AVO analysis and complex attributes for a Glauconiticgas sandbar

Hai-man Chung and Don C. Lawton

ABSTRACT

A Glauconitic gas-bearing sandbar from Southern Alberta was studied in terms ofits AVO behavior and complex attributes. The seismic data reveals an interesting anomalyin the full-offset stack. In particular, the anomaly appears clearly in the near-offset stack(0-1050 meters) but is poorly outlined in the mid-offset stack (1050-2100 meters).Subsequent AVO modelling indicates that in the subject area under study, for theGlauconitic Formation, thin-bed tuning effects overwhelm any AVO effects due to lateralchanges of lithology for offsets up to 3000 meters. The differences of the anomaly as itappears in the near-offset and mid-offset stacks can be largely attributed to a lowerfrequency content in the mid-offset stack.

The complex attributes plots show corresponding differences of the anomaly in thevarious offsets attributes plots. In particular, the instantaneous frequency plot for the near-offset stack reveals the anomaly distinctly, but it is virtually absent on the correspondingplot for the mid-offset stack. It is evident that complex attributes plots are very sensitive tothe frequency content of the seismic data.

PROJECT OBJECTIVE

In recent years, the subject of amplitude variation with offset(AVO) analysis hasattracted much attention among exploration geophysicists. This is mainly due to thepotential capability of AVO analysis of seismic data to delineate clastic gas reservoirs, asdiscussed by Ostrander (1984). However, in addition to the lateral change of the Poisson'sratio due to the presence of gas sands, many other physical factors, such as thin-bedtuning, also contribute to AVO effects. Therefore, the method should be studied further toenhance our understanding of its usefulness and pitfalls.

The applications of complex attributes to seismic signal analysis were firstdiscussed by Taner (1977,1979). In the ensuing years up to the present time, there havebeen only three other papers published in Geophysics on the same subject, of which thepaper by Robertson and Nogami (1984) is the only one that deals with thin-bed effects.No doubt, more papers need to be published on the successes and failures of the usage ofcomplex attributes in exploration in order to promote further understanding anddevelopment of the subject.

One of the objectives of the Crewes project is to study the AVO analysis andcomplex attributes of both P-wave and S-wave seismic data. In particular, their ability todelineate thin beds will be investigated. This paper is the first step in the study, andinvolves only P-wave data. It is a case study of a gas-bearing Glauconitic sandbar inSouthern Alberta. The reader is assumed to be familiar with the theories of both AVOanalysis and complex attributes, as they will not be discussed here. For review of thesesubjects, the papers by Ostrander (1984), Taner (1977,1979), and Robertson and Nogami(1984) are excellent references.

289

INTRODUCTION

In early 1986, Summit Resources and Alberta Energy Company decided to templatea Glauconitic gas well in Southern Alberta with a 3 km long seismic line. The purpose wasto investigate the seismic signature of the gas-bearing sandbar. Since the well does nothave a sonic log, no forward seismic modelling could be performed. Geophysically,shooting a seismic line over the well is perhaps the most effective way to study the sandbarin the absence of a corresponding sonic log.

A 21-fold seismic line was subsequently shot over the well in the latter part ofJanuary, 1986. The seismic data revealed a significant anomaly over the well location.There was drape over the sandbar, phase reversal due to saturated gas sands, and anapparent Mississippian low structure below the gas sands.

The objective of the case study is to investigate the channel sand anomaly in termsof AVO effects and complex attributes. In particular, complex attributes as a function ofsource-receiver offsets will be examined. Because of confidentiality, all seismic shotpointlocation numbers will be omitted, and the wells will be referred to symbolically withoutrevealing their true locations.

GEOLOGICAL BACKGROUND

In the subject area under study, the Upper Mannville Glauconitic Formation isrepresented by two lithofacies, (a) regional sequence, and (b) channel features which areshale and sand-filled. The regional sequence consists of a shoaling upward cycle from theOstracod limestones and Bantry shales. The shoaling sequence carries through to deltaplain carbonaceous shales and coals. Delta front sands and localized shoreface sandswithin the regional sequence can form thin reservoirs if trapping by channel truncationand/or sufficient stmctural reversal occurs. The entire sequence from Ostracod to deltaplain is rarely more than 40 meters in thickness.

Following the deposition of the regional sequence, a series of major channelsdown-cut through it and generally, but not always, also through the underlying Ostracodand Bantry Formations. Within the channels, large discrete bars of varying thicknesseswere deposited which can completely fill the channel with a single clean, medium to coarsegrain quartzose sandbar. The sandbars can be up to 40 meters thick with an areal extent ofup to a section, though half a section is more common.

The subject well penetrated a 40-meter channel section and complete truncation andremoval of the Oslracod/Bantry section has occurred. The channel fill consists of a thick21-meter basal sand over which lies 18 meters of silty/sandy shale and a 1-meter layer ofcarbonaceous shale which caps the entire channel fill sequence. Logs indicate that thesandbar has an average porosity of 23%. Production testing and log analysis indicate thateight meters of gas pay are present over a 13-meter water leg within the channel sand.Reserves are estimated at about five billion cubic feet.

Figure 1 shows the interpreted channel position in the area, the locations of thesubject well (E) and neighboring wells, and the location of the template seismic line.Figure 2 is a structural cross-section through the subject well (E) and some of theneighboring wells. It clearly illustrates the channeling event through the subject well, withthe porous sands highlighted in yellow and the gas-producing zone in red.

290

ACQUISITION AND PROCESSING

The seismic line was acquired with P vibrators as the energy source, and wasrecorded as sign-bit signals. The following chart summarizes the acquisition parameters:

spread: 1600-25-0-25-1600 meters(0-25-3200 meters forfirst roll-in and lastroll-out shots)

source interval: 75 meters

receiver interval: 25 meters

source: Vibroseis, 4 Mertz vibratorsover 32 meters, 12 sweep at14 seconds, 13-55 Hz, 16-59Hz, 18-63 Hz, 22-67 Hz, 26-71 Hz, 28-75 Hz, upsweep anddownsweep

receivers: Mark L-28, 10 Hz, 9 at 2.7meters

instruments: Geocor IV, 128 trace

fold coverage : 21

samplerate: 2ms

field filters: out-out, notch out

The seismic line was processed by Seis-pro Consultants Ltd. The various stacksections indicate that the data have good signal-to-noise ratio. Because we are interested inanalyzing amplitude variation with offset, every effort was made to preserve true relativeamplitudes. This includes application of gain to compensate for spherical divergencewithout any trace equalization before deconvolution. For the deconvolution, a surface-consistent shot deconvolution was applied. In other words, one single deconvolutionoperator, which was obtained as an average over all traces belonging to the same shot, wasapplied to those traces. This contrasts with the normal procedure of obtaining onedeconvolution operator for each trace and applying it to that trace alone. The surface-consistent deconvolution is an attempt to preserve the amplitude characteristics for eachwavelet corresponding to each shot. After stacking, long-window rms scaling was appliedto each trace to ensure that their rms amplitudes do not significantly differ from each other.The following chart summarizes the processing steps:

1. demultiplex2. gain - spherical divergence only, no trace

equalization3. geophone phase compensation

291

4. surface-consistent shot deconvolution5. elevation and weathering corrections6. NMO correction - first pass7. surface-consistent statics8. NMO correction - second pass9. gather10. trim statics11. stack - 21 fold12. filter - bandpass, 10/15 - 75/85 Hz13. scaling- multigate window, 50 to 350 ms, 350 to 1600 ms

Figure 3 shows a normal polarity display of the final stack. The gas-beatingsandbar is represented by the boxed anomaly. Here one can observe drape over the gassands, phase reversal probably due to their low velocity, and a Mississippian low.Moreover, the amplitudes of the peaks along the drape above the sandbar decrease over thesandbar, while the reversal also shows clear amplitude variations. In the following twosections, we will analyze this anomaly in terms of amplitude-versus-offset effects andcomplex attributes.

AMPLITUDE VARIATION WITH OFFSET ANALYSIS (AVO)

To investigate the AVO effects of the gas-bearing sandbar, three partial stacks weregenerated. The first one, the near-offset stack, covers the distance of 0 to 1050 meters, thesecond one, the mid-offset stack, 1050 to 2100 meters, and the third one, the far-offsetstack, 2100 to 3175 meters. However, to cancel noise due to surface waves, the far-offsetstack was muted to below the zone of interest, which is about 1000 ms. Hence, we willnot discuss the far-offset stack, and effectively, we have offsets only up to 2100 meters.

Figures 4, 5, and 6 show the seismic anomaly for full-offset stack, near-offsetstack, and mid-offset stack respectively. The anomaly appears to be significantly differenton the near-offset stack than on the mid-offset stack. The full-offset stack is the average ofthe other two. The anomaly in the near-offset stack has a distinct phase reversal signature.The drape on top of the sandbar is also very evident in this stack. However, the apparentdelay in the Mississippian event that is clear on both the full-offset and mid-offset stacksdoes not appear in the near-offset stack. On the other hand, in the mid-offset stack, theanomaly appears as a very broad and lower-frequency wavelet without showing anyoverlying drape nor clear reversal character.

To attempt to explain the differences in the two stacks, the first step is to investigatethe possibility of the differences as a result of changes in Poisson's ratio. Since low-velocity saturated gas-sands have relatively low Poisson's ratios, they often show up asamplitude anomalies when the corresponding seismic data are displayed in some offset-dependent format such as CDP gather panels (Ostrander, 1984). Figure 7 shows theOstrander gather (Ostrander, 1984) for four CDP locations along the seismic template line.To obtain the gather, six CDP panels across each location are summed with a three-offsetsrange for each output trace. In other words, each trace on the Ostrander gather in figure 7is the sum of eighteen traces. This effectively smears the reflections over six CDPlocations, that is over 62.5 meters and for an offset range of 75 meters. However, thesignal-to-noise ratio is also enhanced by a factor of four. Note that there is no tracebalancing applied to the data. CDP locations A and B are regional Glauconitic locations,and CDP location C is the channel edge location. All three locations show similar

292

characters at the Glauconite to Mississippian interval, with relative strong peaks for theGlauconitic reflections. However, CDP location D, which is where the gas-bearingsandbar is situated, shows remarkably different results. The Glauconitic reflections havevery low amplitudes for the first three near-offset traces, then appear as low-amplitudedoublets for the mid-offset traces, and finally turn into high-amplitude single peaks withsome evidence of NMO stretch. In fact, it appears that the Glauconitic reflections reversepolarity over the mid-offset traces.

To further understand these seismic character changes, AVO analyses wereperformed on three wells, using the Hampson-Russell AVO modelling package. In thispackage, ray tracing is performed, and reflection coefficients are calculated by solving theZoeppritz equations for specified offsets up to the critical distance. One inputs the soniclog and density log, and the package will assign an initial value of 0.25 for the Poisson'sratio for all layers, which can be modified as desired. A peak frequency of 31 Hz and amaximum offset of 3000 meters are used in all the AVO synthetics. For each well chosenfor AVO analysis, two synthetics will be generated. The first one will be a NMO correctedsynthetic, where ray-tracing is first performed, and then NMO correction is applied. Thesecond one will be a plane-wave solution synthetic, where every trace is a zero-offset trace,although the reflection coefficients are still offset-dependent. This synthetic will allow usto investigate AVO effects as predicted by Zoeppritz equation without the interference ofthin-bed tuning and NMO stretch effects.

The chosen wells are (B) (figure 1), a Glauconitic channel gas well; (C), aGlauconitic channel shale well (also a gas well from another formation); and (F), aGlauconitic regional well. The well (B) has lithologies similar to that of the subject well(E); since (E) does not have a sonic log, the thicknesses of the two lithological units in theGlauconitic Formation in (B) are modified to reflect the corresponding thicknesses in (E).Figures 8 and 9 show the results of the AVO modelling on this modified sonic log for twovalues of the Poisson's ratio for the gas-bearing sands. In figures 8a, 8b, and 8c, thePoisson's ratio is 0.1. In figures 9a, 9b, and 9c, it is 0.25. In both cases, the P-wavereflection coefficients are negative, reflecting the low-velocity porous sands. However, infigure 8a, where the Poisson's ratio is 0.1, the P-wave reflection coefficients do not changeappreciably for offsets less than 2000 meters. The corresponding NMO corrected synthetic(figure 8b) shows some noticeable amplitude changes at an offset of 1800 meters or larger,while NMO-stretch effects are observable at about 2400 meters and larger offsets. On theother hand, the plane-wave solution synthetic (figure 8c) reveals hardly any amplitudechanges. When the Poisson's ratio is changed to 0.25, reflecting the situation of non-gaseous sands, the P-wave reflection coefficients (figure 9a) are still negative but theyshow more variations with offset than that of the gaseous sands. However, the situationwith the synthetics is the same as before. The NMO-corrected synthetic (figure 9b) startsto show some significant amplitude changes at an offset of 1800 meters, with strong NMOstretch effects at 2400 meters and larger offsets. The corresponding plane-wave solutionsynthetic reveals no significant amplitude changes with offset.

Figures 10a, 10b, and 10c are the models for (C), which is a channel silty-shalewell. The P-wave reflection coefficients (figure 10a) for the Glauconitic Formation arepositive and start to change significantly only for offsets greater than 2400 meters. Boththe NMO corrected synthetic (figure 10b) and the plane-wave solution synthetic (figure10c) reveal insignificant amplitude changes as a function of offset.

Figures 1 la, 1lb, and 1lc are the models for (F), which is a Glauconitic regionalwell. The P-wave reflection coefficients (figure 1la) show similarity to those of (C). Thetwo synthetics (figures 11b and 1lc) show similar results to those of figures 8b and 8c. Inother words, the NMO corrected synthetics show some significant amplitude changes at1800 meters or larger offsets, but the plane-wave solution synthetic shows negligibleamplitude changes with offset.

The three sets of models seem to imply that any observable AVO effects for offsetsof 2500 meters or less are probably thin-bed tuning effects. The magnitude of the offset-

293

dependent reflection coefficients calculated from Zoeppritz equations do not show anysignificant P-wave changes for offsets below 2400 meters in all cases except in the case offigure 9a. Nevertheless, in all four cases, once thin-bed tuning effects are eliminated as inthe plane-wave solution synthetics, there are hardly any detectable amplitude variationswith offset.

Since the near-offset stack appears to have a higher frequency content than the mid-offset stack, the near-offset stack was then filtered four times, each time with a differentfilter. The results are shown in figures 12 to 15. In figure 12, where the filter is 8/16-35/40 Hz, the phase reversal disappears completely, while the drape is barely observablebut with clear amplitude changes. Obviously, the frequency content is too low to reveal theanomaly. In figure 13, the filter is 8/16-40/45 Hz. The anomaly on this stack shares someinteresting similarity to that in the mid-offset stack. Both stacks show broad wavelets forthe anomaly with similar looking bandwidths. This seems to imply that the differences inthe anomaly as it appears in the near-offset and mid-offset stacks can be partly attributed toa lower frequency content in the mid-offset stack. Figures 14 and 15 have filters 8/16-45/50 and 8/16-/50/55 respectively. They indicate that with frequencies higher than 45 Hzpresent in the data, the near-offset stack starts to develope a relatively distinct character forthe anomaly.

COMPLEX ATTRIBUTES

Another objective of this case history is to examine the complex attributes of theanomaly. In particular, we want to examine the complex attributes of the anomaly forvarious offsets. To make the discussion clearer, we shall discuss the attributes separately.The attributes are calculated by the GMA Grits package.

(a) Instantaneous amplitude

The instantaneous amplitude is sometimes called the amplitude envelope or thereflection strength. It is simply the amplitude of the complex trace, and is phase-independent (Taner, 1979). Figure 16, 17, and 18 are the amplitude envelopes for the full-offset stack, the near-offset stack, and the mid-offset stack respectively. The channel isclearly visible in both the full-offset and near-offset amplitude envelopes, but is virtuallyunobservable in the mid-offset stack amplitude envelope. Since the mid-offset stack has alower frequency content than the other two stacks, its amplitude envelope also appears tohave lower frequency content than the other two envelopes. The lower frequency contentseems to cause the disappearance of the channel signature in the mid-offset amplitudeenvelope.

(b) Instantaneous phase

Figures 19, 20, and 21 are the instantaneous phase plots for the full-offset stack,the near-offset stack, and the mid-offset stack respectively. As pointed out by both Tanner(1979) and Robertson et al. (1984), instantaneous phase is a very effective tool fordelineating discontinuities, faults, pinchouts, angularities and events with different dipattitudes. This is mainly due to the fact that the instantaneous phase is independent ofamplitude. Hence, discontinuities that are difficult to observe on conventional seismic datadue to low amplitudes will show up more clearly on phase displays. Therefore, the subject

294

channel anomaly should be nicely outlined on the instantaneous phase plots for the full-offset stack and the near-offset stack. In particular, on the one for the near-offset stack(figure 20), along the time of 0.98 to 1.0 second, discontinuities that are likely channeledge effects can be seen at trace 22 and between traces 71 and 85. However, the channelanomaly is not detectable in the mid-offset instantaneous phase plot (figure 21), nor arethere any channel edge effects. It also appears to have lower frequency content than theother two phase plots.

Incidentally, there is an event detected in the mid-offset stack phase plot thatappears to be very different in the near-offset phase plot. The event at about 1.020 seconds /between traces 53 to 71 on the mid-offset phase plot appears to have been truncated at bothends, and causes some drape on top. The corresponding event in the near-offset phaseplot, however, is rather continuous. This event is probably the Mississippian reflections.Given the differences in the Mississippian reflection characters on the two stacks, it is notsurprising that the corresponding instantaneous phase plots also exhibit differences. Sincethe Glauconitic reflection is typically only about 30 ms above the Mississippian reflection,it would be very useful to explain the AVO behavior of the latter.

(c) Instantaneous frequency

Figures 22, 23, and 24 are the instantaneous frequency plots for the full-offsetstack, the near-offset stack, and the mid-offset stack respectively. Since the instantaneousfrequency is the time derivative of the instantaneous phase, it is also amplitude-independent. For this attribute, only the near-offset instantaneous frequency plot showsthe anomaly unequivocally with a dipping event between traces 23 and 40 at about 0.98seconds. It is virtually absent on the other two frequency plots. This seems to imply that,among the three attributes, the instantaneous frequency is the best tool to outline thechannel anomaly. This is probably true whenever a thin bed is involved. Robertson et al.(1984) reported that as a bed thins to a quarter period of the dominant wavelength, there isan anomalous increase in instantaneous frequency, which remains high as the bedcontinues to thin. This agrees with Widess's (1973) conclusion that when the bed thins tobelow tuning thickness, the wavelet shape will assume the shape of its derivative andremain constant until the thickness approaches zero, while its amplitude will also decreaseto zero. As mentioned earlier, our subject gas-bearing sandbar is only 8 meters thick, andis below tuning thickness, assuming a peak frequency of 40 Hz. and a Glauconitic sandvelocity of 3700 meters/second, which gives a tuning thickness of 23 meters. Hence, itshould show up in the instantaneous frequency plot. However, it is interesting to note thatthe gas-sand anomaly appears only in the near-offset stack instantaneous frequency plot,but not in the other two frequency plots. Obviously, the instantaneous frequency attributeis very sensitive to the frequency content of the data. Note that the "low-frequencyshadow" reported by Taner (1979) is not observable on any of the frequency plots.

DISCUSSION

The Glauconitic sandbar in the subject well (E) exhibited an interesting anomaly onconventional seismic data. Although the channel section is 40 meters thick, only 8 metersof it are gas-bearing. Assuming a peak frequency of 40 Hz and a sand velocity of 3700meters/second, the tuning thickness would be about 23 meters. Hence, the gas-bearingzone is well below tuning thickness. This, in turn, means that to fully understand theamplitude behavior of the anomaly, one should also investigate the effect of tuning on

295

amplitude changes at non-normal incidence. In particular, in Alberta, the AVO effects ofmany thin hydrocarbon-bearing reservoirs are often negated by their tuning effects.Therefore, for thin beds below tuning thicknesses, no AVO analyses are complete withoutcorresponding thin-bed tuning analyses. Nevertheless, the AVO analyses performed on thethree wells are still informative in a qualitative manner. First of all, the analyses imply thatfor the Glauconitic Formation in Southern Alberta, AVO effects due to a lateral change oflithology can be observed, if at all, only for offsets well over 2500 meters. For our seismictemplate line, the largest offset is 2100 meters, with any larger offsets being muted out forthe Glauconitic reflection in order to cut down noise due to surface waves. Hence, onecould not observe any conclusive evidence for AVO effects as a result of a lateral change inthe Poisson's ratio of the gas-bearing sand between the near-offset stack and the mid-offsetstack, notwithstanding the presence of other character differences. This also means that inany Southern Alberta area where the surface-waves noise requires a relatively severe muteso that any offsets above 2500 meters have to be muted to below the Glauconitic reflection,the AVO method may not be an appropriate tool. Even for areas where the near-surfacecondition is excellent resulting in a very shallow mute, offsets much larger than 2500meters are probably required before any AVO effects could be clearly detected, if one takesinto consideration other noise problems present in seismic data.

Secondly, the AVO analyses clearly indicate that, in the subject area under study, asfar as the Glauconitic Formation is concerned, thin-bed tuning effects dominate any AVOeffects due to lateral changes in the Poisson's Ratio. Furthermore, NMO-stretch effects forlong offset traces are evident on all the synthetics. This also makes any AVO effects due tolateral changes of lithology more difficult to observe on large offset traces, since it lowersthe frequency of those traces. The differences of the anomaly between the near-offset stackand the mid-offset stack are probably largely due to lower frequency content in the mid-offset stack. On the other hand, the differences in the Mississippian reflection between thetwo stacks simply cannot be explained by a difference in frequency alone, nor can it beexplained by AVO effects, since apparent structural differences exist. Is it possible thatdifferent offsets would reveal structural elements preferentially? It would be interesting toinvestigate this possibility.

The complex attributes for the various offsets indicate that the channel can berecognized clearly on data with frequency of 45 Hz or higher present. The instantaneousphase outlines the lateral discontinuities at the channel edge remarkably well for the near-offset stack. In particular, the instantaneous frequency is very useful in delineating bedsthat are thinner than tuning thickness. Furthermore, the differences between theinstantaneous phase and frequency plots for the near-offset and mid-offset stacks suggestthat the attributes are very sensitive to the frequency content of the seismic data. Thus, itwould be a very useful exercise to study how the frequency content of the attributes relateto the frequency content of the corresponding seismic data.

Another potential use of the instantaneous phase and frequency is the possibility ofdelineating a wedge-shaped thin bed. Since the wavelet shape undergoes a definite patternchange as the bed thins from above tuning thickness to zero thickness, the instantaneousphase and frequency plots will outline the change very clearly, since they are independentof the amplitude.

296

CONCLUSION

This case study does not reveal any significant AVO effects due to a lateral decreasein the Poisson's ratio of the Glauconitic gas-bearing sands for the modified well (B) foroffsets up to 3000 meters. However, the differences in the characters of the anomalybetween the three offset stacks suggest that, even if we do not have offsets more than 3000meters, it is still worthwhile to study anomalies with partial stacks. Their usefulness isfurther enhanced by the use of their complex attributes, which outline discontinuities andthin beds preferentially with frequency content. Since near-offset traces tend to have higherfrequency content than mid-offset and far-offset tlaces, complex attributes are appropriatetools to be used in conjunction with other geophysical methods for analyzing partial stacks.

CURRENT RESEARCH

At present, the differences between full-wave equation modelling and ray-tracingmodelling are being investigated. The emphasis is on the differences between the P-waveand S-wave synthetics of thin-bed models generated by the ray-tracing method and thosegenerated by the full-wave equation method. In particular, AVO effects and complexattributes of these synthetics will be studied and compared. The results will also becompared to those generated by physical models. To generate synthetics, the Sierra ray-tracing modelling and full-wave equation modelling packages will be employed.

REFERENCES

Ostrander, W.J., 1984, Plane-wavereflection coefficients for gas sands at non-normalangles of incidence:Geophysics v.49, 1637-1648.

Robertson, J.D., Nogami, H.H., 1984, Complex seismic trace analysis of thin beds: Geophysics, v.49,p.344-352.

Taner, M.T., Koehler, F., and Sheriff, R.E., 1979, Complex seismic trace analysis: Geophysics, v.44,p.1041-1063.

Widess, M.B., 1973, How thin is a thin bed?: Geophysics, v.38, p.1176-1180.

297

ACKNOWLEDGEMENT

We thank Seis-pro Consultants Ltd. for providing excellent processing of theseismic line and offering many valuable suggestions during the development of the project.We also want to thank Alberta Energy Company and Summit Resources for allowing theuse of one of their seismic lines. Special thanks are extended to Mr. Andrew Mah and Mr.Brian Doherty for organizing the various materials.

298

__ m Glauconite-- Mississippian

CoO

full-offset stack

0-2100 meters

Figure 4. Full-offset stack of template seismic anomaly; normal polarity.

Glauconite-- Mississippian

C_

near-offset stack

O- 1050 meters

Figure 5. Near-offset stack of template seismic anomaly, normal polarity

_llpr

(l( /////r((l(llll /(1 (/(.f/f/( /i* /(/( //f(( /( ///_iI* '_

C_

....... _F<( _r (zr-_-rf _........... frrrrrr "[il r( r_r_*%_\l_.\%\_.kll/Jlll'_f_/_J]ll/'(,\L(]l_l__,. ,,,,..,.,,,,,

II(((/I(lllll(I

B&&,M III b_, ! Pl IIg & i &k BLi i_ l_D_l_&k//B

_ I _ ( )Ylf/ I _ I I _111 Ifll III III I /f I 7 I{ I I 11 1(¢ f _ I I ( _ _ I I

,.mid-offset stack

1050-2100 meters

,Figure 6. Mid-offset stack of template seismic anomaly, ,normal polarity

CDP B"; CDP D " CDP C CDP A

regional channel sandbar channel edgo .regional

O. 700 Oo7(20

0.000 0o000

0.900 0,900

Glauconlte

o loMlsslssipplan

I.I00 i.i00

Figure 7. Ostrander gather, summing over six CDP]ocations and

three-offset range with trace-balancing, normal polarity

J

channel gas well, 8 meters sand, Poisson's ratio 0.1ZOEPPRITZ SIHGLE II41ERFQCE

• S S D S-14AUE1 POISSOt.I Event: 36 Tire',e: 882 Depth: 1438= us/m US/r_: /c," u_/ta EU Q.200

............... 15 8.12016 R

.........l..................'v. _°'°_°...........................................17 ' :_&'HI ._T_SR:

680... 0.040 "l''"_'_.............18 _ .............................

........................................ t -0.040i

._. _ 21 y-0.080........ 23 _-0.128o

o_ 24 _-0.160

.._ ......_....._.........._ ........__ -o._oo_1 .=] _ J PR_ .__ _ 27 -0.240 ...........................................

.°°.-<\ ::___ 1 31 680 1268 1800 2400

...._ ....._ J_..._ ....._. ....... OFFSET(meter.)32 Re£ I ected F'-l|ave35

_4o I--_,_...l_:,;iL I ......_'.o,',eo*ed_-,_vo220 2 30 2.7 410 .5

Figure 8a. Reflection coefficients versus offsets for

the Glauconlte Formation in modified well (B)

channel gas well, 8 meters sand, Poisson's ratio 0.1ZOEF'PRITZ REFLECTIUITY i'IODEL

S - ,S _) S-HRUEI POISSOH HOBEL1 (rae_'er-)500 ,,J_."r,,_ us/rf, g/cc u_/n_ EU 2896 2482 2068 1655 1241 827 413 0

r- _ _J L d

.... ....... ...... .......... ............ 17ooo...-_, ........ 5 ........ I ........... q .............. ioh 'l h 20

-_-_ _ S I :_8 0 0 • _ ....... q ...... ] ......... _ ......... I .....

.._._ L._, 1..1 L_ 31-_-_. .I J J_--S_'-- _ I _i _Glau =onlte,_,

..... I_°°__o .....230 2.7 410 .5

Figure 8b. NMO corrected synthetic for modified well (B)

• i "•,

channel gas well, 8 meters sand, Poisson's ratio 0.1i

ZOEPPRITZ REFLECTIUITY IIODEL

S S "D S-_.IFIUEIPOISSOII MObELI (mel'ers)U_,/I,I US.-'m ,Zl/CC: uslh', EU 2"482 20G9 ! G55 | 241 _2? 4 | ;3 lj

500 .......... - _ '"----"- 14

........J ......_F............]...,...._.....__i-.

...................... 18GO0............... •

7r_o... :_ ....... .] ........ ] ......... ] ............. zs- _ k I L

o t_ ....,]........._--I...........27

iiiiI

220 230 2,7 410 .5 ,

Figure 8c. Plane-wave solution synthetic for modified well (B)

channel gas well, 8 meters sand, Poisson's ratio 0.25ZOEPPRITZ SIHGLE IIITERFRCE

'S S D $-WAUEi POISSOH EL_I Even_: 36 Time: 882 Dep'th: 1438

"v'v'-FT -F--F 23 o.,00........_....................................."\........h.........i.....h........,.._4o.1_o.....:.................................

_so _=.__ 2_ _ o.090 _Ri........

.... .... ...... ............° / .... i i i27 c 0.030 :i................................

/v -0.0313 ..................

80 . . • -¢ -0.06g : •.e.s.nPRY -e.¢�e _-_._@..... i ........ :....o 31 -0.120 ..... : ........ : ........ :........LO

-0450 .-.!........ :........ _...... '..!........

_"° .... _ ..... -I .... q .......... 't ........... -o.1_,o ...! ........ _....... _........ _..............,_;......._......_......._ .........,_ o_,_,o._oo,>

_-_ _ _ --_ " 37 -- R.e'f'lec'l:ed P-Have

,_o......._ .........J....r.I...........-rJ.........3__oo 23° _._ ,,,0 ._

Figure 9a. Reflection coefficients versus offsets for theGtauconite Formation in mofiltted well (B)

/.

(

channel gas well, 8 meters sand, Poisson's ratio 0.25ZOEF'PRITZ REFLECTIUITY rlOBEL

S S B S-I,IRUE1 POISSON IIOTEL1 (f_e_ers)5_0 U__zr,i U_/r_l 9/CC us/r,_ ELI ._896 2482 2068 1655 1241 82"7 413 El

I _.J..'.J LE [ l".....-..,........:].......I ..........-3........ 15.... ]6

18

.... ........ ........f ......... ........_ ....... 221_° _........L ........_........ ="_o ..................... 26

=..-_ __ _ 27

_ee...__ .........................

9e_..... -_ ...... .-'/','_..T/_-_.. 38b48 I b38 I'b.3 '} ).

220 230 2.7 418 .5

Figure 9b, NMO corrected synthetic .for modified well (B)

channel gas well, 8 meters sand, Poisson's ratio 0.25ZOEF'PRITZ REFLECTIUITY MODEL

S "S B S-I.IAUEIPOISSON MOIIELI (rne't'ers) IUS/h'a u:_/'rn I"/'2'z U_-/r, EU :3808 2588 2172 1758 1344 9.31 517 I0:_'_'50O

,_' F_ FJ -- 14%- k h 15

860...'_,_........ "l'................. l" .............. 18

)' r_............d........_.....-_'7oo .... _............................ .1 ........ ,..... 2s

k [ L I 74-'=. _] I _ I -

31

.....__=........................................

900 ...... _ .......

4_q;_-_--I l _o lb. I:220 230 2.7 410 .5

Figure 9c. Plane-wave solution synthetic for modified well (B)

channelshale wellolIl_LE IHTERFRCEZOEPPRITZ (" "

S S I ,D S-WRUE1 POISSOH Euer, t: 55 Time: B74 Depth: 14005_ us/m us/m g/co U_/ra ELI 015_ .........................................

u_---"-_ "-7 --- 31 0.450 ........"........................."........

G ........ L_ .... 32 0.400 ........ i ........ ; ....... i ........ :........

•........... i i :iiiiiiiiiiiiiiiiiiiiii i y-t'" 3_ R 0.350

I 3_, _ 0.300688 ............. •...... _ _._50 ........:........_......._........:.......

3_37_e0.200_._ ......._;.......Ji........_....i"_................._..__'=_............_ ............... 0.150

39 0.050 ................. _....... ; ........ :........ :700................................ 4e ! i : _SR:

41 Y 0.000

....................... 4E;_ 47 -0.100

80o..._ _ _ _ Sl -O.lS0.............. ................... -0.200

'___ _ 53 G00120018002400

_i r_ _I ] , OFFSET (.e±er.)54 Re¢Iec_ed P-Have57

"I'_ hj " - ..... Re¢lec*ed S-Wave210 220 2.7 390 .5

Figure lOa. Reflection coefficients versus offsets for theGlauconite Formation In channel shale well (C)

channel shale well

ZOEPPRITZ REFLECTIUITY MODEL

£ S D S-HRUEi POISSOH MOBEL1 ( raeter.s )us/r,1 9,',-_c us/n_ EU 2896 2482 2068 1655 t241 827 413 8508

ii _" . . . , 32

:3435

co -_ _ r

c_ _ _ I I 475t

800" • "_ ................-.c... 53

210 228 2.7 390 .5

Figure lOb. NMO corrected synthetic for well (C)c

channel shale wellZOEPPRITZ REFLECTIUITY IIOI'IEL

.S S D S-WAUE I PO ISSOII MODELI (rnei'er_>us/rn us..'r,_ Q/ec us/n_ EU 2482 2069 t655 1241 827 413 El

5oo---_._- --7_ --:--T- --7

6g36

.... _:_ ....... :].......... ] ....... q ........ j ..... 3_t [ l [--- 38

7,3e..._.. ....... ] .......... 1 ......... ] ............ 39

---.-_. _ _-_ _.

808 ....... "_L...................

_ _J _- _J ela. ._e ,_4

..... i, _--4-_--"-132°h i_" L .....I 210 220 2,7 398 .5

Figure 10c. Plane-wave solution synthetic for well (C)

regional wellZOEPPRITZ SIIIGLE IIITERFFICE

S S D S-LIRUEI POISSOII Even¢: 51 Time: 854 Depth: 1382us/n_ us/m g/cc us/m EU 0.200 .........................................500

_- 28 0.180 ........ :........ ; ....... _ ........ ;........0. 180

i i iiiii Iii iill I .................................i........_. 8.128 ........ :........ } ........ : ........ "........ _PR

_0 i e.le_........ i........ i ........ i............. _K31 e 0.080 ................ _ ........ • ........ ...._...

o.860 :-_e_:-, ..,....... :.......32i 8.848 ........ :......-='_-u:'_B-BB.la_i_T-...; ............................... 33 V ....i 8. 028 ........ :........ ;. ....... - ........ :........

34 i- 0.880 _SR"

..... 36 -0.020_j 37 -8,040

38 -0.080

i -8.080 ................ "......... ' ........ "........41 -0. 100 ....... : " : .;..

"_-'_ _-'J __" _ _-_"_- 4842 _:e£, ec±ed0 6801200,8882488P-Have

88 ..........................

43 OFFSET (rile'f,er.s)45

--- =_ s2L. --1-- i

190 228 2.7 390 ,5

Figure 11a. Reflection coefficients versus offsets for. the

Glauconite Formation in regional well (F)

regional wellZOEPPRITZ REFLECTIUITY I'IOBEL

'S S B S-IIAUE1 POISSON MODELI ,(r_e't'ers).-,00=" us/r_'tz u._/Jn g/co u_/m ELl 2482 2068 1655 1241 827 41:_'3

-:" ._' 28

.... _ ......... _ ......... " ....................... 30f

600 ..... ! ..............................

j.u 32

...._ ............... I.................... 3433TOO.-.;,_........,j......_ ..........._....

-_ E_ E- E- 36

41

.:,-__ _= _ _= 4s...... _ .......... 3 ...... _. .......... ,_ z,o.nlt_'8-zc •

_--. _ --.----7 _-----_ 527 ,

!90 22_ 2.7 398 .5

Figure 11b. NMO corrected synthetic for well (F)

regional wellZOEF'PRITZ REFLECTIUITY I'IODEL

S S " D S-LIRUEI POISSOH IIODELI (n_ei-ers)

580 u'-_/u, u_.-"r,_ g./c,: u_/r_', ELi = 2896 2482 2069 1855 1241 827 413

,_ _ _ ..__ , , , ,,_ ;c_..... ,,,,,r, r, , , ,

I' II] il I I I

t rr, I_o I I, _, ,6[38( •., .... ].l.q.l.IJ..... ..........

._" []. ] 31 _I _ , i F i, r I I I I . , , ,i_ I I i i_

" ?'4 " " I r ! I I 4 I I _ , , , I I '

'_ k "_t _Il'"!,,'_'_''___"t_''___'

r- (

198 228 2.7 398 .5

Figure 11c. Plane-wave solution synthetic for well (F)

_ _,,,, _,, __._,_ _J1_ Glauconite

near-offset stack

8/16-35t40 Hz

Figure 12. Filtered near-offset stacl_

___ Glauconite_ Mississippian

near-offset stack

8/16-40/45.Hz

Figure 13. Filtered near-offset stack

m Glauconite

ssissippian

0

near-offset stack

•8/16-45/50 Hz

Figure 14. Filtered near-offset stack

____ --O,a,,ooo,,e_M=sslssippian

near-offset .stack

8/16-50/55 Hz

Figure 15. Filtered near-offset .,stack

Figure 16. The instantaneous amplitude plot for the full-offset stack

.*OCO+*COO(O*•<*•Ok.(OoCOa9T3Q.

E«o(O3OO)C(O+*C(O+*WO£Hhi^Q)k.3O)

32

3

Figure 18. The instantaneous amplitude plot for the mid-offset stack.

Figure 19. The instantaneous phase plot for the full-offset stack

(.0rocr. Figure 20. The Instantaneous phase plot for the near-offset stack,

iMirtW^iv-^ww^NV^tt \v^^J!!^^^;^^^^\^^\W^^ftW'tott

, . 4 i f \ « ^ i * t i t * » t • . » ,\nUl' » " '•"" Mt t 'M* \ t . | ^ | ^,„„. . . . ^ mm\\m^m\m\m

W^\^^^^^^>^^w ln^w^\^^^v^^^^^^'\^^^\^'^^U^^^u<

> .^>»/-j'.'-VXvS)A.'...- .!AfAf.^

*»•*»*•••**»*•*•*•••<•««••••••••*-

ivnri

Figure 21. The instantaneous phase plot for the mid-offset stack

Figure 22. The instantaneous frequency plot for the full-offset stack

Figure 23. The instantaneous frequency plot for the near-offset stack

Figure 24. The instantaneous frequency plot for the mid-offset stack,