saturation appears to have less impact on Swave velocities. In fact, Guest et al. (1998) interpreted anomalies in S-wave splitting over a gas reservoir in Oman as evidence of an effect of gas on shear waves.
FLUID DESCRIPTION

FIG. 10. (a) A second 3C-3D survey conducted in 1999 over Blackfoot eld, using VectorSeis digital geophones (Goodway and Tessman, 2000) gave a V P /VS anomaly (red indicates low VP /VS values similar to Figure 9b) that correlates with (b) net sand thickness (where thicker sands are indicated by yellow/red).

Thompson et al. (2000) presented early results from a 30-km 2D multicomponent line, in 750 m of water, shot over the Fles Dome, offshore Norway. There is a at spot on the P-P data set that could be an event caused by a uid contact (Figure 20). However, it could also be generated by a lithologic change. The continuity of dipping strata in the P-S section (lack of a at spot) supports the possibility that the P-P anomaly is caused by uids not a lithologic change. Stewart and Todorov (2000) used the Blackfoot 3C-3D seismic data to estimate oil column height. They rst constructed a P-P time-thickness (isochron) map for an interval (Mannville to Mississippian) containing the reservoir. This map provided a regular and dense data set complementary to the sparse interval-thickness values picked from the well logs over the same Mannville-Mississippian interval. The isochron values were then co-kriged with the isopach values from well logs. The isochron map was thus converted to a regularly sampled and dense isopach map. The V P /VS value was next computed using the isochron ratios from the P-P and P-S maps over the same depth interval. This V P /VS value was, in turn, co-kriged with the gammaray index to give an estimate of the shale content by using sand and shale lines and linear interpolation of values between them. The clastic interval (the isopach map under consideration) was assumed to consist only of shale and sand. Thus, the sand estimate was 100 minus the shale percentage. Amplitude inversions were derived from both P-P and P-S volumes to determine P- and S-wave velocities, respectively. The P-P and P-S attribute volumes with V P and VS inversion volumes plus porosity logs were then used to predict porosity values over the interval using a neural net (similar to a procedure described in Russell et al., 2002). The average water saturation (SW ) in the reservoir, which is 25%, was used to infer the oil saturation (100 minus the percentage water). Finally, the oil column height (OCH) for every seismic point was estimated using the product of these factors: OCH = interval thickness percent sand porosity oil saturation, as shown in Figure 21. By using a 3.0-m cutoff value, they estimated the volume of oil in place at about 1.2 106 m3 . The engineering report using the original 3D seismic and about one year of production data estimated a value of 1.36 106 m3 .
RESERVOIR MONITORING

FIG. 11. S-wave velocity versus density from well logs in the Zuata heavy oil eld, Venezuela. The S velocity of the sands is considerably higher than that of the shales (Michelena et al., 2001).

Isaac (1996) showed P-P and P-S sections from a heavy-oil reservoir at Cold Lake, Alberta, undergoing steam ooding (Figure 22). There are variations in the reservoir rock properties associated with temperature and saturation changes. These, in turn, are associated with changes in the seismic character of both P-P and P-S sections. Using surveys in 1993 and repeated in 1994, she found that the variation in V P /VS values correlated with the temperature of the reservoir. The V P /VS values stay constant in areas away from the injection wells (CDPs 2070

48

Stewart et al. WHATS LEFT TO DO

in Figure 23). However, in the areas steamed in 1994, there is an increase in V P /VS that causes the ratio of the V P /VS values from the two years to drop (Figure 23c). Spitz et al. (2000) also discussed results from the 4C-4D Teal South survey, Gulf of Mexico, consisting of two 4C-3D surveys, one in 1997 and the other in 1999. Their pioneering study shows time-lapse differences in the fast converted wave (P-S1 ) over the Teal South eld (Figure 24).

Converted-wave exploration has come a long way in recent years, but there is still plenty of room for progress. Further applications await surveys and resultant images in new environments. The P-S method will undoubtedly become more widely practised and useful as costs decrease. The expense of land multicomponent surveys is decreasing signicantly, but

FIG. 12. Section from the Zuata heavy-oil eld, Venezuela. The S-wave inversions have been converted to sand/shale estimates by a neural-net process. SP and gamma-ray logs are overlain for comparison (Michelena et al., 2001).

FIG. 13. (a) P-S amplitudes at the reservoir level (top Morrow) from the Eva South 3C-3D survey. High amplitudes are in red. (b) Net pay thicknesses from well information (after Miller and Wheeler, 2000). Sand thickness contours are in 10-ft (3.3-m) intervals. The maximum thickness is about 35 ft (11 m).

Converted-wave Seismic Exploration

49

OBC costs are still well above those of towed-streamer methods. Processing and analysis of P-S data have become much more effective and sophisticated, especially by incorporating prestack techniques and anisotropy (Thomsen, 1999). Processing P-P and P-S data together to provide consistent images in depth (Mikhailov et al., 2001) and improved rock property estimates (Margrave et al., 2001) are critical current developments. For example, Spitz (2001) showed a case from the North Sea where P-P and P-S inversion was used to derive a density estimate (which compared favorably with a density log in the area). Further renements await. More detailed analysis of existing images may provide us with greater understanding of the targets under consideration or even new ones. Better interpretation tools are under development, especially with respect to correlation and depth conversion, but additional advances would be welcomed. Continued education and experience will further unravel what converted waves have to show us (Cary, 2001). Looking farther ahead, we anticipate making use of other modes that propagate in a seismic survey, such as a wave, otherwise P, that has an S-wave leg through a high-velocity region. In cases where there are high-velocity layers (basalts, carbonates, salts, or even permafrost in the near surface), seismic imaging may be complicated or compromised. We can make sections

(Figure 25) from this more complicated conversion as in the Gulf of Mexico, where a P-wave converting to an S-wave inside a salt volume and then back to P-wave has been used to create an image. The possibility of high-quality, fully elastic and anisotropic images of the subsurface opens many doors to new interpretation. Accurately repeating these surveys (4C-4D) to look for changes associated with uid movement is a very exciting prospect (Grechka, 2001; Jack, 2001). Permanent seismic monitoring of oil elds, either with active sources or passive listening, using surface and/or borehole measurements will provide considerably more guidance for reservoir production.
CONCLUSIONS

The reection seismic method has used P-waves for many yearswith great success. The extension of the reection method to include P-S waves has been effective in yielding new cases of improved imaging of resource targets. Particularly well documented cases exist for gas-cloud imaging, sand/shale discrimination, and anisotropy analysis. However, there is more to be done in converted-wave exploration seismology, especially in making full use of these new pictures.

FIG. 14. Map of the automatically determined V P /VS value from the Cave West, Kansas survey (after Van Dok and Gaiser, 2001). Low V P /VS values are in red. The interval of V P /VS analysis includes the Morrow formation and can be interpreted in terms of possible sand distributions. Note the high values in the northwest corner that are interpreted to indicate high shale content (conrmed by well results). Gross thickness of the sand interval, from well data, is indicated by 10-ft (3.3-m) contour lines.

50 ACKNOWLEDGMENTS

Stewart et al.

We express our deep appreciation to the sponsors of the Consortium for Research in Elastic Wave Exploration Seismology (the CREWES Project) for their commitment to the develop-

ment of multicomponent seismology. We also express our gratitude to Neil Jones, Rich Van Dok, and Robert Bloor of WesternGeco for providing data examples, and to Heloise Lynn for her insightful interpretations of the 3C Madden survey. Reinaldo Michelena of PDVSA generously provided us with

FIG. 15. Maps of the St. Genevieve horizon. (a) Coherency plot derived from the P-wave data indicates a channel system, with (b) the correspondingV P /VS plot automatically calculated over a 100-ms window around the horizon. The red and yellow colors indicate a low V P /VS value and possible sand accumulations (after Van Dok and Gaiser, 2001).

Converted-wave Seismic Exploration

51

FIG. 16. Portions of the (a) P-S and (b) P-P sections from Carrot Creek, Alberta. Note the amplitude anomaly on the P-S section at the Cardium conglomerate level (from Nazar and Lawton, 1993).

FIG. 17. Anisotropic rotations on a receiver gather from the Teal South survey. The slow S-wave (P-S2 ) on the right has been compressed to t the fast converted wave on the left. The total amount of anisotropy and its direction is shown in the middle panel (after Gaiser, 2000).

the Zuata heavy-oil eld example. We thank Robert Kendall of Veritas DGC for supplying us with his Mahogany examples, and Diamond Geophysical for the Mahogany streamer data. We are grateful to the Amoco (now BP) group of companies for providing us with the Valhall 4-C data set for our analysis. We used Hampson-Russells GEOSTAT program (donated to the University of Calgary) for analysis in the Blackfoot eld oilestimation case. Joanie Whittemore and Carla Osborne (formerly with CREWES) ably assisted with the production of this manuscript.

REFERENCES Alford, R. M., 1986, Shear data in the presence of azimuthal anisotropy: Dilly, Texas: 56th Ann. Internat. Mtg., Soc. Expl. Geophys., Houston, Expanded Abstracts, 476479. Ata, E., and Michelena, R. J., 1995, Mapping distribution of fractures in a reservoir with P-S converted waves: The Leading Edge, 14, 664 676. Bancroft, J. C., 2000, A practical understanding of pre- and post-stack migration: Soc. Explor. Geophys. Barkved, O. I., Mueller, M. C., and Thomsen, L., 1999, Vector interpretation of the Valhall 3D/4C OBS dataset: 61th Ann. Conf., Eur. Assn. Geosci. Eng., Extended Abstracts, 642.

52

Stewart et al. Nebieridze, S., and Boelle, J.-L., 2001, Comparison of different strategies for velocity model building and imaging of PP and PS real data: The Leading Edge, 20, 984995. Isaac, J. H, 1996, Seismic methods for heavy oil reservoir monitoring: Ph.D. thesis, Univ. of Calgary. Jack, I., 2001, The coming of age for 4D seismic: First Break, 19, 1, 2428. Jin, S., and Michelena, R. J., 2000, Prestack inversion of multicomponent data from complex structure: Mahogany Field example, Gulf of Mexico: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Kendall, R. R., Gray, S. H., and Murphy, G. E., 1998, Subsalt imaging using prestack depth migration of converted waves: Mahogany Field, Gulf of Mexico: 68th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 20522055. Kristiansen, P., 2000, 4 years experience with 4C seismic: What we have learned: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Le Stunff, Y., Gosselet, A., Nguyen, S., Berthet, S., and Boelle, J. L., 2000, 3D PP and PS travel time reection tomography on a case study: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Li, X., Dai, H., Mueller, M. C., and Barkved, O. I., 2001, Compensating for the effects of gas clouds on C-wave imaging: A case study from Valhall: The Leading Edge, 20, 10221028. Li, X., Kuhnel, T., and MacBeth, C., 1996, Converted-wave AVO and its implications: 58th Ann. Conf., Eur. Assn. Geosci. Eng., Extended Abstract, M046. Li, X., Yuan, J., Ziolkowski, A., and Strijbo, F., 1999, Estimating VP /VS ratio from converted wavesa 4C case example: 61th Ann. Conf., Eur. Assn. Geosci. Eng., Extended Abstracts, P066. Macbeth, C., Michelena, R., and Spitz, S., 2001, Introduction to Recent advances in shear wave technology for reservoir characterization: A new beginning?: The Leading Edge, 20, 972. MacLeod, M. K., Hadley, M. J., Reynolds, K. J., and Tura, A., 1999a, Multicomponent analysis of OBC data: 31st Ann. Off. Tech. Conf., OTC 10940. MacLeod, M. K., Hanson, R. A., Bell, C. R., and McHugo, S., 1999b, The Alba eld ocean bottom cable seismic survey: Impact on development: Off. Eur. Conf., SPE 56977. Margrave, G. F., Stewart, R. R., and Larsen, J. A., 2001, Joint PP and PS seismic inversion: The Leading Edge, 20, 10481052. Mazur, M. J., Hildebrand, A. R., Hladiuk, D., Schafer, A., Pilkington, M., and Stewart, R. R., 2002, The Steen River crater seismic refraction project: Presented at the 33rd Lunar and Planetary Science Conf., Houston. Michelena, R. J., Donati, M. S., Valenciano, A. A., and DAgosto, C., 2001, Using multicomponent seismic for reservoir characterization in Venezuela: The Leading Edge, 20, 10361041. Mikhailov, O., Johnson, J., Shoshitaishvili, E., and Frasier, C., 2001, Practical approach to joint imaging of multicomponent data: The Leading Edge, 20, 10161021. Miller, S. L. M., 1996, Multicomponent seismic data interpretation: M.Sc. thesis, Univ. of Calgary. Miller, W. A., and Wheeler, D. M., 2000, 3C-3D seismic characterization of the Eva South Morrow Sand unit, Texas County, Oklahoma: Presented at the 6th Ann. 3D Seismic Symp. Rocky Mountain Assn. of Geologists and Denver Geophys. Soc. Nazar, B. D., and Lawton, D. C., 1993, AVO analysis of a thin conglomerate deposit: J. Seis. Expl., 2, 333348. Probert, T., Ronen, S., and Bryan, R., 2000, A case study of azimuthal anisotropy analysis from a N. Sea 3D 4C project: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Purnell, G. W., 1992, Imaging beneath a high-velocity layer using converted waves: Geophysics, 57, 14441452. Rodriguez, C., 2000, Advanced marine methods: Ocean-bottom and vertical cable analyses: Ph. D. thesis, Univ. of Calgary. Russell, B. H., Hampson, D., Todorov, T., and Lines, L. R., 2002, Combining geostatistics and multi-attribute transforms: A channel sand case study, Blackfoot oileld (Alberta): J. Petrol. Geol., 25, 97117. Spitz, S., 2001, Seismic analysis with multi-components: Can. Soc. Expl. Geophys. Recorder, 26, 7, 5761. Spitz, S., Gratacos, B., Haggard, W., and Vuillermoz, C., 2000, Reservoir monitoring using multicompoinent seismic: Processing the Teal South 4D-4C: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Stewart, R. R., Ferguson, R., Miller, S. L. M., Gallant, E., and Margrave, G., 1996, The Blackfoot seismic experiments: Broad-band, 3C-3D, and 3-D VSP surveys: Can. Soc. Expl. Geophys. Recorder, 21, no. 6, 710. Stewart, R. R., Gaiser, J. E., Brown, R. J., and Lawton, D. C., 2002, Converted-wave seismic exploration: Methods: Geophysics, 67, 7, 116.

FIG. 18. (a) Radial and (b) transverse receiver components for east-west propagation from a 3D survey at Madden eld in Wyoming. Equivalent sections for north-south propagation are used with 4C rotations and layer-stripping techniques to estimate fracture orientation and intensity as shown in Figure 19 (after Gaiser, 1999).
Berg, E., Svenning, B., and Martin, J., 1994, SUMIC: Multicomponent sea-bottom seismic surveying in the North SeaData interpretation and applications: 64th Ann. Intl. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 477480. Berteussen, K. A., Langhammer, J., Andreassen, K., Sognnes, H. I., and Henneberg, K., 1999, Multicomponent sea bottom data in gas hydrate exploration: 61th Ann. Conf., Euro. Assn. Geosci. Eng., Extended Abstracts, 6-20. Cary, P. W., 2001, Multicomponent seismic exploration in Canada One persons perspective: Can. Soc. Expl. Geophys. Recorder, 26, 6267. Cary, P. W., and Couzens, R. A., 2000, Processing 4-C data from Mahogany Field, Gulf of Mexico: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Chaisri, S., and Krebes, E. S., 2000, Exact and approximate formulas for P-SV reection and transmission coefcients for a nonwelded contact interface: J. Geophys. Res., 105, 2804528054. Crampin, S., 2001, Shear-wave anisotropy: A new window into the crack-critical rockmass: Can. Soc. Expl. Geophys. Recorder, 26, no. 7, 1127. Dufour, J, Squires, J., Goodway, W. N., Edmunds, A., and Shook, I., 2002, Integrated geological and geophysical interpretation case study, and Lame rock parameter extractions using AVO analysis on the Blackfoot 3C-3D seismic data, southern Alberta, Canada: Geophysics, 67, 2737. Gaiser, J. E., 1996, Multicomponent VP /VS correlation analysis: Geophysics, 61, 11371149. 1999, Applications for vector coordinate systems of 3-D converted-wave data: The Leading Edge, 18, 12901300. 2000, Advantages of 3-D PS-wave data to unravel S-wave birefringence for fracture detection: 70th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 12011204. Garotta, R., Marechal, P., and Magesan, M., 1985, Two-component acquisition as a routine procedure for recording P-waves and converted waves: Can. J. Expl. Geophys., 21, 4053. Goodway, W., and Tessman, D. J., 2000, Blackfoot 3C/3D, a VectorSeis case history: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Granli, J. R., Sollid, A., Hilde, E., and Arnsten, B., 1995, Imaging through gas-lled sediments with marine S-wave data: 65th Ann. Internat. Mtg., Soc. Explor. Geophys., Expanded Abstracts, 352 355. Grechka, V., 2001, Seismic anisotropy: Yesterday, today, tomorrow: Can. Soc. Expl. Geophys. Recorder, 26, no. 7, 910. Guest, S., Vander Kolk. C., and Potters, H., 1998, The effect of fracture lling uids on shear-wave propagation: 68th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 948951. Herrenschmidt, A., Granger, P.-Y., Audebert, F., Gerea, C., Etienne, G., Stopin, A., Alerini, M., Lebegat, S., Lambare, G., Berthet, P.,

Converted-wave Seismic Exploration Stewart, R. R., and Todorov, T., 2000, Oil volume estimation using 3C3D seismic data and well logs: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho. Tatham, R. T., 1982, VP /VS and lithology: Geophysics, 47, 336 344. Thomsen, L., 1999, Converted-wave reection seismology over inhomogeneous anisotropic media: Geophysics, 64, 678 690. 2002, Understanding seismic anisotropy in exploration and exploitation: Soc. Expl. Geophys. and Eur. Assn. Geosci. Eng. Distinguished Instructor Series 5. Thompson, C., Helgesen, H. K., and Battie, J. E., 2000, 2D-4C seismic exploration data for risk reduction, Fles prospect, offshore MidNorway: Presented at the SEG/EAGE Summer Research Workshop, Boise, Idaho.

53

Van Dok, R. R., and Gaiser, J. E., 2001, Stratigraphic description of the Morrow formation using mode-converted shear waves: Interpretation tools and techniques for three land surveys: The Leading Edge, 20, 10421047. Van Dok, R. R., Gaiser, J. E., Jackson, A.R., and Lynn, H.B., 1997, 3-D converted-wave processing: Wind River Basin case history: 67th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 12061209. Winterstein, D. F., and Meadows, M. A., 1991a, Shear-wave polarizations and subsurface stress directions at Lost Hills eld: Geophysics, 56, 13311348. 1991b, Changes in shear-wave polarization azimuth with depth in Cymric and Railroad Gap oil elds: Geophysics, 56, 13491364. Yilmaz, O., 2001, Seismic data analysis: Processing, inversion, and interpretation of seismic data: Soc. Explor. Geophys.

54

Stewart et al.

FIG. 19. Seismic anisotropy calculated by Alford rotation and layer stripping from the data in Figure 18, after removal of the birefringence effects of the overburden. (a) Fast S-wave (P-S), polarization direction and (b) percent azimuthal anisotropy for the target layer between 2.2 and 3.3 s. Regions of stronger anisotropy (9% or more) correlate well with the known east-west trending faults superposed on the maps (after Gaiser, 1999).

Converted-wave Seismic Exploration

55

FIG. 20. Flat spot analysis on P-P and P-S from the Fles prospect, offshore Norway (Thompson et al., 2000). The top section shows the P-wave data from the 2-D OBC survey, the middle section the P-S data from the OBC survey, and the bottom section a line extracted from a 3D towed streamer volume. There is no obvious at spot on the P-S data, suggesting that the P-wave anomaly is a uid contact, not a lithology change.

FIG. 21. OCH estimate for the Blackfoot, Alberta eld (Figure 8) as determined geostatistically from interval thickness percent sand porosity oil saturation.

FIG. 22. Comparison of the (a) 1993 and (b) 1994 3C seismic lines from the Cold Lake, Alberta, steam-injection site. Note the similar data quality and resolution among all lines (from Isaac, 1996). The area of interest is indicated by the vertical bar. CDP/CCP trace spacing is 8 m.

56

Stewart et al.

FIG. 23. V P /VS plots for (a) 1993 lines, (b) 1994 lines, and (c) the ratio of those two. Note that the V P /VS value is fairly constant in the unsteamed regions away from the wells (CDP numbers 2070) (from Isaac, 1996). However, the V P /VS values increase in 1994 with steaming. CDP/CCP trace spacing is 8 m. The lines were 1800 m long.

FIG. 24. Fast-converted-wave (P-S1 ) time slices from the Teal South, Gulf of Mexico, 4C-4D survey at the reservoir level. The P-S image from 1997 (a) is subtracted from the 1999 image to give the (b) difference (Spitz et al., 2000).

Converted-wave Seismic Exploration

57

FIG. 25. Comparison of prestack depth-migrated streamer data using (a) P-wave and (b) S-wave salt-velocity models. Migration with an S-wave salt-velocity model images PSSP waves that have converted at the top of the salt. The position of the S-wave event at the base of the salt is in agreement with the P-wave image. (Courtesy WesternGeco.)

You might also like